Need a FORTRAN compiler for Win7 (or XP)

I have a FORTRAN program dating back to 1964, and would like to compile 
it. Win7 is my first choice, but XP Pro is OK. Does such an animal 
(compiler) exist?
0
wolftracks (253)
8/16/2011 6:05:54 PM
comp.lang.fortran 11340 articles. 37 followers. Post Follow

70 Replies
1404 Views

Similar Articles

[PageSpeed] 56
On Tue, 16 Aug 2011 11:05:54 -0700, W. eWatson wrote:

> I have a FORTRAN program dating back to 1964, and would like to compile
> it. Win7 is my first choice, but XP Pro is OK. Does such an animal
> (compiler) exist?

google gfortran
google 'windows 7 fortran'

-- 
steve
0
sgk (134)
8/16/2011 6:35:00 PM
W. eWatson <wolftracks@invalid.com> wrote:

> I have a FORTRAN program dating back to 1964, and would like to compile 
> it. Win7 is my first choice, but XP Pro is OK. Does such an animal 
> (compiler) exist?

If it is real Fortran II, then it might take a little work with
modern compilers, but probably not so much.   There might be a few
features of Fortran IV that didn't make it into later standards.

gfortran is the popular free Fortran compiler these days.

-- glen
0
gah (12850)
8/16/2011 7:16:50 PM
On 8/16/2011 1:05 PM, W. eWatson wrote:
> I have a FORTRAN program dating back to 1964, and would like to compile it. Win7 is my first choice, but XP Pro is OK. Does such an
> animal (compiler) exist?

www.openwatcom.org has an F77 compiler that will
give you a fighting chance on Windows.

Lynn

0
lmc (231)
8/16/2011 7:33:58 PM
On 8/16/2011 11:35 AM, Steven G. Kargl wrote:
> On Tue, 16 Aug 2011 11:05:54 -0700, W. eWatson wrote:
>
>> I have a FORTRAN program dating back to 1964, and would like to compile
>> it. Win7 is my first choice, but XP Pro is OK. Does such an animal
>> (compiler) exist?
>
> google gfortran
> google 'windows 7 fortran'
>
Well, the first many hits with just gfortran and windows gave me gcc. 
I'm not looking to install Cygwin. The second was pretty much like the 
first. I have no idea if they will accept old ftn.
0
wolftracks (253)
8/16/2011 8:46:32 PM
On 8/16/2011 12:16 PM, glen herrmannsfeldt wrote:
> W. eWatson<wolftracks@invalid.com>  wrote:
>
>> I have a FORTRAN program dating back to 1964, and would like to compile
>> it. Win7 is my first choice, but XP Pro is OK. Does such an animal
>> (compiler) exist?
>
> If it is real Fortran II, then it might take a little work with
> modern compilers, but probably not so much.   There might be a few
> features of Fortran IV that didn't make it into later standards.
>
> gfortran is the popular free Fortran compiler these days.
>
> -- glen
Is gftn ready to go in a pure Windows environment w/o Cygwin? There's a 
lot of noise on some forums about ftn, but little to tell me where I can 
find the compiler.
0
wolftracks (253)
8/16/2011 8:49:37 PM
On 8/16/2011 12:33 PM, Lynn McGuire wrote:
> On 8/16/2011 1:05 PM, W. eWatson wrote:
>> I have a FORTRAN program dating back to 1964, and would like to
>> compile it. Win7 is my first choice, but XP Pro is OK. Does such an
>> animal (compiler) exist?
>
> www.openwatcom.org has an F77 compiler that will
> give you a fighting chance on Windows.
>
> Lynn
>
I don't see 77 mentioned there, and it looks like one gets source code. 
If that's right, then what compiles it?  The README talks to DOS, OS2, 
Win32, ... and others. Will this run under Win XP and above? Hmm, I 
think I'm looking at a README for C++. Is C++ somehow used as the 
compiler to get a F77 compiler (executable)?
0
wolftracks (253)
8/16/2011 8:57:57 PM
W. eWatson wrote:
> Well, the first many hits with just gfortran and windows gave me gcc.
> I'm not looking to install Cygwin.

Well, instead of Cygwin, you could/should use the MinGW version. See 
also http://gcc.gnu.org/wiki/GFortranBinaries

There are also other options such as buying one of the several 
commercial compilers or (less maintained) other free/open-source compilers.

> The second was pretty much like the
> first. I have no idea if they will accept old ftn.

It mainly depends on how old - for Fortran I you probably need some 
modifications, e.g. to comment the FREQUENCY statement. And it depends 
whether the program was developed to run on several platforms/with 
several compilers or not.

If it was programmed for a single compiler, the chances are high that 
the code relies on specific behaviours of the compiler - or on a 
not-widely-supported vendor extension.

Otherwise, Fortran (the language standard) is largely backwards 
compatible and the few things which were deleted are still mostly 
supported by the compilers.

gfortran - as many other Fortran compilers - supports several vendor 
extension. Hence, the change is rather high that your program works.

You might, however, need to use special flags such as -fno-automatic (if 
the program assumes that all local variables are in static memory/have 
the SAVE attribute) or such as -finit-local-zero (if the program assumes 
zero initialization of variables). Or, you might need to add some 
EXTERNAL statements for user functions which are also intrinsic 
functions of the compiler.

If you unlucky, the vendor extension you need might be unsupported. For 
instance, doing a single trip for "DO i = 1, 0" is not supported by 
gfortran, which follows Fortran 77+ (?) by running such a loop zero 
times. (Other compilers have some flag forcing this behaviour which some 
Fortran 66 compilers had.)

Tobias
0
burnus (564)
8/16/2011 9:26:02 PM
On 8/16/2011 3:57 PM, W. eWatson wrote:
....

> I don't see 77 mentioned there, and it looks like one gets source code.

Well, since OW is a F77+ compiler, what is/isn't mentioned doesn't 
matter _too_ much... :)

Looks like you might need to just slow down and actually read a little 
of what is on a page...you _can_ download source but there's no need.

> If that's right, then what compiles it? The README talks to DOS, OS2,
> Win32, ... and others. Will this run under Win XP and above? Hmm, I
> think I'm looking at a README for C++. Is C++ somehow used as the
> compiler to get a F77 compiler (executable)?

That isn't right (only, any way) so question is moot.

OW can generate executables for all of the above OS'es; you select the 
target desired.

<http://www.openwatcom.org/index.php/Download>

--

0
none1568 (7455)
8/16/2011 9:30:33 PM
On 8/16/2011 4:30 PM, dpb wrote:
....

> <http://www.openwatcom.org/index.php/Download>
....

Looks like you would want

<ftp://ftp.openwatcom.org/open-watcom-f77-win32-1.9.exe>

BTW, as others have noted; Fortran Standards have maintained backward 
compatibility of most features since F66; even those features which have 
been deleted or deprecated since F90+ have almost universally been 
retained by various compilers.

OW is no different in that regard; it is a F77-compliant compiler and 
has many extensions beyond; what your code might have used is anybody's 
guess and will depend on what machine/OS/compiler(s) was/were used.

You'll simply have to see what is actually in the code; the easiest way 
is to first just throw it at the compiler and see what it complains 
about then see what you have to change either in the source code or in 
options within the compiler if it happens to support either used 
extensions or obsolete features.

There are some things that will possibly be painful but there's no use 
speculating until one has actual data to look at (namely either source 
or better yet, error diagnostics to go with the code)...

I'd give OW a shot, certainly...

--

--
0
none1568 (7455)
8/16/2011 9:51:48 PM
On Aug 16, 5:49=A0pm, "W. eWatson" <wolftra...@invalid.com> wrote:
> On 8/16/2011 12:16 PM, glen herrmannsfeldt wrote:> W. eWatson<wolftra...@=
invalid.com> =A0wrote:
>
> >> I have a FORTRAN program dating back to 1964, and would like to compil=
e
> >> it. Win7 is my first choice, but XP Pro is OK. Does such an animal
> >> (compiler) exist?
>
> > If it is real Fortran II, then it might take a little work with
> > modern compilers, but probably not so much. =A0 There might be a few
> > features of Fortran IV that didn't make it into later standards.
>
> > gfortran is the popular free Fortran compiler these days.
>
> > -- glen
>
> Is gftn ready to go in a pure Windows environment w/o Cygwin? There's a
> lot of noise on some forums about ftn, but little to tell me where I can
> find the compiler.

Just in case:

http://tdm-gcc.tdragon.net/download

download On-Demand Installer (Downloads and installs the components
you select)

tdm-gcc-webdl

run tdm-gcc-webdl.exe

Select all (which I did), or gfortran + OpenMP, or...

done (at least for me, on Win7)

Warning: this just works, I'm not an expert.

c:\>gfortran -v
Using built-in specs.
COLLECT_GCC=3Dgfortran
COLLECT_LTO_WRAPPER=3Dc:/mingw32/bin/../libexec/gcc/mingw32/4.5.2/lto-
wrapper.exe
Target: mingw32
Configured with: ../../src/gcc-4.5.2/configure --build=3Dmingw32 --
enable-language
s=3Dc,c++,ada,fortran,objc,obj-c++ --enable-threads=3Dwin32 --enable-
libgomp --enabl
e-lto --enable-fully-dynamic-string --enable-libstdcxx-debug --enable-
version-sp
ecific-runtime-libs --with-gnu-ld --disable-nls --disable-win32-
registry --disab
le-symvers --disable-werror --prefix=3D/mingw32tdm --with-local-prefix=3D/
mingw32tdm
 --enable-cxx-flags=3D'-fno-function-sections -fno-data-sections' --with-
pkgversio
n=3Dtdm-1 --enable-sjlj-exceptions --with-bugurl=3Dhttp://tdm-
gcc.tdragon.net/bugs
Thread model: win32
gcc version 4.5.2 (tdm-1)

HTH

Fernando.
0
ftinetti (148)
8/16/2011 10:07:42 PM
Tobias Burnus <burnus@net-b.de> wrote:

(snip)
> It mainly depends on how old - for Fortran I you probably need some 
> modifications, e.g. to comment the FREQUENCY statement. And it depends 
> whether the program was developed to run on several platforms/with 
> several compilers or not.

(snip)
> Otherwise, Fortran (the language standard) is largely backwards 
> compatible and the few things which were deleted are still mostly 
> supported by the compilers.

The Fortran II READ and PRINT statements that didn't make it
into Fortran 66 were added back later.  

If it does tape or drum I/O, though, it might need some work.

> gfortran - as many other Fortran compilers - supports several vendor 
> extension. Hence, the change is rather high that your program works.

> You might, however, need to use special flags such as -fno-automatic (if 
> the program assumes that all local variables are in static memory/have 
> the SAVE attribute) or such as -finit-local-zero (if the program assumes 
> zero initialization of variables). Or, you might need to add some 
> EXTERNAL statements for user functions which are also intrinsic 
> functions of the compiler.

Remove the trailing F from all the function names.

-- glen
0
gah (12850)
8/16/2011 10:30:43 PM

"W. eWatson"  wrote in message news:j2ebiq$v80$1@dont-email.me...

I have a FORTRAN program dating back to 1964, and would like to compile
it. Win7 is my first choice, but XP Pro is OK. Does such an animal
(compiler) exist?

As others have written, almost any modern Fortran compiler should work. For 
ease of installation, there are packages of Gfortran, g95 and OpenWatcom 
that are point and click easy.

The most trivial installation is the 2.95 version of G77 but that is Fortran 
77 with extensions. I would recommend that only if you are totally stuck 
with not being able to use one of the others.

You might also want to look for a compiler with good debugging facilities, 
such as Silverfrost (aka Salford). One version is free for non-commercial 
use.

What other problems might you have that are NOT going to be solved by your 
compiler? I can think of a few.
These are system specific and architecture specific.

1. Unit numbers that need to be changed.
2. Differences in I/O for example in OPENING files, and so on.
3. Carriage control on output.
4. Using arrays INTEGERs or REALs for what now would be CHARACTER variables. 
Also various schemes depending on word size, character set, packing and so 
on. [Yes, Hollerith will still compile.]
5. Machine specific constants.
6. Bit manipulation.
7. Equivalence and Common that depend on architecture.
8. Vendor specific extensions. Non-standard DATA statements, intrinsic 
functions, etc.
9. Non-standard declarations for single vs double precision, and so on.
10. Issues related to data files versus cards?
11. Mag tapes?

Fortunately there are a number of people who frequent this newsgroup with 
lots of experience getting old programs to run again. In particular, 
Terence, who labored mightily to bring a version of BMD back up, and others. 
Lynn McGuire porting enormous programs to OpenWatcom. And others.

-- e




0
epc8 (1259)
8/16/2011 10:47:14 PM

"glen herrmannsfeldt"  wrote in message 
news:j2er2j$62e$1@speranza.aioe.org...

Tobias Burnus <burnus@net-b.de> wrote:

(snip)
> It mainly depends on how old - for Fortran I you probably need some
> modifications, e.g. to comment the FREQUENCY statement. And it depends
> whether the program was developed to run on several platforms/with
> several compilers or not.

(snip)
> Otherwise, Fortran (the language standard) is largely backwards
> compatible and the few things which were deleted are still mostly
> supported by the compilers.

The Fortran II READ and PRINT statements that didn't make it
into Fortran 66 were added back later.

[snip]

Remove the trailing F from all the function names.

??? ISTR using FORTRAN IV rather than FORTRAN II back then (62-63). No 
trailing F on function names at all.
Later there were CALL OPEN() and CALL EXIT() and so on....

-- e




0
epc8 (1259)
8/16/2011 10:53:36 PM
The code ran on a CDC 1604. What is below is from an OCR I program and 
scanner (Sony V500) I used to extract some code. As can be seen it needs 
to be cleaned up a little. For example the ATAN line below, 3-4 lines, 
has some <>, which are supposed to be ().  However, I'm just trying to 
give some idea of what code looks like. It is not indented properly 
either, but I supposed a program like tidy can do that. There are two 
code samples.

This could be a mess if some of the OCR inserts really odd characters 
for what it thinks it sees. The left margin does not line up well. I'm 
copying this from an rtf file, which probably has some strange 
characters in the left margin. There's certainly room for improvement.

Code sample 1.

