I had a Ti-89 for a couple of weeks and while it was interesting it had a
few problems :

1) All the regression packages were far too simplistic and limited. Why have
different commands for linear, quadratic, etc., they are all the same
algorithm. No general linear least squares, no uncertainties for the
coefficients, no covariances, etc. . No nonlinear package. 

2) There are a lot of things you can't do just on lists, stats and graphing.
I understand that the calculator has to store the variables but surely it
could do this automatically.  It was most frustrating when working with
large data variables and wanting to plot certain parts of it.  They have to
be copied to other variables first, before they can even be sorted which is
lame.

3) No higher math functions. On a calculator this advanced I assumed there
would be bessel functions and the like.

4) The program editor is horrible. If the program is any length at all it
becomes very tedious to manipulate as you can't set marks and thus moving
around it difficult.

5) The manual is horrible, it doesn't even explain many of the basic things
that you need to do like in the program IO screen if the data goes off the
side or bottom how do you scroll the screen to see it. 

6) History editing is really limited. No command/filename completion, past
command editing etc. . 

There were a couple of things I did like :

1) In the programming you can execute strings as commands and reference
strings as data variables, this is a fairly powerful and flexible feature

2) There is a data variable which is essentially a matrix variable with
uneven rows and cols. It is great when working with groups of closely
related data.

3) You can readily copy commands to a text file for ease of reference later
on, and even execute them there as scripts.

So does the HP 48/49 improve on any of the first part while allowing the
second part.

Does anyone have any experience with the others like the Casio 9850 and so
on. These are a lot cheaper, how much more limited are they than something
like a HP-48 G.

-- 
 Cliff Stamp                   
 sstamp@physics.mun.ca             http://www.physics.mun.ca/~sstamp/

 The one unforgivable sin, the offence against one's own integrity, 
 is to accept anything at all simply on authority -- Maureen Johnson Long

 Anyone can hold the helm when the sea is calm.   -- Publilius Syrus

sstamp@physics.mun.ca wrote:
> Does anyone have any experience with the others like the Casio 9850 and so
> on. These are a lot cheaper, how much more limited are they than something
> like a HP-48 G.
> 
I have a casio cfx9850gb-plus as well as a hp38g, hp39g and a hp49g+.
The casio is a piece of s*** compared to the hps.

To program the casio, need to enter in token which have to be accessed through
menus. If you are not too familiar with the menu layout, it can take you ages
to find anything. No ML programming and  very basic stats features. It does 
have some finance functions. Scroll back history is a little difficult to use.
Displaying of equations are a bad to put it bluntly. The keys lack the tactile
response of a hp. Thats all I can think of at the moment. 

Comparing a casio to a hp48g is like comparing a .... well
the difference is so great it fail to be comparable, you can't even say its 
like comparing apples to oranges 'cos it would be an insult to fruit.

The only why I have one at all is cos it was what the school recomended. Used
it for 6 months and I ditched it for a hp38g. Still keep it as a reminder of
how bad graphics calcs can get.
 
-- 

Wing Wong.
Webpage: http://wing.ucc.asn.au

sstamp@physics.mun.ca wrote:


> 3) You can readily copy commands to a text file for ease of reference later
> on, and even execute them there as scripts.
> 
> So does the HP 48/49 improve on any of the first part while allowing the
> second part.
> 
> Does anyone have any experience with the others like the Casio 9850 and so
> on. These are a lot cheaper, how much more limited are they than something
> like a HP-48 G.
> 

question:  working operation like tanh(a+ib), asin( >1) or with
complex argument on TI89

(hp41cv?), (hp33s?), HP42s, hp48x and hp49x do it,
but not TI83 for example.

ie. can TI89 handle complex numberspace on all trig, Y^X, log, exp etc. 
functions as well as HP ???




/TE

sstamp@physics.mun.ca wrote:
> I had a Ti-89 for a couple of weeks and while it was interesting it had a
> few problems :
> 
> 1) All the regression packages were far too simplistic and limited. Why have
> different commands for linear, quadratic, etc., they are all the same
> algorithm. No general linear least squares, no uncertainties for the
> coefficients, no covariances, etc. . No nonlinear package. 

There is a nonlinear regression package by Don Phillips on ticalc.org.

