making a case for Fortran 
Author Message
 making a case for Fortran

greetings,

I usually never read the fortran group, being involved in the c and object  
oriented worlds instead, but i might be induced to change now.

I am planning to attend graduate school in statistics and i am looking for  
a general purpose language that once i get good at will do everything i  
need.

i need:

easy and powerful matrix math.

good numerical anlysis routinues

good error trapping and debugging.

good graphics, the kind i can whipp together and paste into a paper.

the ability to be able to build off of what i've done before, ie build  
routinues in one class and use them for all others.

does Fortran get me there?  if so which compiler do you recommed.  i'll go  
out a buy a new computer in the fall and i was thinking of a powerPC or a  
windows clone.  pluses if its also portable to a unix box back and forth.

thnaks for your advice.

joe alotta.



Sun, 01 Dec 1996 04:18:01 GMT  
 making a case for Fortran

    K> greetings,
    K> I usually never read the fortran group, being involved in the c and object  
    K> oriented worlds instead, but i might be induced to change now.

    K> I am planning to attend graduate school in statistics and i am looking for  
    K> a general purpose language that once i get good at will do everything i  
    K> need.

    K> i need:

    K> easy and powerful matrix math.

    K> good numerical anlysis routinues

    K> good error trapping and debugging.

    K> good graphics, the kind i can whipp together and paste into a paper.

    K> the ability to be able to build off of what i've done before, ie build  
    K> routinues in one class and use them for all others.

    K> does Fortran get me there?  if so which compiler do you recommed.  i'll go  
    K> out a buy a new computer in the fall and i was thinking of a powerPC or a  
    K> windows clone.  pluses if its also portable to a unix box back and forth.

Well, I hate to say it on a fortran group, but I don't think fortran
is what you want.

But I could be wrong.

Fortran does not include any graphics. Its possible to get good
de{*filter*}s, but I think the C world has fortran beat there.

There are very very good numerical analysis routines widely available.

Fortran's arrays are easy to use but there's no easy dynamic memory
allocation which means an awful lot on a low-memory machine like a PC.

I would suggest that you probably will need to know fortran if you are
doing heavy statistical analysis (although I have done some in fortran
and some in C and perfer C if you can get the math packages to do it
there).

Sounds to me like what your really looking for is one of the advanced
PC statistical analysis packages, like sigmaplot (I think that name is
right). Some of the modern spreadsheets can really kick some
statistical {*filter*}also, like excel. These tools have the graphics
built-in in a way that fortran will never hope to match.

Good luck.

--

 (301) 286-2396                          fax: (301) 286-1754
Geek code: GAT d? -p+ c++++ l u+++ e- m+ s+++/++ n+ h--- f? !g w+ t++ r y++



Sun, 01 Dec 1996 21:29:18 GMT  
 making a case for Fortran

Quote:
>Fortran's arrays are easy to use but there's no easy dynamic memory
>allocation which means an awful lot on a low-memory machine like a PC.

When did they remove dynamic memory allocation from the Fortran (F90)
standard?

Dan
--
Dan Pop
CERN, CN Division

Mail:  CERN - PPE, Bat. 31 R-004, CH-1211 Geneve 23, Switzerland



Sun, 01 Dec 1996 23:27:45 GMT  
 making a case for Fortran

Quote:


> K> ...

> K> I am planning to attend graduate school in statistics and i am looking for  
> K> a general purpose language that once i get good at will do everything i  
> K> need.

> K> i need: ...[list of requirements omitted]
> ...[responses on several requirements omitted]

> Fortran's arrays are easy to use but there's no easy dynamic memory
> allocation which means an awful lot on a low-memory machine like a PC.

I agree with most of Mr. Hartnett's responses (especially his
suggestion to look into the various PC and MAC statistical analysis
packages), but I'd like to add my $.02 on this particular point:

   * These days a PC is not necessarily a low-memory machine.
   * Dynamic memory allocation _is_ available in Fortran 90.

Several Fortran 90 implementations are already available for PCs and
the Lahey Fortran 90 is due out by fall, I believe.

Quote:
> Good luck.

Ditto.

--

Stanford Linear Accelerator Center, MS 97 | reflect those of SLAC,
P.O. Box 4349; Stanford, CA  94309        | Stanford or the DOE



Mon, 02 Dec 1996 00:35:04 GMT  
 making a case for Fortran

Quote:

>     K> I am planning to attend graduate school in statistics and i am looking for  
>     K> a general purpose language that once i get good at will do everything i  
>     K> need.