PERIOD = 0.0
EFN = -EFN 661
EA = 2.*ATANF(SQR7F{EFN/EFP>*<ESINTA/<EX+ECOS7A>))
DDELT = (EA - EX* SINF(EA))*SQRTF(ACUBF) / 0.01720209895
T = XJDM-DDELT
PRINT 803*OBS(1),METNO<1)
PR I NT 804,ARAD,DRAD,XRAD* YRAD,ZRAD,SINEQ,COSZD* ARC,DECRC , XRC , 
1YRC , ZRC, ARCC, DECRCC,XRCC,YRCC ,ZRCC,SDELTZ,V INA
PRINT 805,VCOR,VCC,XRB ,YRB, ZRB,RX,RXP,RYiRYP^RZ*RZP,VHELIO
PRINT 806 , A ,EX,QUE, T , SMEGA tCOMEGA^OI iPEZRI OD
END
SUBROUTINE PRECCO (T)
DIMENSION PREC (3,3)
COMMON PREC
T2=T*T
T3=T2#T
PREC( 1*1 ) = !.-( .00029696*T2)-( .00000014*T3)
PREC(1,2)=.02234941#T+.00000676*T2-.00000221*T3
PREC(1,3)=.00971691*T-.00000206*T2-.00000098*T3
PREC ( 2 9 1 ) :=-PREC ( 1 t 2 )
PREC(2,2)=1*~.00024975*T2-.COQ00015*T3
PREC(2,3)=-�00010858*72
PREC(3 ,1 )=PREC(1,3)
PREC(3,2)=PREC(2,3)
PREC(3,3)=1.-.00004721*T2~.00000002*T3
RETURN
END
FUNCTION FPRIMF(FP1,FM1,F2P1,F2M1)
FPRIMF = ((FP1-FM1)/2.-(F2P1-F2M1)/l2.)/4.
END
FUNCTION XINTPF(FO,F1,F20,F21,S)
XINTPF = FO + S*(F1-FO >-(S* < S-l � )*(S-2. )*F20 + S*(S+1�)*(S-l�)*F21 
)/6,

Code sample 2.
....

END
FUNCTION CONVF(SiTM�O)
CONVF = CD#3600�+TM*60*+S)/206265.4
END
SUBROUTINE PRECES <X1�Y1�Z1)
DIMENSION PREC (3,3)
COMMON PREC
X2=PREC( 1 � 1 )#X1+PREC( i �2 )*Y1+PREC( 1 �3)*Z1
Y2 = PREC( 2, 1 )*Xl+PREC<2�2)*Yl+PREC(2f3)*Zl
Z2 = PREC( 3t 1 )*X1+PREC(3�2)*Y1 + PREC ( 3 , 3 ) *Z 1
X1=X2
Y1=Y2
Z1=Z2
RETURN
END
SUBROUTINE CLEAR
DIMENSION DUMMY 1 (9) �ANE(3,4)
COMMON DUMMY 1 �ANE�BYPASS
BYPASS=0.
DO 122 L=l*3
DO 122 M=1 � 4
122 ANE<L�M)=0.
RETURN
END
SUBROUT INE NORMEC"3( XN,A,B,CfW,K,KE)
DIMENSION DUM2(9)tANE(3t4)�BP(100)
COMMON DUiV2� ANE � BYPASS �BP
IF(KE)715,124,125 125 BYPASS = BYPASS-H .
BP(K)=1,
GO TO 126
124 BP(K)=0.
ANE < 1 � 1 ) =ANE(1*1)+A#A*W
ANE (1*2) =ANE (1,2 )-�-A*B*W
ANE(1,3)=ANE(1�3)+A*C*W
ANE(1�4)=ANE(1*4)+A*XN*W

....
0
wolftracks (253)
8/17/2011 12:38:16 AM
On 2011-08-16 19:53:36 -0300, e p chandler said:

> "glen herrmannsfeldt"  wrote in message news:j2er2j$62e$1@speranza.aioe.org...
> 
> Tobias Burnus <burnus@net-b.de> wrote:
> 
> (snip)
>> It mainly depends on how old - for Fortran I you probably need some
>> modifications, e.g. to comment the FREQUENCY statement. And it depends
>> whether the program was developed to run on several platforms/with
>> several compilers or not.
> 
> (snip)
>> Otherwise, Fortran (the language standard) is largely backwards
>> compatible and the few things which were deleted are still mostly
>> supported by the compilers.
> 
> The Fortran II READ and PRINT statements that didn't make it
> into Fortran 66 were added back later.
> 
> [snip]
> 
> Remove the trailing F from all the function names.
> 
> ??? ISTR using FORTRAN IV rather than FORTRAN II back then (62-63). No 
> trailing F on function names at all.
> Later there were CALL OPEN() and CALL EXIT() and so on....
> 
> -- e

In Fortran II the trailing F indicated the use of a nonstandard calling
sequence. The usual calling sequence was for any number of arguements and
used a list of addresses. The nonstandard method was to put the single
argument in the accumulator.

So one called SINF but the library had a SIN that took its arguement from
the accumulator.



0
8/17/2011 12:47:14 AM
W. eWatson <wolftracks@invalid.com> wrote:

> The code ran on a CDC 1604. What is below is from an OCR I program and 
> scanner (Sony V500) I used to extract some code. 

(snip)

> PERIOD = 0.0
> EFN = -EFN 661
> EA = 2.*ATANF(SQR7F{EFN/EFP>*<ESINTA/<EX+ECOS7A>))
> DDELT = (EA - EX* SINF(EA))*SQRTF(ACUBF) / 0.01720209895

Fortran II function names with trailing F.

> PRINT 805,VCOR,VCC,XRB ,YRB, ZRB,RX,RXP,RYiRYP^RZ*RZP,VHELIO

Fortran II I/O statements, which are supported.

> FUNCTION FPRIMF(FP1,FM1,F2P1,F2M1)
> FPRIMF = ((FP1-FM1)/2.-(F2P1-F2M1)/l2.)/4.
> END
> FUNCTION XINTPF(FO,F1,F20,F21,S)
> XINTPF = FO + S*(F1-FO >-(S* < S-l ??? )*(S-2. )*F20 + S*(S+1???)*(S-l???)*F21 
> )/6,

Again, trailing F on function names.  I believe the rule is that
INTEGER functions start with X, and REAL without X.

That you will have to fix.  Declare the X functions INTEGER where they
are called, and INTEGER FUNCTION XINTP(...) here.

-- glen
0
gah (12850)
8/17/2011 12:50:18 AM
On 8/16/2011 5:50 PM, glen herrmannsfeldt wrote:
> W. eWatson<wolftracks@invalid.com>  wrote:
>
>> The code ran on a CDC 1604. What is below is from an OCR I program and
>> scanner (Sony V500) I used to extract some code.
>
> (snip)
>
>> PERIOD = 0.0
>> EFN = -EFN 661
>> EA = 2.*ATANF(SQR7F{EFN/EFP>*<ESINTA/<EX+ECOS7A>))
>> DDELT = (EA - EX* SINF(EA))*SQRTF(ACUBF) / 0.01720209895
>
> Fortran II function names with trailing F.
>
>> PRINT 805,VCOR,VCC,XRB ,YRB, ZRB,RX,RXP,RYiRYP^RZ*RZP,VHELIO
>
> Fortran II I/O statements, which are supported.
>
>> FUNCTION FPRIMF(FP1,FM1,F2P1,F2M1)
>> FPRIMF = ((FP1-FM1)/2.-(F2P1-F2M1)/l2.)/4.
>> END
>> FUNCTION XINTPF(FO,F1,F20,F21,S)
>> XINTPF = FO + S*(F1-FO>-(S*<  S-l ??? )*(S-2. )*F20 + S*(S+1???)*(S-l???)*F21
>> )/6,
>
> Again, trailing F on function names.  I believe the rule is that
> INTEGER functions start with X, and REAL without X.
>
> That you will have to fix.  Declare the X functions INTEGER where they
> are called, and INTEGER FUNCTION XINTP(...) here.
>
> -- glen
So it's ftn II.

It's an EPSON V500.

It's been a very long time since I used FORTRAN.
0
wolftracks (253)
8/17/2011 1:06:20 AM
On Aug 17, 6:46=A0am, "W. eWatson" <wolftra...@invalid.com> wrote:
> > google gfortran
> > google 'windows 7 fortran'
>
> Well, the first many hits with just gfortran and windows gave me gcc.
> I'm not looking to install Cygwin. The second was pretty much like the
> first. I have no idea if they will accept old ftn.

Silverfrost FTN95 has an option to compile using the old
fixed-source form of F77, and has very good debugging
facilities.  Quite happily runs on XP.  It's free for your own non-
commercial use.
0
louisa.hutch (142)
8/17/2011 1:07:53 AM
On 8/16/2011 2:51 PM, dpb wrote:
> On 8/16/2011 4:30 PM, dpb wrote:
> ...
>
>> <http://www.openwatcom.org/index.php/Download>
> ...
>
> Looks like you would want
>
> <ftp://ftp.openwatcom.org/open-watcom-f77-win32-1.9.exe>
>
> BTW, as others have noted; Fortran Standards have maintained backward
> compatibility of most features since F66; even those features which have
> been deleted or deprecated since F90+ have almost universally been
> retained by various compilers.
>
> OW is no different in that regard; it is a F77-compliant compiler and
> has many extensions beyond; what your code might have used is anybody's
> guess and will depend on what machine/OS/compiler(s) was/were used.
>
> You'll simply have to see what is actually in the code; the easiest way
> is to first just throw it at the compiler and see what it complains
> about then see what you have to change either in the source code or in
> options within the compiler if it happens to support either used
> extensions or obsolete features.
>
> There are some things that will possibly be painful but there's no use
> speculating until one has actual data to look at (namely either source
> or better yet, error diagnostics to go with the code)...
>
> I'd give OW a shot, certainly...
>
> --
>
> --
The ftp link produced zero bytes of download. I'm using FF. Maybe I need 
to somehow adjust the link to handle the ftp?

Note that I supplied a post at the top of this thread that gives some 
sample code. One poster concluded it's ftn II.

Yes, it's been my plan to throw some of the code at it the compiler and 
see what happens. However, the code is in a book, and I'm using an OCR 
scanner, Epson Perfection V500, and AABBYY software to scan the pages.

It's obvious in the code post that there are many wrinkles to get by on 
OCR. There are 1400 lines of code in the program. There are three 
programs in that code. The book provides complete input and output, 
fortunately(!!), or I'm not sure I'd be trying to do this. I hope to 
just try the simplest of the three for starters. Possibly I can take a 
few functions and subroutines, and use them with another main program of 
my own making to try matters out.

I just got the V500, and need to get AABBYY tech support to ask a few 
questions. I still maintain hope that I can swing this.
0
wolftracks (253)
8/17/2011 1:29:13 AM
On 8/16/2011 7:50 PM, glen herrmannsfeldt wrote:
....

>
>> FUNCTION FPRIMF(FP1,FM1,F2P1,F2M1)
>> FPRIMF = ((FP1-FM1)/2.-(F2P1-F2M1)/l2.)/4.
>> END
>> FUNCTION XINTPF(FO,F1,F20,F21,S)
>> XINTPF = FO + S*(F1-FO>-(S*<  S-l ??? )*(S-2. )*F20 + S*(S+1???)*(S-l???)*F21
>> )/6,
>
> Again, trailing F on function names.  I believe the rule is that
> INTEGER functions start with X, and REAL without X.
>
> That you will have to fix.  Declare the X functions INTEGER where they
> are called, and INTEGER FUNCTION XINTP(...) here.
....

Are you sure about this?  That isn't what I recall (or _think_ I recall 
from FORTRAN II and doesn't seem to match the IBM manual at 
computerhistory.org...

It says for new statement FUNCTION in part...

"...The function name consists of 1 to 6 alphanumeric characters, the 
first of which is alphabetic; the first character must be I, J, K, L, M, 
or N if and only if the value of the function is to be fixed point, and 
the final character must not be F if the total number of characters is 
4, 5, or 6. ... "

<http://archive.computerhistory.org/resources/text/Fortran/102653989.05.01.acc.pdf>

I didn't have any luck in finding early CDC manual, only beginning w/ 
IV, unfortunately.

That's the point at which my FORTRAN experience began, roughly, I don't 
know the prior rules...

So, I don't know if the above function is or isn't supposed to be an 
integer-valued function or not w/o more study, but just casually looking 
at the source doesn't make me think it is...

--
0
none1568 (7455)
8/17/2011 1:38:26 AM
W. eWatson <wolftracks@invalid.com> wrote:

> Yes, it's been my plan to throw some of the code at it the compiler and
> see what happens. However, the code is in a book, and I'm using an OCR
> scanner, Epson Perfection V500, and AABBYY software to scan the pages.

Ouch. OCR of code is painful. Sometimes it can be so painful that it
ends up being more work to find and fix all the problems relating to the
OCR. It can sometimes work out ok, but sometimes not. Code is very
sensitive to picky things like O vs 0, and 1 vs I. The contextual clues
to disambiguation aren't always the same as those of English text, there
being things like AMAX1.

-- 
Richard Maine
email: last name at domain . net
domain: summer-triangle
0
nospam47 (9747)
8/17/2011 1:40:28 AM
On 8/16/2011 6:40 PM, Richard Maine wrote:
> W. eWatson<wolftracks@invalid.com>  wrote:
>
>> Yes, it's been my plan to throw some of the code at it the compiler and
>> see what happens. However, the code is in a book, and I'm using an OCR
>> scanner, Epson Perfection V500, and AABBYY software to scan the pages.
>
> Ouch. OCR of code is painful. Sometimes it can be so painful that it
> ends up being more work to find and fix all the problems relating to the
> OCR. It can sometimes work out ok, but sometimes not. Code is very
> sensitive to picky things like O vs 0, and 1 vs I. The contextual clues
> to disambiguation aren't always the same as those of English text, there
> being things like AMAX1.
>
Ouch is right. I'll run with it as long as I think it's  useful. The 
fallback position is to recode it in Python or VB.  That has some 
appeal, since 1/3 of the code is replaceable by some C++ code I have. 
However, I really want simplicity here. I don't need the  torture. :-)
0
wolftracks (253)
8/17/2011 2:18:26 AM
On 8/16/2011 8:29 PM, W. eWatson wrote:
> On 8/16/2011 2:51 PM, dpb wrote:
....

>> Looks like you would want
>>
>> <ftp://ftp.openwatcom.org/open-watcom-f77-win32-1.9.exe>
....

> The ftp link produced zero bytes of download. I'm using FF. Maybe I need
> to somehow adjust the link to handle the ftp?

Worked just fine and installed cleanly here from the download (I'd never 
downloaded OW; I have the last published commercial version and afaik 
there's been no reak development on the Fortran side since).

But, I used FF here, too (altho not sure which version, I haven't 
updated much recently but noticed it did some behind-the-scenes update 
just the other day so must not be _too_ out-of-date).