> 2) There are a lot of things you can't do just on lists, stats and graphing.
> I understand that the calculator has to store the variables but surely it
> could do this automatically.  It was most frustrating when working with
> large data variables and wanting to plot certain parts of it.  They have to
> be copied to other variables first, before they can even be sorted which is
> lame.

I understand what you mean about SortA/SortD, but I don't really
understand the other stuff (e.g. you can graph a function without
storing it to a variable). Can you clarify?

> 3) No higher math functions. On a calculator this advanced I assumed there
> would be bessel functions and the like.

MathTools has Bessel functions and various other special functions.

> 4) The program editor is horrible. If the program is any length at all it
> becomes very tedious to manipulate as you can't set marks and thus moving
> around it difficult.

It's easier on a V200 or TI-92+.

> 5) The manual is horrible, it doesn't even explain many of the basic things
> that you need to do like in the program IO screen if the data goes off the
> side or bottom how do you scroll the screen to see it. 

The help entry for the Pause command mentions this.

> 6) History editing is really limited. No command/filename completion, past
> command editing etc. . 

Samuel's "Complete" does command/filename completion. I'm not sure
what you mean by "past command editing" (you can paste history items
to the author line and edit them).

Cheers,
Bhuvanesh.

Torbj�rn Ekstr�m <toek@algonet.se> wrote:
> can TI89 handle complex numberspace on all trig, Y^X, log, exp etc. 
> functions as well as HP ???

Yes.

--
Bhuvanesh

The TI89 does correctly do most of the trig/transendentals with complex
args.  Better, in some ways.

For example, on the HP 49G in EXACT mode,
2 asin ->num
returns (1.5707.....,-1.3169.....)

On the TI89, asin(2) returns an exact expression:
(pi/2) - (ln(4*sqrt(3)+7)/2) * i

"Torbj�rn Ekstr�m" <toek@algonet.se> wrote in message
news:avJWc.101636$dP1.357374@newsc.telia.net...
> sstamp@physics.mun.ca wrote:
>
>
> > 3) You can readily copy commands to a text file for ease of reference
later
> > on, and even execute them there as scripts.
> >
> > So does the HP 48/49 improve on any of the first part while allowing the
> > second part.
> >
> > Does anyone have any experience with the others like the Casio 9850 and
so
> > on. These are a lot cheaper, how much more limited are they than
something
> > like a HP-48 G.
> >
>
> question:  working operation like tanh(a+ib), asin( >1) or with
> complex argument on TI89
>
> (hp41cv?), (hp33s?), HP42s, hp48x and hp49x do it,
> but not TI83 for example.
>
> ie. can TI89 handle complex numberspace on all trig, Y^X, log, exp etc.
> functions as well as HP ???
>
>
>
>
> /TE
>
>

> 1) All the regression packages were far too simplistic and limited. Why have
> different commands for linear, quadratic, etc., they are all the same
> algorithm. No general linear least squares, no uncertainties for the
> coefficients, no covariances, etc. . No nonlinear package. 

Have you tried the statistics and list editor flash app?


> 6) History editing is really limited. No command/filename completion, past
> command editing etc. . 

I wrote a program that adds auto completion to the main entry line. 
It is avaialable here:

http://www.stearley.org/calc.html





> 
> There were a couple of things I did like :
> 
> 1) In the programming you can execute strings as commands and reference
> strings as data variables, this is a fairly powerful and flexible feature
> 
> 2) There is a data variable which is essentially a matrix variable with
> uneven rows and cols. It is great when working with groups of closely
> related data.
> 
> 3) You can readily copy commands to a text file for ease of reference later
> on, and even execute them there as scripts.
> 
> So does the HP 48/49 improve on any of the first part while allowing the
> second part.
> 
> Does anyone have any experience with the others like the Casio 9850 and so
> on. These are a lot cheaper, how much more limited are they than something
> like a HP-48 G.

<sstamp@physics.mun.ca> wrote in message
news:cgfdl9$dmv$1@coranto.ucs.mun.ca...
> I had a Ti-89 for a couple of weeks and while it was interesting it had a
> few problems :
>
> 1) All the regression packages were far too simplistic and limited. Why
have
> different commands for linear, quadratic, etc., they are all the same
> algorithm. No general linear least squares, no uncertainties for the
> coefficients, no covariances, etc. . No nonlinear package.