>     K> i need:
>     K> easy and powerful matrix math.
>     K> good numerical anlysis routinues
>     K> good error trapping and debugging.
>     K> good graphics, the kind i can whipp together and paste into a paper.
>     K> the ability to be able to build off of what i've done before, ie build  
>     K>     routinues in one class and use them for all others.

>     K> does Fortran get me there?  if so which compiler do you recommed.  i'll go  
>     K> out a buy a new computer in the fall and i was thinking of a powerPC or a  
>     K> windows clone.  pluses if its also portable to a unix box back and forth.

> I would suggest that you probably will need to know fortran if you are
> doing heavy statistical analysis (although I have done some in fortran
> and some in C and perfer C if you can get the math packages to do it
> there).

> Sounds to me like what your really looking for is one of the advanced
> PC statistical analysis packages, like sigmaplot (I think that name is
> right). Some of the modern spreadsheets can really kick some
> statistical {*filter*}also, like excel. These tools have the graphics
> built-in in a way that fortran will never hope to match.

If you are C/C++ literate, the simplest solution *might* be to go ahead and
use the fortran libraries, but link them to a C program.  We utilize hybrid
programs here with good effect.  The down side is that you'd have to get
two compilers and you'd need to learn how to hook C and Fortran code (really
not a problem as long as you stay away from passing character strings and you
have compatible compilers).  You may already have a C compiler you're comfortable
using.
As far as a operating system goes, I haven't used it before, but you might be able
to get good use out of Linux, the freely available unix OS for PCs.  I know
of the existance of a number of sources of freely available numerical routines
for C, C++, and Fortran (write me if interested...), but I haven't checked any
of them out.

Bryce



Mon, 02 Dec 1996 02:09:52 GMT  
 making a case for Fortran
: greetings,

: I usually never read the fortran group, being involved in the c and object  
: oriented worlds instead, but i might be induced to change now.

: I am planning to attend graduate school in statistics and i am looking for  
: a general purpose language that once i get good at will do everything i  
: need.

: i need:

: easy and powerful matrix math.

: good numerical anlysis routinues

: good error trapping and debugging.

: good graphics, the kind i can whipp together and paste into a paper.

: the ability to be able to build off of what i've done before, ie build  
: routinues in one class and use them for all others.

: does Fortran get me there?  if so which compiler do you recommed.  i'll go  
: out a buy a new computer in the fall and i was thinking of a powerPC or a  
: windows clone.  pluses if its also portable to a unix box back and forth.

: thnaks for your advice.

: joe alotta.

Try learning MATLAB.  It is a matrix language which has reasonably good
graphics.  It is easy to build a big program out of smaller pieces and
the little pieces are pretty simple to debug.

I'm sure that commercial packages will work quite well for statistics,
but if you decide to take a controls class, or do FFTs MATLAB will do
the trick as well.

On a cost basis, MATLAB is fairly expensive outside of academics (there
is a low cost student version which is limited in ability).  But there is
a shareware version out on the net for PCs.  Fortran prices vary from
$100 to $1200+, but once again there are academic discounts.  There have
been several posts to this group about all of the choices, but since
I just use Lahey 16 bit Fortran ($250) I can't speak as to which one
best compares to MATLAB in terms of price.

Mark
Newton, MA

PS  You can, of course, do everything in Fortran too.  It will just be
    more time consuming and error prone.



Mon, 02 Dec 1996 06:24:51 GMT  
 making a case for Fortran
I highly recommend that you use C or better yet C++. There is a LOT of
shareware/free code libraries around to help with the things you
mentioned. Also, if you want graphics on a PC platform, C is the way to go.

It is VERY easy to call FORTRAN routines from C. The reason I mention
this is that there are quite a few GOOD math libraries written in
FORTRAN. You may want/have to use some of them. MOST of the good
FORTRAN libraries have also been translated to C.

---
========================================================================
=                       =                       =                      =
= Charles W. Stump II   = Leverage              = Reverse Engineering, =
=                       = Technologists Inc.    = Reengineering, and   =
=                       =                       = Quality Engineering  =
= Sr. Software Engineer = PO Box 4638           = tools and services.  =
=                       = Rockville, MD         =                      =

=                       = (301)309-8783         =                      =
========================================================================



Mon, 02 Dec 1996 23:58:05 GMT  
 making a case for Fortran

Quote:

>I am planning to attend graduate school in statistics and i am looking for  
>a general purpose language that once i get good at will do everything i need.
>i need:
>easy and powerful matrix math.
>good numerical anlysis [sic] routinues
>good error trapping and debugging.
>good graphics, the kind i can whipp [sic] together and paste into a paper.
>the ability to be able to build off of [sic] what i've done before, ie build  
>routinues [sic] in one class and use them for all others.