> Note that I supplied a post at the top of this thread that gives some
> sample code. One poster concluded it's ftn II.

Well, I'm not sure about that but I won't argue it isn't although the 
IBM FORTRAN II manual I looked at didn't have the restriction on 
function names Glen thought...by the time I really began writing 
FORTRAN, FORTRAN IV generation was the compiler on the machines I used 
early on (Philco followed by CDC 66/7600s) so my direct knowledge 
earlier than that is skimpy at best.

> Yes, it's been my plan to throw some of the code at it the compiler and
> see what happens. However, the code is in a book, and I'm using an OCR
> scanner, Epson Perfection V500, and AABBYY software to scan the pages.
>
> It's obvious in the code post that there are many wrinkles to get by on
> OCR. There are 1400 lines of code in the program. There are three
> programs in that code. The book provides complete input and output,
> fortunately(!!), or I'm not sure I'd be trying to do this. I hope to
> just try the simplest of the three for starters. Possibly I can take a
> few functions and subroutines, and use them with another main program of
> my own making to try matters out.
....

As Richard says on down that's a big order for OCR...1400 lines sounds 
like a lot but it's "only" something like 20 pages or so; you might find 
it quicker in the end to simply either type the code from the book than 
try to fix the OCR or take the intent and read the code and rewrite it 
in a newer dialect of Fortran; like at least F95 instead of worrying 
about actually implementing this old code.

Just for curiosity's sake, what _does_ it try to do?

--
0
none1568 (7455)
8/17/2011 2:21:32 AM
W. eWatson wrote:
> On 8/16/2011 2:51 PM, dpb wrote:
>> On 8/16/2011 4:30 PM, dpb wrote:
>> ...
>>
>>> <http://www.openwatcom.org/index.php/Download>
>> ...
>>
>> Looks like you would want
>>
>> <ftp://ftp.openwatcom.org/open-watcom-f77-win32-1.9.exe>
>>
>> BTW, as others have noted; Fortran Standards have maintained backward
>> compatibility of most features since F66; even those features which have
>> been deleted or deprecated since F90+ have almost universally been
>> retained by various compilers.
>>
>> OW is no different in that regard; it is a F77-compliant compiler and
>> has many extensions beyond; what your code might have used is anybody's
>> guess and will depend on what machine/OS/compiler(s) was/were used.
>>
>> You'll simply have to see what is actually in the code; the easiest way
>> is to first just throw it at the compiler and see what it complains
>> about then see what you have to change either in the source code or in
>> options within the compiler if it happens to support either used
>> extensions or obsolete features.
>>
>> There are some things that will possibly be painful but there's no use
>> speculating until one has actual data to look at (namely either source
>> or better yet, error diagnostics to go with the code)...
>>
>> I'd give OW a shot, certainly...
>>
>> --
>>
>> --
> The ftp link produced zero bytes of download. I'm using FF. Maybe I need 
> to somehow adjust the link to handle the ftp?
> 
Pasted the link in firefox, downloaded oke.
0
8/17/2011 3:05:19 AM
Gordon Sande <Gordon.Sande@gmail.com> wrote:

(snip)
> In Fortran II the trailing F indicated the use of a nonstandard calling
> sequence. The usual calling sequence was for any number of arguements and
> used a list of addresses. The nonstandard method was to put the single
> argument in the accumulator.

> So one called SINF but the library had a SIN that took its arguement from
> the accumulator.

Bitsavers has the IBM Fortran II General Information manual, which
gives the processor independent explanations.  It looks like the
trailing F is there for statement functions, built-in functions,
and library functions, along with the initial X for INTEGERs.

Fortran I put the first argument in AC, the second in MQ, and later
in memory.  Note call by value.   Also, the F doesn't count toward
the length of the name, seven characters being allowed when variables
can only have six.

IBM Fortran II was implemented on many different machines with
different storage formats (binary and decimal), and different
calling conventions.  

I presume, but so far haven't seen documented, that FUNCTIONs
and SUBROUTINEs were not call by value.

-- glen
0
gah (12850)
8/17/2011 3:31:34 AM
dpb <none@non.net> wrote:

> As Richard says on down that's a big order for OCR...1400 lines sounds
> like a lot but it's "only" something like 20 pages or so; you might find
> it quicker in the end to simply either type the code from the book than
> try to fix the OCR or take the intent and read the code and rewrite it
> in a newer dialect of Fortran; like at least F95 instead of worrying 
> about actually implementing this old code.

I also note that I failed to actually finish my sentence before
stumbling on to the next one. I said that sometimes fixing the OCR
version was "more work", but the "than whatever" part doesn't seem to
have made it from my mind to the keyboard. I had intended to say "more
work than just typing the code" as per dpb's comment above.


-- 
Richard Maine
email: last name at domain . net
domain: summer-triangle
0
nospam47 (9747)
8/17/2011 3:33:02 AM
Louisa <louisa.hutch@gmail.com> wrote:
(snip)

> Silverfrost FTN95 has an option to compile using the old
> fixed-source form of F77, and has very good debugging
> facilities.  Quite happily runs on XP.  It's free for your own non-
> commercial use.

Fixed source form is still standard in Fortran 2008. 
I don't see any suggestions that it be removed in the next revision.

-- glen
0
gah (12850)
8/17/2011 3:33:47 AM
Richard Maine <nospam@see.signature> wrote:

(snip on OCR)

> Ouch. OCR of code is painful. Sometimes it can be so painful that it
> ends up being more work to find and fix all the problems relating to the
> OCR. It can sometimes work out ok, but sometimes not. Code is very
> sensitive to picky things like O vs 0, and 1 vs I. The contextual clues
> to disambiguation aren't always the same as those of English text, there
> being things like AMAX1.

The first, last, and only time I tried to run OCR code, I decided
that it took about as long as it would to just type in.

Also, typing it in is more interesting than debugging OCR code.

-- glen

0
gah (12850)
8/17/2011 3:36:30 AM
glen herrmannsfeldt <gah@ugcs.caltech.edu> wrote:

> Louisa <louisa.hutch@gmail.com> wrote:
> (snip)
> 
> > Silverfrost FTN95 has an option to compile using the old
> > fixed-source form of F77, and has very good debugging
> > facilities.  Quite happily runs on XP.  It's free for your own non-
> > commercial use.
> 
> Fixed source form is still standard in Fortran 2008. 
> I don't see any suggestions that it be removed in the next revision.

Well, it is obsolescent (or anyway it was in f2003; I didn't check to
see whether that might have changed). Obsolescence constitutes a
suggestion that it might be removed in the next version. I would not
place any large bets on that actually happening, and even if it was
removed from the standard, I wouldn't bet on it getting removed from
most compilers, but the suggestion has been made and at one time got
enough support to be incorporated in the standard. (I remember voting in
favor of making fixed source form obsolescent, but obviously I wasn't
the only one).

-- 
Richard Maine
email: last name at domain . net
domain: summer-triangle
0
nospam47 (9747)
8/17/2011 3:41:18 AM
dpb <none@non.net> wrote:

(snip, I wrote)

>> Again, trailing F on function names.  I believe the rule is that
>> INTEGER functions start with X, and REAL without X.

>> That you will have to fix.  Declare the X functions INTEGER where they
>> are called, and INTEGER FUNCTION XINTP(...) here.

> Are you sure about this?  That isn't what I recall (or _think_ I recall 
> from FORTRAN II and doesn't seem to match the IBM manual at 
> computerhistory.org...

The manual I found on bitsavers (after that post) indicates that
one naming convention (I believe with X and F) is used for statement
functions, built-in functions, and library functions, and that
another (not described) is used for FUNCTIONs.

> It says for new statement FUNCTION in part...

> "...The function name consists of 1 to 6 alphanumeric characters, the 
> first of which is alphabetic; the first character must be I, J, K, L, M, 
> or N if and only if the value of the function is to be fixed point, and 
> the final character must not be F if the total number of characters is 
> 4, 5, or 6. ... "

The sample does use SQRTF(), the Fortran II name.

Maybe someone was being consistent and also using X and F for
Fortran FUNCTIONs, but that wouldn't work if INTEGER functions
needed I, J, K, L, M, or N.

-- glen
0
gah (12850)
8/17/2011 3:43:03 AM
Richard Maine <nospam@see.signature> wrote:

(snip, I wrote)
>> Fixed source form is still standard in Fortran 2008. 
>> I don't see any suggestions that it be removed in the next revision.

> Well, it is obsolescent (or anyway it was in f2003; I didn't check to
> see whether that might have changed). Obsolescence constitutes a
> suggestion that it might be removed in the next version. 

But it is still in 2008, isn't it?  I have a 2008 PDF, but on
a different computer.  Until I use an actual Fortran 2003 compiler
I haven't been so interested in reading the 2008 standard.

Besides being on the obsolescence list, has there been discussion
for actually removing it?  It seems to me that some are on the
list to discourage their use in new code, but with little interest
in actually removing them.

> I would not
> place any large bets on that actually happening, and even if it was
> removed from the standard, I wouldn't bet on it getting removed from
> most compilers, but the suggestion has been made and at one time got
> enough support to be incorporated in the standard. (I remember voting in
> favor of making fixed source form obsolescent, but obviously I wasn't
> the only one).

It would be nice to have a conversion program before it was removed.
The conversion program I used before would only convert Fortran 2007.
(The complication being where the blanks are and aren't.)

-- glen
0
gah (12850)
8/17/2011 3:51:26 AM
glen herrmannsfeldt <gah@ugcs.caltech.edu> wrote:

> Richard Maine <nospam@see.signature> wrote:
> 
> (snip, I wrote)
> >> Fixed source form is still standard in Fortran 2008. 
> >> I don't see any suggestions that it be removed in the next revision.
> 
> > Well, it is obsolescent (or anyway it was in f2003; I didn't check to
> > see whether that might have changed). Obsolescence constitutes a
> > suggestion that it might be removed in the next version. 
> 
> But it is still in 2008, isn't it?

Yes. I was responding only to "I don't see any suggestion" by pointing
out that there is a clearly visible suggestion that I'm pretty sure you
must have seen. That's all. I'm not commenting on whether it is likely
to happen - just that obsolescence is such a suggestion. See the
standard itself for what it means by "obsolescent."

-- 
Richard Maine
email: last name at domain . net
domain: summer-triangle
0
nospam47 (9747)
8/17/2011 4:04:07 AM
W. eWatson wrote:
> On 8/16/2011 12:16 PM, glen herrmannsfeldt wrote:
>> W. eWatson<wolftracks@invalid.com>  wrote:
>>
>>> I have a FORTRAN program dating back to 1964, and would like to compile
>>> it. Win7 is my first choice, but XP Pro is OK. Does such an animal
>>> (compiler) exist?
>>
>> If it is real Fortran II, then it might take a little work with
>> modern compilers, but probably not so much.   There might be a few
>> features of Fortran IV that didn't make it into later standards.
>>
>> gfortran is the popular free Fortran compiler these days.
>>
>> -- glen
> Is gftn ready to go in a pure Windows environment w/o Cygwin?

Yes, the MinGW gfortran compiler will run in a pure Windows environment 
without Cygwin.  It's relatively painless if you're fine with one of the 
prepackaged setups (and not that much worse in the unlikely event that 
one of the outstanding bugs is a showstopper for you and you have to 
rebuild from source with the bug patched).
0
enigma (394)
8/17/2011 4:15:35 AM
OK,I stumbled my way into Watfor, and now need to know what to do next. 
As I recall in the install there are example programs. Well, OK, maybe 
what I need more is a document that describes this tool. Back on the 
Watfor site?
0
wolftracks (253)
8/17/2011 4:31:35 AM
W. eWatson wrote:
> OK,I stumbled my way into Watfor, and now need to know what to do next.
> As I recall in the install there are example programs. Well, OK, maybe
> what I need more is a document that describes this tool. Back on the
> Watfor site?

I'd stick with Openwatcom F77 rather than Watfor, although there is no 
harm in trying both.

This download link wored for me a minute ago-
  http://openwatcom.mirrors.pair.com/open-watcom-f77-win32-1.9.exe

I don't envy the task of fixing up source code that went through an OCR 
process. I tried once or twice. Letter "O" versus number "0", lowercase 
"l" versus number "1", etc. Strongly suggest using "Implicit None" and 
explicitly declare all variables at start of each routine.
0
user17588 (148)
8/17/2011 1:36:22 PM
On 8/17/2011 6:36 AM, user1 wrote:
> W. eWatson wrote:
>> OK,I stumbled my way into Watfor, and now need to know what to do next.
>> As I recall in the install there are example programs. Well, OK, maybe
>> what I need more is a document that describes this tool. Back on the
>> Watfor site?
>
> I'd stick with Openwatcom F77 rather than Watfor, although there is no
> harm in trying both.
>
> This download link wored for me a minute ago-
> http://openwatcom.mirrors.pair.com/open-watcom-f77-win32-1.9.exe
>
> I don't envy the task of fixing up source code that went through an OCR
> process. I tried once or twice. Letter "O" versus number "0", lowercase
> "l" versus number "1", etc. Strongly suggest using "Implicit None" and
> explicitly declare all variables at start of each routine.
I barely know what I have. :-) I used the link you mentioned, and don't 
know about plain Watfor. I figured anything I got would be Watfor. I 
found the manual

Yes, the OCR will be a challenge. I began by using it on my HP C6100 AOI 
printer. It consistently failed on equal signs. I figured I'd use a 
higher dpi than 300. Nope. Not allowed. As mentioned above, I happened 
to get my new Epson V500 and found I could get to 9600 dpi. I'm 
presently using it at 600 dpi.

Another tool I'll need is tidy, which squares away errant ftn txt 
formatting. I tried it yesterday, and found the program would not 
install. I sent the presenter keeper, a Canadian, a msg about it. I have 
a feeling he may not be reachable. I think the code dates at 2000.I can 
probably write a Python program that helps.

I found the ftn77 manual, v1.8, so I should be able to get a better 
handle on matters.

ABBYY is the maker of the OCR software, FineReader 6.0, and I will talk 
to their tech people today. Just missed their hours yesterday at 6 pm. 
They may have suggestions.

I was able to contact the author of the book and program. Surprisingly 
he lives 45 miles from me. He was very pleased that anyone would try to 
turn the code into a computer readable form. If this works, we will be 
collaborating on a project using the code. Who would have thought I'd be 
using ftn again!
0
wolftracks (253)
8/17/2011 2:04:38 PM
On 8/17/2011 9:04 AM, W. eWatson wrote:
> On 8/17/2011 6:36 AM, user1 wrote:
....