Both calcs have 3rd party FREE programs to remedy this.

> 2) There are a lot of things you can't do just on lists, stats and
graphing.
> I understand that the calculator has to store the variables but surely it
> could do this automatically.  It was most frustrating when working with
> large data variables and wanting to plot certain parts of it.  They have
to
> be copied to other variables first, before they can even be sorted which
is
> lame.

HP 49 does it better.

> 3) No higher math functions. On a calculator this advanced I assumed there
> would be bessel functions and the like.

Again you need a third party program for both calcs.
I was hoping that ProductLog eg. Lambert W would become standard.

> 4) The program editor is horrible. If the program is any length at all it
> becomes very tedious to manipulate as you can't set marks and thus moving
> around it difficult.

You need yet another program for that for both calcs.

> 5) The manual is horrible, it doesn't even explain many of the basic
things
> that you need to do like in the program IO screen if the data goes off the
> side or bottom how do you scroll the screen to see it.

Neither calc has a complete 34763 page manual.

> 6) History editing is really limited. No command/filename completion, past
> command editing etc. .

Command completion needs a 3rd party program for both calcs.

> There were a couple of things I did like :
>
> 1) In the programming you can execute strings as commands and reference
> strings as data variables, this is a fairly powerful and flexible feature

HP 49G has even more this kind of flexibility.

> 2) There is a data variable which is essentially a matrix variable with
> uneven rows and cols. It is great when working with groups of closely
> related data.

Yes

> 3) You can readily copy commands to a text file for ease of reference
later
> on, and even execute them there as scripts.

Yes

> So does the HP 48/49 improve on any of the first part while allowing the
> second part.

Not that much except for greater flexibility in programming.

> Does anyone have any experience with the others like the Casio 9850 and so
> on. These are a lot cheaper, how much more limited are they than something
> like a HP-48 G.

Casio is not at the level of the HP & TI
If you go cheap go Sharp.
{VPN}

Bhuvanesh <lalu_bhatt@yahoo.com> wrote:

> There is a nonlinear regression package by Don Phillips on ticalc.org.

As the calculator is programmable, I realize that you can find utilities to
do most anything. As a programer myself (mainly fortran), I would not have a
problem in writing the routines I needed, already have for the main
workstation I use. However on a calculator which osts this much, I would
expect a more powerful builtin functionality. Some of the things I mentioned
are so basic you hit them in introductory stat and physics courses for
example.

>> 2) There are a lot of things you can't do just on lists, stats and graphing.
>> I understand that the calculator has to store the variables but surely it
>> could do this automatically.  It was most frustrating when working with
>> large data variables and wanting to plot certain parts of it.  They have to
>> be copied to other variables first, before they can even be sorted which is
>> lame.
> 
> I understand what you mean about SortA/SortD, but I don't really
> understand the other stuff (e.g. you can graph a function without
> storing it to a variable). Can you clarify?

If you have a large data variable you can't do histograms of parts of it
with commands like :

hist data(1),data(2)
hist data(3),data(4)
hist data(5),data(6)

but you have to do something like

data(1) -> x, data (2) -> y

hist x,y

repeat

[yes I know this isn't the syntax]

Yes you can jam them into the stats/list editor, but its limited to six
lists, plus I would like a simple command line functionality so you can do
things like :

hist data(5),(data(6)-data(4))*scale

for a bunch of scales, and so on, without having to create distinct lists
for them.

It is even limiting if you are just working with a few lists say x,y,z,t
and want to plot a few variations on them like x vs ln t, then x vs sqrt t,
etc. . 

> Samuel's "Complete" does command/filename completion. I'm not sure
> what you mean by "past command editing" (you can paste history items
> to the author line and edit them).

Most operating systems, have past command editing in that you can edit past
commands by number !n (calls the n'th history command) and then modify them
with !n:s/tvar/svar/. It would be nice to have the same editor able to
stream through programs as a simple search and replace.

-- 
 Cliff Stamp                   
 sstamp@physics.mun.ca             http://www.physics.mun.ca/~sstamp/

 The one unforgivable sin, the offence against one's own integrity, 
 is to accept anything at all simply on authority -- Maureen Johnson Long

 Anyone can hold the helm when the sea is calm.   -- Publilius Syrus