What about using a commerciallly-available graphically-oriented tool:
 - BMDP/Diamond,
 - SAS/JMP,
 - PV~wave,
which already has graphics-output, a scripting-language, and interfaces to
high-quality statistical-analysis routines, rather than doing "roll-your-own"?


Sat, 07 Dec 1996 07:47:44 GMT  
 making a case for Fortran


->
->>Fortran's arrays are easy to use but there's no easy dynamic memory
->>allocation which means an awful lot on a low-memory machine like a PC.
->>
->When did they remove dynamic memory allocation from the Fortran (F90)
->standard?

Fortran has never had dynamic memory allocation until the Fortran 90 compilers
were released, although several vendors did add it as a non-standard extension.  Unfortunately a LOT of sites don't have F90 yet and may not for quite some time
to come.



Sun, 08 Dec 1996 07:28:37 GMT  
 making a case for Fortran

Quote:


>->
>->>Fortran's arrays are easy to use but there's no easy dynamic memory
>->>allocation which means an awful lot on a low-memory machine like a PC.
>->>
>->When did they remove dynamic memory allocation from the Fortran (F90)
>->standard?

>Fortran has never had dynamic memory allocation until the Fortran 90 compilers

Not true. Fortran has never had dynamic memory allocation until the Fortran 90
standard was released.

Quote:
>were released, although several vendors did add it as a non-standard extension.  Unfortunately a LOT of sites don't have F90 yet and may not for quite some time
>to come.

Have I ever stated that dynamic memory allocation was present in Fortran
_before_ F90? I don't think so.

The fact that F90 is not available everywhere does not affect the fact that
Fortran has dynamic memory allocation.

Dan
--
Dan Pop
CERN, CN Division

Mail:  CERN - PPE, Bat. 31 R-004, CH-1211 Geneve 23, Switzerland



Sun, 08 Dec 1996 16:13:08 GMT  
 making a case for Fortran

Quote:
Schell) writes:

Krist,

Fortran can sorta do what you want.  It depends, really, on how much
you want to get into it.  I work in statistics in Fortran, but I'd
not depend on my own graphics programs for graphics:  I'd not spend
the man-years of effort to duplicate what is available on the market
<grin>.

If you seriously want to do object-oriented real statistics with
graphics and all, I suggest you consider S-Plus.  You will get more
statistics done, and have your output faster, and have more
off-the-shelf tools, and excellent graphics.  (I am not affiliated
with StatSci).  You can contact StatSci and they'll send you a Video
Tape about their system.  It will really blow you away.  In no way
could you achieve such capabilities "rolling your own" programs,
unless, of course, you want to spend man-years and go into
competition with StatSci <grin>.

I've used SAS, SPSS, BMDP, and the rest, and I've been around Fortran
since FORTRAN II.  You'll see a majority of the statistical programs
have historically been written in Fortran (although early on it was
Algol, and recently some C, C++, and Pascal).  You'd probably be well
served to learn Fortran.  It is non-proprietory and quite portable
and will be around forever.  If you want to program for a living,
then Fortran and C++ will be "a good thing".  You'd be able to use
Fortran in Dynamic Link Libraries (or whatever) and use C++ for the
graphical interface and event-driven programming of modern operating
systems.  For a grad student, I'd skip the C++ stuff unless you want
to add another year to your studies.  

If you want to do statistics that have already been described (not
research stuff), I'd suggest Systat.  It's relatively inexpensive,
fast, has good graphics.

For what it's worth, in my work I use Fortran, C, and SAS.

Anyway, for S-Plus (unix and Windows) info you should call:

   Statistical Systems, Inc.
   1700 Westlake Ave N.
   Suite 500
   Seattle WA 98109
   (205) 283-8802

My opiniions are my own, and in no way consititute an endor{*filter*}t by
anyone else, especially the government.

  Jon Richards - Operations Research Analyst
  MESC-NBS-DOI



Tue, 10 Dec 1996 23:37:01 GMT  
 
 [ 11 post ] 

 Relevant Pages 

1. making a case for FORTRAN

2. Is this safe? (Making string comparisons ignore case)

3. making the Python case to my dept.

4. making the Python case to my dept

5. Making an engineering case for Lisp

6. Upper case / Lower case I'm a lost case

7. CASE vs case vs Case...

8. Making REXX using FORTRAN

9. PLEASE HELP re: making Windows App's using Fortran subroutines

10. Making fortran libraries

11. Why is Fortran not case-sensitive?

12. making library in Lahey fortran

 

 
Powered by phpBB® Forum Software