>> I'd stick with Openwatcom F77 rather than Watfor, although there is no
>> harm in trying both.
>>
>> This download link wored for me a minute ago-
>> http://openwatcom.mirrors.pair.com/open-watcom-f77-win32-1.9.exe
>>
....

> I barely know what I have. :-) I used the link you mentioned, and don't
> know about plain Watfor. I figured anything I got would be Watfor. I
> found the manual

You've lost me entirely...

Did you or did you not download the OpenWatcom F77-Win32 install .exe?

....

> I found the ftn77 manual, v1.8, so I should be able to get a better
> handle on matters.
....

All the OW documentation is installed in the install and includes both 
MS .hlp and (if chosen) a pdf-format file as well.  This includes 
documentation for both the compiler and the IDE and a programmer's 
manual that talks about the compiler usage and how to build for the 
target system as well as the LRF (Language Reference Manual)

Other than (perhaps) going to bitsavers or some other archive site for 
early FORTRAN syntax if something arcane shows up (altho what you posted 
looks pretty plain-vanilla and shouldn't create too much trouble w/ 
F77-compiler) you should have all you need in the package to use OW F77. 
  If you've not used Fortran in quite some time, you'll probably find a 
general text worthwhile for general reference, however.

Hmmmm.....just went and looked at the download install of OW...you 
should have V1.9 documents and I see that the install didn't put icons 
for the help files in the program folder automagically--that should be 
fixed.

The .pdf files are in the \docs subdirectory; you'll want

fuguide -- Fortran Users Guide, the Programmers' Manual
f77lr   -- F77 Language Reference Manual
ftools  -- Linker, Librarian, etc., etc., ...

and the holy grail...

fbooks  -- F77 Bookshelf index to all online help documents is a help 
file in the \binnt subdirectory.  Unfortunately, there doesn't seem to 
be an equivalent in the pdf documents yet.

I'd make a shortcut to the fbooks help file firstest thing I did...

--



for starters
--


0
none1568 (7455)
8/17/2011 3:03:21 PM
W. eWatson wrote:

> I barely know what I have. :-) I used the link you mentioned, and don't
> know about plain Watfor. I figured anything I got would be Watfor. I
> found the manual
>

History mainly. You seem to have OpenWatcom F77. There is a much older 
product called Watfor.


0
user17588 (148)
8/17/2011 6:20:11 PM
user1 <user1@example.net> wrote:
(snip)

> History mainly. You seem to have OpenWatcom F77. There is a much older 
> product called Watfor.

As I understand it, WATFOR ran on the 7090, and then later the IBM 360.
The improved version was named WATFIV, pronounced wat-five, and 
supported many Fortran 77 features in 1973.  Those came from the
University of Waterloo in Canada.     (One interesting feature
of 7090 WATFOR was the use of parity bits to detect use of
variables before they had been given a value.  As this wasn't
available for S/360, they converted to using a specific value
and testing before every use.)

At some point, commercial versions of the compilers for Intel
systems were sold by the company named Watcom.  I don't know the 
exact connection between Watcom and U Waterloo, though.  I believe 
the WATFOR name was revived, though I am not sure exactly how.

Sometime later, among various buying and selling of companies,
the source was released and OpenWatcom started working on and
distributing it.  In a somewhat similar, but unrelated, way,
StarOffice, originally bought by Sun, was released to start OpenOffice.

-- glen


0
gah (12850)
8/17/2011 7:51:18 PM
On 8/17/2011 8:03 AM, dpb wrote:
> On 8/17/2011 9:04 AM, W. eWatson wrote:
>> On 8/17/2011 6:36 AM, user1 wrote:
> ...
>
>>> I'd stick with Openwatcom F77 rather than Watfor, although there is no
>>> harm in trying both.
>>>
>>> This download link wored for me a minute ago-
>>> http://openwatcom.mirrors.pair.com/open-watcom-f77-win32-1.9.exe
>>>
> ...
>
>> I barely know what I have. :-) I used the link you mentioned, and don't
>> know about plain Watfor. I figured anything I got would be Watfor. I
>> found the manual
>
> You've lost me entirely...
>
> Did you or did you not download the OpenWatcom F77-Win32 install .exe?
>
> ...
Yes. ftn77.
>
>> I found the ftn77 manual, v1.8, so I should be able to get a better
>> handle on matters.
> ...
>
> All the OW documentation is installed in the install and includes both
> MS .hlp and (if chosen) a pdf-format file as well. This includes
> documentation for both the compiler and the IDE and a programmer's
> manual that talks about the compiler usage and how to build for the
> target system as well as the LRF (Language Reference Manual)
>
> Other than (perhaps) going to bitsavers or some other archive site for
> early FORTRAN syntax if something arcane shows up (altho what you posted
> looks pretty plain-vanilla and shouldn't create too much trouble w/
> F77-compiler) you should have all you need in the package to use OW F77.
> If you've not used Fortran in quite some time, you'll probably find a
> general text worthwhile for general reference, however.
>
> Hmmmm.....just went and looked at the download install of OW...you
> should have V1.9 documents and I see that the install didn't put icons
> for the help files in the program folder automagically--that should be
> fixed.
>
> The .pdf files are in the \docs subdirectory; you'll want
>
> fuguide -- Fortran Users Guide, the Programmers' Manual
> f77lr -- F77 Language Reference Manual
> ftools -- Linker, Librarian, etc., etc., ...
Win7 search finds no ftools, ftools, or fuguide anywhere on my PC. Same 
with fbooks.
>
> and the holy grail...
>
> fbooks -- F77 Bookshelf index to all online help documents is a help
> file in the \binnt subdirectory. Unfortunately, there doesn't seem to be
> an equivalent in the pdf documents yet.
>
> I'd make a shortcut to the fbooks help file firstest thing I did...
>
> --
> for starters
> --
>
>
So what's the story with the Pdf I have? Open Fortran 77, about 55 
pages. The StartUp menu for Win7 shows a lot of sub items, DDE Spy, Text 
Editor, Zoom, README, etc. I think this will take awhile to settle in.

I'm a bit baffled by this NG. Each time I go to another one, FF seems to 
unsubscribe me, so when subscribe, there are no "have read" indicators. 
Well, I'll hopefully get that squared away with a FF NG.
0
wolftracks (253)
8/17/2011 8:36:30 PM
W. eWatson wrote:

> So what's the story with the Pdf I have? Open Fortran 77, about 55
> pages. The StartUp menu for Win7 shows a lot of sub items, DDE Spy, Text
> Editor, Zoom, README, etc. I think this will take awhile to settle in.
>

It sounds like you are struggling with really basic stuff. Hopefully you 
allowed the installer to set the needed environment variables. You may 
find it easier to compile short programs from a command prompt (cmd.exe) 
as follows, where I am compiling and running a program named test1.for ...


C:\temp>dir
  Volume in drive C is Local Disk1
  Volume Serial Number is D436-BEF6

  Directory of C:\temp

08/17/2011  05:43 PM    <DIR>          .
08/17/2011  05:43 PM    <DIR>          ..
08/17/2011  05:42 PM                44 test1.for
                1 File(s)             44 bytes
                2 Dir(s)  399,520,464,896 bytes free

C:\temp>type test1.for
        print *,'Hello World'
        end


C:\temp>wfl386 test1.for
Open Watcom F77/32 Compile and Link Utility Version 1.9
Portions Copyright (c) 1990-2002 Sybase, Inc. All Rights Reserved.
Source code is available under the Sybase Open Watcom Public License.
See http://www.openwatcom.org/ for details.
         wfc386 test1.for
Open Watcom FORTRAN 77/32 Optimizing Compiler Version 1.9
Portions Copyright (c) 1984-2002 Sybase, Inc. All Rights Reserved.
Source code is available under the Sybase Open Watcom Public License.
See http://www.openwatcom.org/ for details.
test1.for: 2 statements, 20 bytes, 1 extensions, 0 warnings, 0 errors
         wlink @__wfl__.lnk
Open Watcom Linker Version 1.9
Portions Copyright (c) 1985-2002 Sybase, Inc. All Rights Reserved.
Source code is available under the Sybase Open Watcom Public License.
See http://www.openwatcom.org/ for details.
loading object files
searching libraries
creating a Windows NT character-mode executable

C:\temp>test1
Hello World

C:\temp>


0
user17588 (148)
8/17/2011 9:49:25 PM
On 8/17/2011 3:36 PM, W. eWatson wrote:
> On 8/17/2011 8:03 AM, dpb wrote:
....

>> Did you or did you not download the OpenWatcom F77-Win32 install .exe?
>>
>> ...
> Yes. ftn77.
>>
>>> I found the ftn77 manual, v1.8, so I should be able to get a better
>>> handle on matters.
>> ...

Again, V1.8 is _NOT_ the lastest version of OW F77 so I don't know what 
you actually have downloaded.

What, specifically, was the file you downloaded and where did it come from?

....

>> The .pdf files are in the \docs subdirectory; you'll want
>>
>> fuguide -- Fortran Users Guide, the Programmers' Manual
>> f77lr -- F77 Language Reference Manual
>> ftools -- Linker, Librarian, etc., etc., ...

> Win7 search finds no ftools, ftools, or fuguide anywhere on my PC. Same
> with fbooks.

Then you haven't installed OW F77 V1.9 fully w/ the included 
documentation or that isn't what you actually downloaded.

Which directory did you place it in?  Go to that directory and under it 
should be a directory structure of

...\binnt
...\binw
...\docs
....

The .pdf files are in the ..\docs subdirectory if you allowed install to 
install them; the .hlp files are in the ..\binnt subdirectory.

Again, if you didn't do a full install or got a partial distribution 
from somewhere other than openwatcom.org or a mirror there's no telling 
what you have actually installed.

The thing is again that everything current _SHOULD_ say V1.9 associated 
with it, _NOT_ v1.8.
....

> So what's the story with the Pdf I have? Open Fortran 77, about 55
> pages. The StartUp menu for Win7 shows a lot of sub items, DDE Spy, Text
> Editor, Zoom, README, etc. I think this will take awhile to settle in.

That would appear perhaps to be the "Getting Started" document; that's 
56 pages here but w/o a specific document name/title how am I supposed 
to know???

The Start menu should have the utilities as well as the IDE; however for 
something like what you're attempting as user1 says it may be as simple 
to just use the text editor and work on compiling your code from the 
command line.

At a command prompt (I presume Win7 still lets you have a command 
window; I've yet to see a Win7 box?) what does just typing

wfc386

give?  Here, I see...

C:\> wfc386
Open Watcom FORTRAN 77/32 Optimizing Compiler Version 1.9
Portions Copyright (c) 1984-2002 Sybase, Inc. All Rights Reserved.
Source code is available under the Sybase Open Watcom Public License.
See http://www.openwatcom.org/ for details.

  Usage: wfc386 {- or /<option>} <file-spec> {- or /<option>}
.....

> I'm a bit baffled by this NG. Each time I go to another one, FF seems to
> unsubscribe me, so when subscribe, there are no "have read" indicators.
> Well, I'll hopefully get that squared away with a FF NG.

You've lost me entirely here..."go to another one" what?

Your headers make it look like you're using Thunderbird at the 
eternal-september newsserver location; don't see why you should be 
having any particular trouble at the clf group...

Connect to the newsserver of choice and subscribe to comp.lang.fortran 
and read/post...not much else I know to say about it.

Again, I'd start at the beginning--it appears you do have some version 
of OW F77 but at least one release out of date.  It doesn't make much 
difference on the functionality as there has been little changed but it 
would surely help to be consistent if trying to figure out what you 
have/don't have.

--
0
none1568 (7455)
8/17/2011 10:38:50 PM
On Aug 17, 4:05=A0am, "W. eWatson" <wolftra...@invalid.com> wrote:
> I have a FORTRAN program dating back to 1964, and would like to compile
> it. Win7 is my first choice, but XP Pro is OK. Does such an animal
> (compiler) exist?

My experience is that Microsoft F77 (V 3.1 of 1983) and CVF/DVF
Fortran 6.6 , running on my M.S. V7, will compile anything FII, FIII
(alias "4/1/61"), mainframe FIV essentially without modifications,
(except for storage of alphanumeric strings as floating point
constants, when a little re-definition to strings constants is
preferred). 1964 is spot on the late I was the local Fortran expert at
IBM UK (London).

If you e-mail me the source I can give it a try and report on what
needs doing, and what i recommend doing for more compatibility down
the line.
I can even automatically re-structure such a program, including re-
ordering the label and format number sequences to make tracing easier.
0
tbwright (1098)
8/18/2011 12:23:31 AM
On Aug 18, 10:23=A0am, Terence <tbwri...@cantv.net> wrote:
> On Aug 17, 4:05=A0am, "W. eWatson" <wolftra...@invalid.com> wrote:
>
> > I have a FORTRAN program dating back to 1964, and would like to compile
> > it. Win7 is my first choice, but XP Pro is OK. Does such an animal
> > (compiler) exist?
>
> My experience is that Microsoft F77 (V 3.1 of 1983) and CVF/DVF
> Fortran 6.6 , running on my M.S. V7, will compile anything FII, FIII
> (alias "4/1/61"), mainframe FIV essentially without modifications,
> (except for storage of alphanumeric strings as floating point
> constants, when a little re-definition to strings constants is
> preferred). 1964 is spot on the late I was the local Fortran expert at
> IBM UK (London).
>
> If you e-mail me the source I can give it a try and report on what
> needs doing, and what i recommend doing for more compatibility down
> the line.
> I can even automatically re-structure such a program, including re-
> ordering the label and format number sequences to make tracing easier.


OK.
I had a look at some of the posted sample code.
I wouldn't worry to much about the OCR errors that I see.
Most are mis-reads such as replacements of a comma by << or asterisk
or #  and the like, where the actual character is non-valid in the
context.

First I correct the OCR-captured version globally with an editor that
replaces  globally with optional stopping for confirmation.
Then I compile (with a variable listing and line mentions) to
a) spot constant storage problems,
b) locate errors and variables with similar names that might be
intended to be the same.
c) use of obsolete radix codes and unique-to-a-vendor Fortran syntax.

This clears up most OCR or compatibility issues.

The next compile might even not even show errors. If test data with a
listing if the expected results is available, getting a correct
version is fairly straight-forward.
0
tbwright (1098)
8/18/2011 1:17:38 AM
On 8/17/2011 3:38 PM, dpb wrote:
> On 8/17/2011 3:36 PM, W. eWatson wrote:
>> On 8/17/2011 8:03 AM, dpb wrote:
> ...
>
>>> Did you or did you not download the OpenWatcom F77-Win32 install .exe?
>>>
>>> ...
>> Yes. ftn77.
>>>
>>>> I found the ftn77 manual, v1.8, so I should be able to get a better
>>>> handle on matters.
>>> ...
>
> Again, V1.8 is _NOT_ the lastest version of OW F77 so I don't know what
> you actually have downloaded.
>
> What, specifically, was the file you downloaded and where did it come from?
It came directly from the link you provided. Well perhaps not. I could 
not get the ftp to work, but noticed the program exe in your link. I 
went to the Watfor site and found it under downloads, and downloaded it, 
  then installed it.

Help About shows. 1.9. Does the manual version keep step with the 
program? Ah, you say yes below. I'll look on the watfor site  again for 
the 1.9 doc.
>
> ...
>
>>> The .pdf files are in the \docs subdirectory; you'll want
>>>
>>> fuguide -- Fortran Users Guide, the Programmers' Manual
>>> f77lr -- F77 Language Reference Manual
>>> ftools -- Linker, Librarian, etc., etc., ...
>
>> Win7 search finds no ftools, ftools, or fuguide anywhere on my PC. Same
>> with fbooks.
>
> Then you haven't installed OW F77 V1.9 fully w/ the included
> documentation or that isn't what you actually downloaded.
>
> Which directory did you place it in? Go to that directory and under it
> should be a directory structure of
>
> ..\binnt
> ..\binw
> ..\docs
> ...
These folder seem more like something in Linux.
I see C:\WATCOM, which appears to be some WatFor diety for ftn. Ah, 
there's binnt and it's brothers, edat, binw. There's a heap of folders 
down there.
>
> The .pdf files are in the ..\docs subdirectory if you allowed install to
> install them; the .hlp files are in the ..\binnt subdirectory.
>
> Again, if you didn't do a full install or got a partial distribution
> from somewhere other than openwatcom.org or a mirror there's no telling
> what you have actually installed.
>
> The thing is again that everything current _SHOULD_ say V1.9 associated
> with it, _NOT_ v1.8.
> ...
>
>> So what's the story with the Pdf I have?
==========
Open Fortran 77
=========== See below.
, about 55
>> pages. The StartUp menu for Win7 shows a lot of sub items, DDE Spy, Text
>> Editor, Zoom, README, etc. I think this will take awhile to settle in.
>
> That would appear perhaps to be the"Getting Started" document; that's
> 56 pages here but w/o a specific document name/title how am I supposed
> to know???
I think I provided the title, perhaps in another post. In fact, the 
manual uses Getting Started in the title, along with Open Watcom F77.
>
> The Start menu should have the utilities as well as the IDE; however for
> something like what you're attempting as user1 says it may be as simple
> to just use the text editor and work on compiling your code from the
> command line.
>
> At a command prompt (I presume Win7 still lets you have a command
> window; I've yet to see a Win7 box?) what does just typing
>
> wfc386
>
> give? Here, I see...
>
> C:\> wfc386
> Open Watcom FORTRAN 77/32 Optimizing Compiler Version 1.9
> Portions Copyright (c) 1984-2002 Sybase, Inc. All Rights Reserved.
> Source code is available under the Sybase Open Watcom Public License.
> See http://www.openwatcom.org/ for details.
>
> Usage: wfc386 {- or /<option>} <file-spec> {- or /<option>}
> ....
>
Let's drop the following topic. I'll work it out.
>> I'm a bit baffled by this NG. Each time I go to another one, FF seems to
>> unsubscribe me, so when subscribe, there are no "have read" indicators.
>> Well, I'll hopefully get that squared away with a FF NG.
>
> You've lost me entirely here..."go to another one" what?
Going to another FireFox (FF) (Newsgroup)NG.
>
> Your headers make it look like you're using Thunderbird at the
> eternal-september newsserver location; don't see why you should be
> having any particular trouble at the clf group...
>
> Connect to the newsserver of choice and subscribe to comp.lang.fortran
> and read/post...not much else I know to say about it.
>
> Again, I'd start at the beginning--it appears you do have some version
> of OW F77 but at least one release out of date. It doesn't make much
> difference on the functionality as there has been little changed but it
> would surely help to be consistent if trying to figure out what you
> have/don't have.
>
> --

I think we are getting closer to wrapping up.

0
wolftracks (253)
8/18/2011 1:36:46 AM
See post above to dbp. I think I'm on track.


0
wolftracks (253)
8/18/2011 1:39:22 AM
On 8/17/2011 6:39 PM, W. eWatson wrote:
> See post above to dbp. I think I'm on track.
>
>
dpb.
0
wolftracks (253)
8/18/2011 1:41:00 AM
W. eWatson <wolftracks@invalid.com> wrote:

> On 8/17/2011 3:38 PM, dpb wrote:
> > On 8/17/2011 3:36 PM, W. eWatson wrote:
> >> I'm a bit baffled by this NG. Each time I go to another one, FF seems to
> >> unsubscribe me, so when subscribe, there are no "have read" indicators.
> >> Well, I'll hopefully get that squared away with a FF NG.
> >
> > You've lost me entirely here..."go to another one" what?
> Going to another FireFox (FF) (Newsgroup)NG.

Firefox doesn't have newsgroups, which is probably the main source of
the confusion. Well, I suppose there probably are newsgroups about
Firefox (didn't check; it just seems likely) but that doesn't fit the
context. You can use Firefox to access newsgroups, but that doesn't make
them Firefox newsgroups (and as dpb noted, the headers look like you are
using Thunderbird).

I could probably be just as confusing and talk about this as a Mac
newsgroup because I'm reading it on my Mac.

-- 
Richard Maine
email: last name at domain . net
domain: summer-triangle
0
nospam47 (9747)
8/18/2011 1:49:04 AM
On Aug 18, 11:17=A0am, Terence <tbwri...@cantv.net> wrote:
> On Aug 18, 10:23=A0am, Terence <tbwri...@cantv.net> wrote:
>
>
>
>
>
> > On Aug 17, 4:05=A0am, "W. eWatson" <wolftra...@invalid.com> wrote:
>
> > > I have a FORTRAN program dating back to 1964, and would like to compi=
le
> > > it. Win7 is my first choice, but XP Pro is OK. Does such an animal
> > > (compiler) exist?
>
> > My experience is that Microsoft F77 (V 3.1 of 1983) and CVF/DVF
> > Fortran 6.6 , running on my M.S. V7, will compile anything FII, FIII
> > (alias "4/1/61"), mainframe FIV essentially without modifications,
> > (except for storage of alphanumeric strings as floating point
> > constants, when a little re-definition to strings constants is
> > preferred). 1964 is spot on the late I was the local Fortran expert at
> > IBM UK (London).
>
> > If you e-mail me the source I can give it a try and report on what
> > needs doing, and what i recommend doing for more compatibility down
> > the line.
> > I can even automatically re-structure such a program, including re-
> > ordering the label and format number sequences to make tracing easier.
>
> OK.
> I had a look at some of the posted sample code.
> I wouldn't worry to much about the OCR errors that I see.
> Most are mis-reads such as replacements of a comma by << or asterisk
> or # =A0and the like, where the actual character is non-valid in the
> context.
>
> First I correct the OCR-captured version globally with an editor that
> replaces =A0globally with optional stopping for confirmation.
> Then I compile (with a variable listing and line mentions) to
> a) spot constant storage problems,
> b) locate errors and variables with similar names that might be
> intended to be the same.
> c) use of obsolete radix codes and unique-to-a-vendor Fortran syntax.
>
> This clears up most OCR or compatibility issues.
>
> The next compile might even not even show errors. If test data with a
> listing if the expected results is available, getting a correct
> version is fairly straight-forward.

I worked on this code sample. It seemed straight-forward.
Locating and correcting the : and ; and i and ! and t replacements of
1 (one); the <> and {} replacements of (); and the # and +
replacements of *; and + for , between indices and so on, was fairly
obvious and quick.

It appears to be FIV for a CDC machine.
It uses function statements

My first suggestion is to use the OCR reading of the source photo, at
least once more, and save both versions of the scanning.
Better still, publish where this was source is to be located, and a
viewable text version would help suggest how to correct some obscure
errors.

Any differences (I have one of those compare programs that let you
select lines or edit directly, on two versions) will spot where a read
error has occurred and suggest others.

My editor easily put back the leading 6 blanks per line and located
leading numerics where the label or continuation digit was displaced.

I repeat my offer to "have a go" if sent the scanned version(s) of the
source.
0
tbwright (1098)
8/18/2011 1:50:31 AM
>
> I could probably be just as confusing and talk about this as a Mac
> newsgroup because I'm reading it on my Mac.
>

Yes, my mistake. TB.
0
wolftracks (253)
8/18/2011 1:53:42 AM
TB NG not the FF NG. There are NGs for each.

0
wolftracks (253)
8/18/2011 1:55:06 AM
On 8/17/2011 8:36 PM, W. eWatson wrote:
> On 8/17/2011 3:38 PM, dpb wrote:
....

>> What, specifically, was the file you downloaded and where did it come
>> from?
> It came directly from the link you provided. Well perhaps not. I could
> not get the ftp to work, but noticed the program exe in your link. I
> went to the Watfor site and found it under downloads, and downloaded it,
> then installed it.

OK, to quit confusing everybody, forget "Watfor"--it's OpenWatcom.Org 
and the compiler is the OpenWatcom F77 compiler.

Watfor was a (long ago antecedent) but is long gone...OpenWatcom is the 
open source organization that was the recipient of the Watcom commercial 
compilers.

>
> Help About shows. 1.9. Does the manual version keep step with the
> program? Ah, you say yes below. I'll look on the watfor site again for
> the 1.9 doc.

That would be "OpenWatcom"...

The documentation is included in the installable .exe and in the version 
I downloaded every document I've opened has also had reference to v1.9. 
  I don't know where the 1.8 stuff must be coming from...

OK, I just tried it; openwatcom.org will accept anonymous ftp.
If you couldn't get the ftp interface through your browser to work, 
there's always the "real" ftp...

C:\ ftp
ftp> open ftp.openwatcom.org
Connected to www.openwatcom.org.
220 ProFTPD 1.3.0 Server (ftp.openwatcom.org) [216.55.181.205]
User (www.openwatcom.org:(none)): anonymous
331 Anonymous login ok, send your complete email address as your password.
Password:
230 Anonymous access granted, restrictions apply.
ftp> ls
200 PORT command successful
150 Opening ASCII mode data connection for file list
archive
devel
manuals
source
TIMESTAMP
license.txt
....[elided parts for brevity]...
open-watcom-c-win32-1.9.exe.md5
open-watcom-f77-dos-1.9.exe
open-watcom-f77-dos-1.9.exe.md5
open-watcom-f77-linux-1.9
open-watcom-f77-linux-1.9.md5
open-watcom-f77-os2-1.9.exe
open-watcom-f77-os2-1.9.exe.md5
open-watcom-f77-win32-1.9.exe
open-watcom-f77-win32-1.9.exe.md5
226-Transfer complete.
226 Quotas off
ftp: binary
ftp> get open-watcom-f77-win32-1.9.exe
....[when done, disconnect]...
ftp> disconnect
ftp> open ftp.openwatcom.org
Connected to www.openwatcom.org.
220 ProFTPD 1.3.0 Server (ftp.openwatcom.org) [216.55.181.205]
User (www.openwatcom.org:(none)): anonymous
331 Anonymous login ok, send your complete email address as your password.
Password:
230 Anonymous access granted, restrictions apply.
ftp> ls
200 PORT command successful
150 Opening ASCII mode data connection for file list
archive
devel
manuals
source
TIMESTAMP
license.txt
open-watcom-c-dos-1.9.exe
open-watcom-c-dos-1.9.exe.md5
open-watcom-c-linux-1.9
open-watcom-c-linux-1.9.md5
readme.txt
wiki
open-watcom-c-os2-1.9.exe
open-watcom-c-os2-1.9.exe.md5
open-watcom-c-win32-1.9.exe
open-watcom-c-win32-1.9.exe.md5
open-watcom-f77-dos-1.9.exe
open-watcom-f77-dos-1.9.exe.md5
open-watcom-f77-linux-1.9
open-watcom-f77-linux-1.9.md5
open-watcom-f77-os2-1.9.exe
open-watcom-f77-os2-1.9.exe.md5
open-watcom-f77-win32-1.9.exe
open-watcom-f77-win32-1.9.exe.md5
226-Transfer complete.
226 Quotas off
ftp> binary
ftp> get open-watcom-f77-win32-1.9.exe

....[ when done, disconnect and quit]...

ftp> disconnect
Connection closed by remote host.
ftp> quit
C:\



>> ...
>>
>>>> The .pdf files are in the \docs subdirectory; you'll want
>>>>
>>>> fuguide -- Fortran Users Guide, the Programmers' Manual
>>>> f77lr -- F77 Language Reference Manual
>>>> ftools -- Linker, Librarian, etc., etc., ...
>>
>>> Win7 search finds no ftools, ftools, or fuguide anywhere on my PC. Same
>>> with fbooks.

Did you search for the .pdf extension or .hlp?  If you downloaded the 
full package and didn't tell install not to, all the doc's should have 
been installed...

....

>> Which directory did you place it in? Go to that directory and under it
>> should be a directory structure of
>>
>> ..\binnt
>> ..\binw
>> ..\docs
>> ...
> These folder seem more like something in Linux.

???

> I see C:\WATCOM, which appears to be some WatFor diety for ftn. Ah,
> there's binnt and it's brothers, edat, binw. There's a heap of folders
> down there.

Well, yes...  :)  Again, forget you ever heard of "WatFor" for this purpose.

My only real puzzle appears yet to be why "1.8" on some of the doc's. 
Probably really doesn't matter but it makes me wonder about a mismatched 
download or somesuch...

OK, I was curious...I opened every .hlp file in ..\winnt and every .pdf 
file in ..\docs subdirectories.  Every one that had a version reference 
was V1.9; quite a number of the tools are generic and have no specific 
version with them. I saw no reference to V1.8, however.

Anyway, looks like you do have a workable compiler even if not the 
latest; good luck w/ your project...

--
0
none1568 (7455)
8/18/2011 2:35:45 AM
W. eWatson <wolftracks@invalid.com> wrote:

> TB NG not the FF NG. There are NGs for each.

Well, no. They are the same newsgroups, which really have little to do
with either TB or FF. They are just newsgroups, or perhaps usenet
newsgroups. They aren't separated out by what particular application you
are using to read them. I'm using neither TB nor FF, for example.

-- 
Richard Maine
email: last name at domain . net
domain: summer-triangle
0
nospam47 (9747)
8/18/2011 2:37:04 AM
On 8/17/2011 7:35 PM, dpb wrote:


....
>
> OK, to quit confusing everybody, forget "Watfor"--it's OpenWatcom.Org
> and the compiler is the OpenWatcom F77 compiler.
>
> Watfor was a (long ago antecedent) but is long gone...OpenWatcom is the
> open source organization that was the recipient of the Watcom commercial
> compilers.
....
> The documentation is included in the installable .exe and in the version
> I downloaded every document I've opened has also had reference to v1.9.
> I don't know where the 1.8 stuff must be coming from...
>
> OK, I just tried it; openwatcom.org will accept anonymous ftp.
> If you couldn't get the ftp interface through your browser to work,
> there's always the "real" ftp...
>
> C:\ ftp
> ftp> open ftp.openwatcom.org
....
>
>>> ...
>>>
>>>>> The .pdf files are in the \docs subdirectory; you'll want
>>>>>
>>>>> fuguide -- Fortran Users Guide, the Programmers' Manual
>>>>> f77lr -- F77 Language Reference Manual
>>>>> ftools -- Linker, Librarian, etc., etc., ...
>>>
>>>> Win7 search finds no ftools, ftools, or fuguide anywhere on my PC. Same
>>>> with fbooks.
>
> Did you search for the .pdf extension or .hlp? If you downloaded the
> full package and didn't tell install not to, all the doc's should have
> been installed...
Nope. Just the name. I just searched for fuguide.pdf. Not found; 
however, sometimes the Win7 search is fussy. The file might not have 
been indexed. Where specifically would I find it? The entire path.
>
> ...
>
....
>
>> I see C:\WATCOM, which appears to be some WatFor diety for ftn. Ah,
>> there's binnt and it's brothers, edat, binw. There's a heap of folders
>> down there.
>
> Well, yes... :) Again, forget you ever heard of "WatFor" for this purpose.
>
> My only real puzzle appears yet to be why "1.8" on some of the doc's.
> Probably really doesn't matter but it makes me wonder about a mismatched
> download or somesuch...
Beats me. Well,not quite thanks to FF history search. 
<http://www.openwatcom.org/index.php/Manuals>  Look at the header Open 
Watcom F77--below it. (Hey, how about that,Open Watcom.) If you drill 
down, you'll find it references a "current" folder.
Did I miss something here for 1.9?

>
> OK, I was curious...I opened every .hlp file in ..\winnt and every .pdf
> file in ..\docs subdirectories. Every one that had a version reference
> was V1.9; quite a number of the tools are generic and have no specific
> version with them. I saw no reference to V1.8, however.
Winnt? docs? What is the complete path to them?
>
> Anyway, looks like you do have a workable compiler even if not the
> latest; good luck w/ your project...
>
> --
v1.9 is the latest, right? That's what the Help->About said.
NOTE: At no time did, I set up matters according the the README file, 
which shows paths, etc. for various OSes. Was it necessary?

Well,as Tom Watson, IBM prez of long ago) said, "Double your error rate, 
and you will succeed." ;-)
0
wolftracks (253)
8/18/2011 4:29:03 AM
On 8/17/2011 6:50 PM, Terence wrote:
> On Aug 18, 11:17 am, Terence<tbwri...@cantv.net>  wrote:
>> On Aug 18, 10:23 am, Terence<tbwri...@cantv.net>  wrote:
>>
>>
>>
>>
>>
>>> On Aug 17, 4:05 am, "W. eWatson"<wolftra...@invalid.com>  wrote:
>>
>>>> I have a FORTRAN program dating back to 1964, and would like to compile
>>>> it. Win7 is my first choice, but XP Pro is OK. Does such an animal
>>>> (compiler) exist?
>>
>>> My experience is that Microsoft F77 (V 3.1 of 1983) and CVF/DVF
>>> Fortran 6.6 , running on my M.S. V7, will compile anything FII, FIII
>>> (alias "4/1/61"), mainframe FIV essentially without modifications,
>>> (except for storage of alphanumeric strings as floating point
>>> constants, when a little re-definition to strings constants is
>>> preferred). 1964 is spot on the late I was the local Fortran expert at
>>> IBM UK (London).
>>
>>> If you e-mail me the source I can give it a try and report on what
>>> needs doing, and what i recommend doing for more compatibility down
>>> the line.
>>> I can even automatically re-structure such a program, including re-
>>> ordering the label and format number sequences to make tracing easier.
>>
>> OK.
>> I had a look at some of the posted sample code.
>> I wouldn't worry to much about the OCR errors that I see.
>> Most are mis-reads such as replacements of a comma by<<  or asterisk
>> or #  and the like, where the actual character is non-valid in the
>> context.
>>
>> First I correct the OCR-captured version globally with an editor that
>> replaces  globally with optional stopping for confirmation.
>> Then I compile (with a variable listing and line mentions) to
>> a) spot constant storage problems,
>> b) locate errors and variables with similar names that might be
>> intended to be the same.
>> c) use of obsolete radix codes and unique-to-a-vendor Fortran syntax.
>>
>> This clears up most OCR or compatibility issues.
>>
>> The next compile might even not even show errors. If test data with a
>> listing if the expected results is available, getting a correct
>> version is fairly straight-forward.
>
> I worked on this code sample. It seemed straight-forward.
> Locating and correcting the : and ; and i and ! and t replacements of
> 1 (one); the<>  and {} replacements of (); and the # and +
> replacements of *; and + for , between indices and so on, was fairly
> obvious and quick.
>
> It appears to be FIV for a CDC machine.
> It uses function statements
>
> My first suggestion is to use the OCR reading of the source photo, at
> least once more, and save both versions of the scanning.
> Better still, publish where this was source is to be located, and a
> viewable text version would help suggest how to correct some obscure
> errors.
>
> Any differences (I have one of those compare programs that let you
> select lines or edit directly, on two versions) will spot where a read
> error has occurred and suggest others.
>
> My editor easily put back the leading 6 blanks per line and located
> leading numerics where the label or continuation digit was displaced.
>
> I repeat my offer to "have a go" if sent the scanned version(s) of the
> source.
Well, this is interesting. By source, do you mean the result of an OCR 
operation? I can mail you that and perhaps some pix of the pages. It 
sounds like you've got some good tools for this sort of thing.

It  appears you are in Britain, correct? I'd almost suggest borrowing 
the book from a college library, but I suspect it's only found in the 
USA. I borrowed mine here, but found a used one on the internet, so 
bought it.

One thing that bothers me a bit is the presence of some odd possibly 
invisible characters that enter into the file. When I tried to edit the 
rtf file where a second line was attached to the end of the previous 
line, I couldn't split them onto two lines.

I talked to ABBYY's tech support today about their FineRead s/w, which 
is what I got with the Epson photo scanner. It's pretty much what I've 
seen in maybe 60 minutes of use. It has the ability to save the 
resulting read text into Word, and some other formats. So far I've just 
used the rtf format.

I was baffled by what I was getting by laying the book flat on the 
scanner. I could easily select the left and right page, but the right 
page produced a big left margin. Sort of like it was both pages, but 
with nothing on the left page. Well, that's the way it works. I'll need 
to make a copy of both pages and cut them in half. That's OK .At least, 
I know what's going on. I have version 6.0 of FineRead, but am 
interested in 9.0, their pro version. Don't know the cost yet, or what 
extended capabilities it has.

Do you have a capability to read rtf?  Maybe Notebook would be sufficient.

I'll continue this tomorrow. It's getting late.

0
wolftracks (253)
8/18/2011 5:28:07 AM
glen herrmannsfeldt <gah@ugcs.caltech.edu> wrote:

> Second, evaluating expressions differently at compile time
> than at run time can be confusing.

....but is sometimes largely unavoidable for cross compilers. GFortran
does support cross compilation. That's likely to be a significant
factor.

-- 
Richard Maine
email: last name at domain . net
domain: summer-triangle
0
nospam47 (9747)
8/18/2011 12:51:22 PM
JB <foo@bar.invalid> wrote:

> IIRC at some point there was discussion about passing scalar
> intent(in) arguments by value for procedures with an explicit
> interface; I don't think this has been done so far.

I won't say it is impossible, but that would have complications that
might not be obvious to some at first glance. Mostly the complication
has to do with having the implementation of a procedure depend on
whether it has an explicit interface. That's a complication because
having an explicit interface is a concept defined only in the scope
where a procedure is referenced - not where the procedure is defined.
There are some properties of a procedure that mean it will require an
explicit interface in order to be referenced, but intent(in) isn't one
of those.

Thus, a procedure with an intent(in) dummy might be referenced with an
explicit interface in some places, but without an explicit interface in
others. Having two versions of the procedure might work in some cases,
but would be tricky if there were saved local variables. So-called
link-time optimization has been mentioned here recently. I suppose that
might be used to procrastinate the decision until "link time", which
could allow the compiler to know things about all the referencing
scopes.

-- 
Richard Maine
email: last name at domain . net
domain: summer-triangle
0
nospam47 (9747)
8/18/2011 1:01:58 PM
On 8/17/2011 11:29 PM, W. eWatson wrote:
....

> Beats me. Well,not quite thanks to FF history search.
> <http://www.openwatcom.org/index.php/Manuals> Look at the header Open
> Watcom F77--below it. (Hey, how about that,Open Watcom.) If you drill
> down, you'll find it references a "current" folder.
> Did I miss something here for 1.9?

_IF_ you will (or did) download

open-watcom-f77-win32-1.9.exe

as I demo'ed for the anon ftp session, all the documentation will be 
included.

Apparently that subdirectory of the OW site wasn't updated at the same 
time as the files in the distribution were refreshed--altho it probably 
doesn't matter for content, perhaps that's why you aren't finding stuff.

....

> Winnt? docs? What is the complete path to them?
>>
>> Anyway, looks like you do have a workable compiler even if not the
>> latest; good luck w/ your project...
>>
>> --
> v1.9 is the latest, right? That's what the Help->About said.
> NOTE: At no time did, I set up matters according the the README file,
> which shows paths, etc. for various OSes. Was it necessary?

If you run the above .exe, it has an installer that lets you choose the 
root directory (I think the default is "\WATCOM"; I chose "OWATCOM" 
because I have the former commercial installation in the "WATCOM" 
directory which didn't want to overwrite).  You have to select the 
options to install the documentation; I think by default it will choose 
to not owing to disk space being at a premium back when...

Here's the directory structure and the Help and PDF files I ended up 
with after the above installation completed...

C:\> cd OWATCOM
C:\OWATCOM> ddir
  Directory of  C:\OWATCOM\*
[binnt]         [binw]          [docs]          [eddat]         [h]
[lib286]        [lib386]        [samples]       [src]
               0 bytes in 0 files and 9 dirs
....

NB:  There is, of course, a directory tree underneath some of these; 
this is just the directories directly in the OpenWatcom root directory.

C:\OWATCOM> binnt\

C:\OWATCOM\binnt> dir *.hlp
  Volume in drive C is unlabeled      Serial number is BC9D:AAD0
  Directory of  c:\owatcom\binnt\*.hlp
clib.hlp      1585464   5/25/10   9:40
cw.hlp          87120   5/25/10   9:41
ddespy.hlp      78227   5/25/10   9:41
drnt.hlp        68551   5/25/10   9:41
drwatcom.hlp    93061   5/25/10   9:41
editor.hlp     187358   5/25/10   9:41
f77graph.hlp   276760   5/25/10   9:41
f77lr.hlp      372199   5/25/10   9:42
fbooks.hlp     119316   5/25/10   9:42
fpguide.hlp    369371   5/25/10   9:42
ftools.hlp     222504   5/25/10   9:42
fuguide.hlp    314587   5/25/10   9:42
f_readme.hlp   216235   5/25/10   9:42
heapwalk.hlp    97265   5/25/10   9:42
ide.hlp        439844   5/25/10   9:43
lguide.hlp     337298   5/25/10   9:43
resacc.hlp      96067   5/25/10   9:43
rescomp.hlp     61789   5/25/10   9:43
resdlg.hlp     117444   5/25/10   9:43
resedt.hlp      71762   5/25/10   9:43
resimg.hlp     128019   5/25/10   9:43
resmnu.hlp      86610   5/25/10   9:43
resstr.hlp      70663   5/25/10   9:43
spy.hlp        102621   5/25/10   9:43
vi.hlp         490716   5/25/10   9:44
wd.hlp         556540   5/25/10   9:44
wfcerrs.hlp    157783   5/25/10   9:44
wprof.hlp       80676   5/25/10   9:44
zoom.hlp        22090   5/25/10   9:44
       6,907,940 bytes in 29 files and 0 dirs
....

C:\OWATCOM\binnt> ..\docs\
C:\OWATCOM\docs> dir *.pdf
  Directory of  c:\owatcom\docs\*.pdf
cguide.pdf     869246   5/25/10   9:51
cguideq.pdf    931451   5/25/10   9:51
clib.pdf      1773310   5/25/10   9:51
clibqnx.pdf   1321209   5/25/10   9:52
clr.pdf        413947   5/25/10   9:52
cpplib.pdf    1112222   5/25/10   9:52
cw.pdf         102938   5/25/10   9:52
c_readme.pdf   391834   5/25/10   9:51
f77graph.pdf   263008   5/25/10   9:52
f77lr.pdf      497596   5/25/10   9:52
fpguide.pdf    441545   5/25/10   9:52
ftools.pdf     240701   5/25/10   9:52
fuguide.pdf    441794   5/25/10   9:52
f_readme.pdf   284609   5/25/10   9:52
guitool.pdf   1687426   5/25/10   9:52
lguide.pdf     380553   5/25/10   9:52
pguide.pdf     497122   5/25/10   9:52
tools.pdf      260382   5/25/10   9:52
vi.pdf         555256   5/25/10   9:52
wd.pdf         561006   5/25/10   9:52
      13,027,155 bytes in 20 files and 0 dirs
....
C:\OWATCOM\docs>

There is _not_ as I said an overall index of the .pdf file 
unfortunately, that is why I suggested to add a link in the STARTUP menu 
OpenWatcom folder to the ftools.hlp file; it does have such an overall 
index (to all the various .hlp files, of course).  I will send an 
enhancement request to OW to have the link automagically created in the 
next update to the installer....

In the above .hlp and .pdf files I found no references to V1.8; where 
there is a version-specific entry it is uniformly V1.9.

Again, I'm simply trying to get you to a consistent, clean install from 
which to start; I'm puzzled as to why not use the above route.

....

Again, good luck...

--
0
none1568 (7455)
8/18/2011 1:08:17 PM
On 8/18/2011 12:28 AM, W. eWatson wrote:
....

> One thing that bothers me a bit is the presence of some odd possibly
> invisible characters that enter into the file. When I tried to edit the
> rtf file where a second line was attached to the end of the previous
> line, I couldn't split them onto two lines.
>
> I talked to ABBYY's tech support today about their FineRead s/w, which
> is what I got with the Epson photo scanner. It's pretty much what I've
> seen in maybe 60 minutes of use. It has the ability to save the
> resulting read text into Word, and some other formats. So far I've just
> used the rtf format.
....

If it can't save as plain text, you'll play heck turning it into 
compilable code because of the things you've just noticed--control 
characters invisible to the screen and incompatible w/ Fortran (or any 
other compiler, for that matter).

Save the original scan as plain text if possible; if it won't allow 
that, first thing in Word or whatever is to save it from there as text 
then only use a progamming editor on it from then on.  (OBTW, note 
you'll need to convert any tabs to get the initial line spacing to 
equivalent blanks; while a few compilers will accept a TAB as an 
extension, they make life generally miserable in code and other 
compilers will barf.)

As Terrence says, you need tools specifically for the purpose; his idea 
of a comparison of two copies is a good one I hadn't thought of doing 
but I've not tried the rescue operation like this in a long, long, long, 
time and he's been through a fairly significant one w/ BMDP suite not 
terribly long ago...

--
0
none1568 (7455)
8/18/2011 1:16:46 PM
"W. eWatson" <wolftracks@invalid.com> wrote in message 
news:4E4CA2E7.6010407@invalid.com...

> Do you have a capability to read rtf?  Maybe Notebook would be sufficient.

IMO it is better to convert directly to a text file if your software can do 
that.



0
epc8 (1259)
8/18/2011 2:29:36 PM
On 8/18/2011 6:08 AM, dpb wrote:
> On 8/17/2011 11:29 PM, W. eWatson wrote:
> ...
>
>> Beats me. Well,not quite thanks to FF history search.
>> <http://www.openwatcom.org/index.php/Manuals> Look at the header Open
>> Watcom F77--below it. (Hey, how about that,Open Watcom.) If you drill
>> down, you'll find it references a "current" folder.
>> Did I miss something here for 1.9?
>
> _IF_ you will (or did) download
Of course I downloaded it, ax explained previously.
>
"it" = open-watcom-f77-win32-1.9.exe
>
> as I demo'ed for the anon ftp session, all the documentation will be
> included.
>
> Apparently that subdirectory of the OW site wasn't updated at the same
> time as the files in the distribution were refreshed--altho it probably
> doesn't matter for content, perhaps that's why you aren't finding stuff.
>
> ...
>
>> Winnt? docs? What is the complete path to them?
>>>
>>> Anyway, looks like you do have a workable compiler even if not the
>>> latest; good luck w/ your project...
>>>
>>> --
>> v1.9 is the latest, right? That's what the Help->About said.
>> NOTE: At no time did, I set up matters according the the README file,
>> which shows paths, etc. for various OSes. Was it necessary?
>
> If you run the above .exe, it has an installer that lets you choose the
> root directory (I think the default is "\WATCOM"; I chose "OWATCOM"
> because I have the former commercial installation in the "WATCOM"
> directory which didn't want to overwrite). You have to select the
> options to install the documentation; I think by default it will choose
> to not owing to disk space being at a premium back when...
It was a clean install. As I recall, there were some choices with 
dialogs along the way. I chose what I thought was right.
>
> Here's the directory structure and the Help and PDF files I ended up
> with after the above installation completed...
>
> C:\> cd OWATCOM
> C:\OWATCOM> ddir ...
> C:\OWATCOM\docs>
>
....
> There is _not_ as I said an overall index of the .pdf file
> unfortunately, that is why I suggested to add a link in the STARTUP menu
> OpenWatcom folder to the ftools.hlp file; it does have such an overall
> index (to all the various .hlp files, of course). I will send an
> enhancement request to OW to have the link automagically created in the
> next update to the installer....
>
> In the above .hlp and .pdf files I found no references to V1.8; where
> there is a version-specific entry it is uniformly V1.9.
>
> Again, I'm simply trying to get you to a consistent, clean install from
> which to start; I'm puzzled as to why not use the above route.

Frankly, I would expect simplicity in an install.

This is beginning to remind me of an ordeal I went through two weeks 
ago. An XP machine I have was installed w/o a login. A boot-up takes one 
right to the desktop. I thought it was time to provide a login, so 
posted to a NG. One of the MVP respondents had me looking everywhere to 
get it done. I finally stepped aside, and found it took two easy steps.

Well, I'll muddle on and see if what I have so far. However, first, I'd 
like to get going on the OCR.

Thanks for the help.
0
wolftracks (253)
8/18/2011 2:35:50 PM
On 8/18/2011 9:35 AM, W. eWatson wrote:
....

> Frankly, I would expect simplicity in an install.
....

It's two clicks--one to download; another to run the installer.
I, frankly, don't think it gets much simpler...particularly when one 
reflects on the purchase price! :)

Choose a few options and Voila!!!  The above took me less than 10 
minutes total I would guess the other day and, other than the one link 
to the .hlp files master index went flawlessly.

It does take a minute or so to look at the installation choices and 
check the one(s) to install the documentation that isn't checked by default.

Since you have the the full download .exe, if it were I, I'd do an 
uninstall from the Windows utility; clean any remaining traces from the 
\WATCOM directory and tree and then rerun the install wizard making sure 
to select all the documentation choices.  Then you will have a single, 
consistent set and the Startup Menu OW folder will have all the links 
(excepting the one I've noted that is easy enough to add).

--
0
none1568 (7455)
8/18/2011 4:18:37 PM
In a few minutes, I will send you three tiff files. Two are the first 
pages of a program, and the third is the last.
0
wolftracks (253)
8/18/2011 4:53:21 PM
Richard Maine <nospam@see.signature> wrote:
> JB <foo@bar.invalid> wrote:

>> IIRC at some point there was discussion about passing scalar
>> intent(in) arguments by value for procedures with an explicit
>> interface; I don't think this has been done so far.

> I won't say it is impossible, but that would have complications that
> might not be obvious to some at first glance. Mostly the complication
> has to do with having the implementation of a procedure depend on
> whether it has an explicit interface. That's a complication because
> having an explicit interface is a concept defined only in the scope
> where a procedure is referenced - not where the procedure is defined.
> There are some properties of a procedure that mean it will require an
> explicit interface in order to be referenced, but intent(in) isn't one
> of those.

You could do it in cases where the explicit interface was already
required.  If a routine has an assumed shape array, then use
pass by value for other intent(in) arguments.  That does seem
strange, though.

> Thus, a procedure with an intent(in) dummy might be referenced with an
> explicit interface in some places, but without an explicit interface in
> others. Having two versions of the procedure might work in some cases,
> but would be tricky if there were saved local variables. So-called
> link-time optimization has been mentioned here recently. I suppose that
> might be used to procrastinate the decision until "link time", which
> could allow the compiler to know things about all the referencing
> scopes.

You could have two entry points to the same procedure, though a
previous post indicates that many object formats don't allow
for multiple entry points.

-- glen

0
gah (12850)
8/18/2011 6:40:51 PM
On Aug 17, 6:55=A0pm, "W. eWatson" <wolftra...@invalid.com> wrote:
> TB NG not the FF NG. There are NGs for each.

Last time I used TB, I seem to recall that read items
are in normal text, unread items bolded.  I don't recall
a separate indicator (I could be wrong).

That will depend somewhat on the view you've selected.
I tend to use the "old fashion" list-of-articles sorted by
topic, rather than any of the threaded or tree views.

Also in TB, there's a difference between "read next"
and "read next unread", and similar.  Experiment
a little.

BTW, if TB isn't keeping track of which articles you've
read, it's got to be some sort of configuration problem,
e.g., some difficulty writing the file that keeps track
of read articles (and subscriptions, etc.).  You'd need
to check the TB FAQ and/or support page for help
troubleshooting that.  (In other words, it *should*
keep track as you expect...)

    -Ken
0
8/18/2011 7:35:12 PM
On 2011-08-18, Richard Maine <nospam@see.signature> wrote:
> JB <foo@bar.invalid> wrote:
>
>> IIRC at some point there was discussion about passing scalar
>> intent(in) arguments by value for procedures with an explicit
>> interface; I don't think this has been done so far.
>
> I won't say it is impossible, but that would have complications that
> might not be obvious to some at first glance. Mostly the complication
> has to do with having the implementation of a procedure depend on
> whether it has an explicit interface. That's a complication because
> having an explicit interface is a concept defined only in the scope
> where a procedure is referenced - not where the procedure is defined.
> There are some properties of a procedure that mean it will require an
> explicit interface in order to be referenced, but intent(in) isn't one
> of those.
>
> Thus, a procedure with an intent(in) dummy might be referenced with an
> explicit interface in some places, but without an explicit interface in
> others. Having two versions of the procedure might work in some cases,
> but would be tricky if there were saved local variables. So-called
> link-time optimization has been mentioned here recently. I suppose that
> might be used to procrastinate the decision until "link time", which
> could allow the compiler to know things about all the referencing
> scopes.

Ah, I was thinking about module procedures and forgot about separate
interface blocks for a second. Of course you're right, there would be
issues for such.  So perhaps adding more ways for passing dummy
argument aliasing, escape and constness etc. information to the
middle-end, as explained by Tobias, is the better way to go. That
being said, gfortran does support the VALUE attribute, so support for
passing arguments by value is already there, in case one would want to
handle module procedures differently.



-- 
JB
0
foo33 (1454)
8/18/2011 7:43:40 PM
On 8/18/11 1:40 PM, glen herrmannsfeldt wrote:
> Richard Maine<nospam@see.signature>  wrote:
>> JB<foo@bar.invalid>  wrote:
>
>>> IIRC at some point there was discussion about passing scalar
>>> intent(in) arguments by value for procedures with an explicit
>>> interface; I don't think this has been done so far.
>
>> I won't say it is impossible, but that would have complications that
>> might not be obvious to some at first glance. Mostly the complication
>> has to do with having the implementation of a procedure depend on
>> whether it has an explicit interface. That's a complication because
>> having an explicit interface is a concept defined only in the scope
>> where a procedure is referenced - not where the procedure is defined.
>> There are some properties of a procedure that mean it will require an
>> explicit interface in order to be referenced, but intent(in) isn't one
>> of those.
>
> You could do it in cases where the explicit interface was already
> required.  If a routine has an assumed shape array, then use
> pass by value for other intent(in) arguments.  That does seem
> strange, though.
>
>> Thus, a procedure with an intent(in) dummy might be referenced with an
>> explicit interface in some places, but without an explicit interface in
>> others. Having two versions of the procedure might work in some cases,
>> but would be tricky if there were saved local variables. So-called
>> link-time optimization has been mentioned here recently. I suppose that
>> might be used to procrastinate the decision until "link time", which
>> could allow the compiler to know things about all the referencing
>> scopes.
>
> You could have two entry points to the same procedure, though a
> previous post indicates that many object formats don't allow
> for multiple entry points.
>
> -- glen
>
Back in the late 80s (or so) the Convex compiler would make clones of 
routines.  Mostly, they looked for routines that were called (hopefully) 
many times and had an integer argument (or 2 or 3) that was a constant. 
  The clone was compiled "as if" the argument(s) were replaced by 
parameters with the same value.  Frequently the optimizer could do a 
very good job when it knew the value of the integers.  The same could be 
done for reals, logicals, etc., but the payoff seems likely to be smaller.

There's no reason why you couldn't do a similar thing for INTENT(IN) 
arguments, compile a clone and call it when it makes clear sense.  But, 
IMO, calling by value is a very very very small optimization for integer 
arguments, especially intent(in).  Any half-way decent optimizer will 
pull the values up into a register during the entry set-up and keep it 
there as long as it is needed.  The net saving on the called side will 
be an instruction issue or two.  And, those cycles will likely have 
occurred on the calling side.

Dick Hendrickson
0
8/18/2011 9:59:14 PM
Dick Hendrickson <dick.hendrickson@att.net> wrote:

(snip on passing INTENT(IN) by VALUE, I wrote)
>> You could do it in cases where the explicit interface was already
>> required.  If a routine has an assumed shape array, then use
>> pass by value for other intent(in) arguments.  That does seem
>> strange, though.

(snip, I also wrote)
>> You could have two entry points to the same procedure, though a
>> previous post indicates that many object formats don't allow
>> for multiple entry points.

> Back in the late 80s (or so) the Convex compiler would make clones of 
> routines.  Mostly, they looked for routines that were called (hopefully) 
> many times and had an integer argument (or 2 or 3) that was a constant. 
>  The clone was compiled "as if" the argument(s) were replaced by 
> parameters with the same value.  Frequently the optimizer could do a 
> very good job when it knew the value of the integers.  The same could be 
> done for reals, logicals, etc., but the payoff seems likely to be smaller.

Sounds like the ones some compilers would inline.  
The argument passing time is a significant fraction of the time
for simpler functions.  Though I still like statement functions
for the simplest ones, which are easiest for compilers to inline.

> There's no reason why you couldn't do a similar thing for INTENT(IN) 
> arguments, compile a clone and call it when it makes clear sense.  

As someone else mentioned, watch for SAVEd variables.

> But, 
> IMO, calling by value is a very very very small optimization for integer 
> arguments, especially intent(in).  Any half-way decent optimizer will 
> pull the values up into a register during the entry set-up and keep it 
> there as long as it is needed.  The net saving on the called side will 
> be an instruction issue or two.  And, those cycles will likely have 
> occurred on the calling side.

More and more keep a few arguments in registers, though usually
also leaving stack space.  That is easier for VALUE.  Then it might
be able to arrange things such that the value is already in the
right register.  But the stack arguments should be in L1 cache, 
anyway, and so pretty fast to access.

The OS/360 Fortran compilers make a local copy of scalar arguments.
That saves one instruction for every access, at the cost of four
(two on entry, two on exit).  (Two if it can stay in a register.)
For arrays, they might make a local copy of the address.

-- glen

0
gah (12850)
8/18/2011 10:57:29 PM
On Aug 18, 3:28=A0pm, "W. eWatson" <wolftra...@invalid.com> wrote:
> On 8/17/2011 6:50 PM, Terence wrote:

> > I repeat my offer to "have a go" if sent the scanned version(s) of the
> > source.
>
> Well, this is interesting. By source, do you mean the result of an OCR
> operation? I can mail you that and perhaps some pix of the pages. It
> sounds like you've got some good tools for this sort of thing.

E-mail the OCR result - it's a text file!

>
> It =A0appears you are in Britain, correct? I'd almost suggest borrowing
> the book from a college library, but I suspect it's only found in the
> USA. I borrowed mine here, but found a used one on the internet, so
> bought it.
>
 Born, educated and gowned there; reside in Australia.

> One thing that bothers me a bit is the presence of some odd possibly
> invisible characters that enter into the file.

Not a problem. They are OCR-to-Word artifacts of the Word symbol set.
There are sizes and positions of dashes and versions of apostrophies
and stars, mis-interpreted by the OCR (which also does not respect the
indents of a Fortran source listing. I found cedillas, braces,
triangular brackets and arrows in the OCR result which could not have
been in the original!

> When I tried to edit the rtf file ....
Don't try that! Cut and paste the text into Notepad, save as
whatever.for and work on the text file.

> I talked to ABBYY's tech support today about their FineRead s/w, which
> is what I got with the Epson photo scanner. It's pretty much what I've
> seen in maybe 60 minutes of use. It has the ability to save the
> resulting read text into Word, and some other formats. So far I've just
> used the rtf format.

That's OK. BUT USE ".RTF" and NOT ".DOC" See cut and past into Notepad
above. That operation removes the word or rtf formatting.

>
> I was baffled by what I was getting by laying the book flat on the
> scanner. I could easily select the left and right page, but the right
> page produced a big left margin.....

Wordstar is perfect for that operation. It allows vertical slicing and
has many other ancient valuable other-wise disappeared features. (Copy
kept in safe..)

>
> Do you have a capability to read rtf? =A0Maybe Notebook would be sufficie=
nt.
Don't care as long as the eye understands the OCR result, in which
case the editor and then the compiler will!
>
> I'll continue this tomorrow. It's getting late.

Already seen and processed OK the first three sample pages.


0
tbwright (1098)
8/19/2011 9:35:56 AM
On Aug 19, 2:53=A0am, "W. eWatson" <wolftra...@invalid.com> wrote:
> In a few minutes, I will send you three tiff files. Two are the first
> pages of a program, and the third is the last.

I got them as tif files - problem - but imager can open it - nothing
else will on my machines (compatability) .
0
tbwright (1098)
8/19/2011 9:47:23 AM
Reply:
Similar Artilces:

help needed with predict function
Hello! I wrote the following code: date=csvread('date.csv'); %the data variable contains 9356 lines and 2 columns %the first column contains the time the measurement took place %the second column contains the measured value a=size(date); y=iddata(date(1:fix(a(1)/2),2)); %i use the 1st half of the data to estimate the model using the ar function mod=ar(y,2); pred=predict(mod,iddata(date(:,2))); %end code Is pred(fix(a(1)/2)+1:end) really the prediction of my second half of data using the model estimated or it's just a simulation of the model? I want to predict the next 100 values...

CG & Music Compilation on DVD
Hello all, I own a video production company called Cinelux Studios that I run in my free time. I'm looking for some fun work to give me something to do, and I'm curious if anyone's interested in putting together a compilation of computer graphics work onto a DVD, creating something like "Beyond the Mind's Eye". The idea that I had was to gather a compilation of the latest and greatest computer graphics and computer generated music from people all over the world, and put it onto a DVD. I've got $20,000 in digital video editing equipment, a love for p...

Photo editing very slow in PSP 9 after upgrading to XP
I recently upgraded my computer to Windows XP Home from Windows 98SE. Everything seems to be working pretty well except that editing digital photos in PSP 9 seems to be extremely slow. I notice it more with larger files, but I didn't have this problem with the same size files in 98. Any ideas? Sharon sharonf wrote: > I recently upgraded my computer to Windows XP Home from > Windows 98SE. Everything seems to be working pretty well > except that editing digital photos in PSP 9 seems to be > extremely slow. I notice it more with larger files, but I > didn't have this p...

I need ...
Hi. It is my first post on this list. I can't find (in network) some program (code) writed in assembler (16bit). This program writed some guy who was in army , and had to many free time. Program used bootsector of disc. When somebody forgot your disc from floppy disc , and restart your Pc , that on monitor was a amination (fire), and text: (I don't know exactly )."Idiot !! In floppy disc is a disc " . I remember that in bootsector was jump to last sector of disc where was a program with animation. Sorry for my English , and mistakes. ...

Need Bondi Rev A iMac motherboard swap advice
I just started using our daughter's "hand-me-down" rev A iMac and I am surprised at how well it works with OS X (192 MB RAM, 20 GB HD). However, as you probably know, the graphics chip in the Rev A (Rage IIC) is pretty slow so I started thinking about swapping in a motherboard from a Rev B, C, or D. I had the impression that the Rev A, B, C and D motherboards could be swapped interchangibly, but an ebay seller of an Rev D motherboard stated in his ad that it was only compatible with the Rev B, C or D machines. Can someone enlighten me on this issue? Can later model moth...

solaris compiler und Kommentare
Hallo, auf einem solaris 8 System benutze ich den Compiler /opt/share/SUNWspro/bin/cc. Ich habe ihn nicht installiert, vielleicht hilft euch diese Ausgabe hier bei der Versionnierung weiter: #/opt/share/SUNWspro/bin/cc --V cc: Warning: option -- passed to ld cc: WorkShop Compilers 5.0 98/12/15 C 5.0 usage: cc [ options] files. Use 'cc -flags' for details Zun�chst einmal unterst�tzt der Compiler kein C99, d.h. lokale Variablen d�rfen nicht im flie�enden Text definiert oder deklariert werden. Damit kann ich leben, aber was ich noch *nie* gesehen habe, dass ein C-Compiler "/...

Linksys BEFXSx41 <-> Windows XP IPsec Tunnel
I have one of the Linksys BEFXSx41 Routers w/ VPN capability. I've been trying to get a setup where a windows xp client connects to the router over the internet to VPN. The router has a static WAN IP. Settings: Local Secure Group: Subnet 192.168.1.0 Remote Secure Group: Any Remote Secure Gateway: Any Encryption: DES Authentication: MD5 Key management: Auto IKE Use PKS (Tried both on and off) Passphrase: test phrase timeout 3600 The method set was to use PKS pre-shared key. I believe the networking on the router is configured properly, when the client connects it says th...

Data Dictionary Tool Needed
I'm looking for a tool (hopefully inexpensive or freeware) that can accept an XML format and produce an HTML, hyperlinked data dictionary. The content of the XML is class/attribute (or table/column) information. Any ideas out there? Thanks! Quinn ...

ca-tools with XP
Sorry my enghish, but i need a help i have a clipper 5.2d and work with ca-tools Lib. In Windows XP, very functions is not work, like a CapsLock and mutch more ca-tools not work with XP???? i need a help, please... thanks advanced On Sat, 28 Feb 2004 00:16:50 -0000, "The Lion" <lion-home@netcabo.pt> wrote: >Sorry my enghish, but i need a help > >i have a clipper 5.2d and work with ca-tools Lib. > >In Windows XP, very functions is not work, like a CapsLock and mutch more > >ca-tools not work with XP???? i need a help, please... > Any function whi...

need LDA / PCA matlab code and tutorials
hello PCA(principal component analysis) and LDA(linear discriminate analysis) processes are used to reduce dimensionality of a vector. can anyone provide tutorials and matlab code of these techniques. thanks mahesh Have you tried searching online? There is a large amount of high-quality material available for free on the World Wide Web. Try some combination of terms such as: "dicriminant analysis" (note the correct spelling) "source code" MATLAB I have found these search engines to be especially useful: A9 (www.a9.com) AlltheWeb (www.alltheweb.com) Alta Vista (www.alt...

C compiler internals
Hai, Can I get some docs or links to learn the C compiler internals from basic to advanced. It must contains good documentation of how compilers allocates memory to variables. Bye Guru Jois On Sun, 27 May 2007 04:52:52 -0700, Guru Jois wrote: > Hai, > > Can I get some docs or links to learn the C compiler internals > from basic to advanced. It must contains good documentation of how > compilers allocates memory to variables. > > Bye > Guru Jois There are a lot of different C compilers out there and some of them are closed source. I don't think the MS...

!!!ASSISTANCE NEEDED!!!
Dear Friend, After deliberation with my children, I decided to contact you for your assistance in standing as a beneficiary to the sum of US$15.5M(Fifteen Million,Five hundred Thousand United States Dollars Only). First, let me start by introducing myself as MRS SUSAN SHABANGU, a mother of three children and the Deputy Minister of Minerals and Energy since 1s= t April 1996 to date under the auspices of the President of South Africa MR THABO MBEKI. You can view my profile at my website www.gov.za/profiles/shabangu.htm THE PROPOSAL: After the swearing in ceremony making me the D...

Freeform database needed
I am not sure that "freeform" is the right term, but I need a database that does not use the standard table/record/field structure. I want to be able to just dump stuff into the program - text, images, URLs, etc - and then be able to locate/organize information based on keywords and logical concepts. Can anyone suggest something? -- Peter Aitken Remove the crap from my email address before using. On Tue, 06 Apr 2004 20:33:34 +0000, Peter Aitken wrote: > I am not sure that "freeform" is the right term, but I need a database > that does not use the...

xHarbour 1.2 - compiling from sources on ubuntu
I'm administrating a bunch of Ubuntu servers in a company which benefits from xHarbour for a long time now. I'm trying to compile from sources the newest version of xHarbour compiler on Ubuntu 8.04. It all goes ok, and even generates .deb file correctly. Trouble is that programmers complain that Windows version of xHarbour has such libs in it like for example libdbfcdxmt.a. I've been trying to look very closely into all of .cf files, to find any clues how to force compiling dbfcdxmt on Ubuntu, but failed. librdd on the other hand has its mt version, as well as many others. My qu...

Do I need TSIG for zone transfer on an intranet env?
is it too much? ACL should do the job? ...

Apache/mySQL/PHP on XP
Hi, I am looking for some sort of easy-to-install binary package to get an environment ready for PHP development on XP. I am starting the development of a site, starting with learning PHP and design etc. The final site will run on some sort Linux/*nix, but for the purpose of the first development phase I will have to use Windows. I know the FoxServ package, but it seems to be very slowly updated. I saw other things (Google) but they seem to be a one-man-show, not really promising if I need to keep an up to date environment for a year or more. I am looking for a precompiled environment wi...

Re: Newbie needs your help! #3
Avinash, I am not total sure why you want a data structure so abnormal as what you desribe. It seems to me that it will provide more trouble than it is worth. However, if you are insistant on such a structure here ya go: data oneb ; if eof then do ; Sex = 'Race' ; Dose1 = . ; Dose2 = . ; output ; end ; set one end = eof ; output ; run ; data three ; set oneb two (rename = (Race = sex)) ; run ; HTH Toby Dunn -----Original Message----- From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU] On Behalf Of Avinash Sent: Tuesday, November 23, 2004 12:03 PM To: SAS...

models needed
does anyone have a decent model of a laptop and / or docking station suitable for renderings? any help would be greatly appreciated. -mr. brown ...

Need Computer Help?
PC Dude computer Help and Discussion forums offer professional help for all your computer problems. Check us out at http://pcdude.proboards86.com Gary335 napisal(a): > PC Dude computer Help and Discussion forums offer professional help for > all your computer problems. Check us out at > http://pcdude.proboards86.com bleh ... Regards -- A=2EB - http://www.dzialkanadmorzem.pl/ http://www.realestate1.eu campingpl=E4tze in Polen urlaub in polen wetter polen Ostsee Polen dzialkanadmor...@onet.eu wrote: > Gary335 napisal(a): > > PC Dude computer Help...

Running on XP without administrator rights
Our standard configuration is for users to run SAS without administrator rights on their PCs. Is there a simple way to give them control over their autoexec.sas without giving them write access to the SAS directory. They can currently create a shortcut to run SAS and add a reference to their personal autoexec in the startup statement. However, this won't work when they click on a SAS file to open it interactively or in batch mode. Our standard configuration doesn't allow them to change file associations. Another approach might be to put a dummy autoexec in the SAS folder ...

Administrator Needs Help!!!!!!!!!!! #2
I administor a small isp with about 1000 customers. We have sun ultra 10 300 mhz with 512mb memory. I have solaris 2.6 running, with apache, dns, sendmail, tacacs-plus running also. My problem is that this server gets really slow at times. I think we have too much running on one server. What I need to know is if I can put some of these things, like sendmail and maybe tacacs-plus on another server. I have 3 more sun ultra 10's 300's. My other problem is I only have one domain name for the isp. All mail is routed to this domain name. If I put sendmail on another server, I will h...

Urgent:help needed
Hello All, I am using Fsh,it has Python Dependecy.It uses several moduels from Python.The problem is"Select" module in missing in the list of modules available in python.So I can't import it.But the library file needed for select module ,select.sl is there in the python(I guess).If anyone has come across such problem before,if do help me.Help me in knowing how to import the select module (from python). With Advance Thanks, Prabu.S --- Outgoing mail is certified Virus Free. Checked by AVG anti-virus system (http://www.grisoft.com). Version: 6.0.712 / Virus Database: 468 -...

Pricing Needed
Hi Julie, Please help! I need HP maintenance costs on a C7145RA-DLT Thanks ATS Chris Modenbach Phone: 800-231-5500F2skZGHV * To join/leave the list, search archives, change list settings, * * etc., please visit http://raven.utc.edu/archives/hp3000-l.html * ...

Immediate Need - Jira or Informatica Admin - Nashville, TN
Hi, I have an urgent opening for Jira or Informatica Admin, kindly send the upd= ated resume ASAP. Job Title: Jira or Informatica Admin Location: Nashville, TN Duration: 12 month+ Job Description: Must Have Skills: JIRA or Informatica ( ETL ) Admin with UNIX Support Servi= ces and Application Support experience Junior to Mid-level Technologist with knowledge of tools shall provide Oper= ational support L1, L2, and evolve to engineering function for WMA shared s= ervice platforms working in concert with WMA IT senior Architects and UBS G= lobal core engineering team suppor...

Non-volatile compiler optimizations
Hello, I'm trying to understand volatile. I have written trivial code where a variable is tested to decide whether to break early out of a loop. I've considered 4 different cases. auto variable static with function scope static with file scope external linkage AFAIU, since the abstract machine model is single threaded, if a variable is not volatile, then its value cannot change inside the loop. (Signals might be an exception?) Thus, in every case, the compiler is allowed to remove the test, because it may assume that the variable cannot change. I tested with gcc...