[comp.lang.fortran] MSDOS Fortran

lowrie@caen.engin.umich.edu (Robert Byron Lowrie) (01/18/91)

I'm looking for a good fortran compiler for my IBM PS/2 50Z.  I've heard
the Lahey compilers are the best.  Does anyone know any good or bad points
about the Lahey Personal Compiler?  Its advertised price is $99.  I'm NOT
looking for a full blown compiler with all the bells and whistles; just
a something for small programs at home.  Any large programs I can do at
work.

If anyone has any suggestions, I'd appretiate them.

Thanks,

	Rob.

gt4512c@prism.gatech.EDU (BRADBERRY,JOHN L) (01/18/91)

In article <1991Jan17.171224.19665@engin.umich.edu> lowrie@caen.engin.umich.edu (Robert Byron Lowrie) writes:
>I'm looking for a good fortran compiler for my IBM PS/2 50Z.  I've heard
>the Lahey compilers are the best.  Does anyone know any good or bad points
>about the Lahey Personal Compiler?  Its advertised price is $99.  I'm NOT
>looking for a full blown compiler with all the bells and whistles; just
>a something for small programs at home.  Any large programs I can do at
>work.
>
Based on what you described as your expectations, it looks like almost
ANY PC compatible compiler will do. However, if your application(s) include
graphics or mixed language requirements, you may find Lahey quite limiting!

Lahey is one of the least expensive at allowing you to use memory beyond
the dreaded 640k boundary, but it seems to be geared more for people who
are writing self contained program solutions that do not require a signifi-
cant of amount of library development.

By the way, beware that some versions of Lahey DO NOT initialize memory
to zero's (in case of array operations). Of course you shouldn't depend on
any compiler to do that anyway, but some people still expect their 
variables to be initialized by angels! 

-- 
John L. Bradberry        |Georgia Tech Research Inst|001100110011001100110011
Scientific Concepts Inc. |Microwaves and Antenna Lab|Int : gt4512c@prism
2359 Windy Hill Rd. 201-J|404 528-5325 (GTRI)       |GTRI:jbrad@msd.gatech.
Marietta, Ga. 30067      |404 438-4181 (SCI)        |'...is this thing on..?'   

BVAUGHAN@pucc.Princeton.EDU (Barbara Vaughan) (01/19/91)

In article <19886@hydra.gatech.EDU>, gt4512c@prism.gatech.EDU (BRADBERRY,JOHN L) writes:

>In article <1991Jan17.171224.19665@engin.umich.edu> lowrie@caen.engin.umich.edu (Robert Byron Lowrie) writes:
>>...a good fortran compiler for my IBM PS/2 50Z...
>> Does anyone know any good or bad points
>>about the Lahey Personal Compiler?  Its advertised price is $99.

>Based on what you described as your expectations, it looks like almost
>ANY PC compatible compiler will do. However, if your application(s) include
>graphics or mixed language requirements, you may find Lahey quite limiting!

I'm not sure this is true.  The F77l version of Lahey Fortran has graphic
library, screen manipulation library (includes functions to simplify
making of menus) and functions to allow passing ofsome arguments to C.
The libraries cost extra $$ but can be used with Personal Fortran as well
as the full-sized full-cost Lahey F77L.

>Lahey is one of the least expensive at allowing you to use memory beyond
>the dreaded 640k boundary

You've got to be kidding. The only Lahey version that allows you to break
640K is the F77L-EM/32, which is $895 ($995 as of April). It also
requires the purchase of QEMM or something on top of that.  The $99
Personal Fortran is strictly a 640K compiler.

>By the way, beware that some versions of Lahey DO NOT initialize memory
>to zero's (in case of array operations). Of course you shouldn't depend on
>any compiler to do that anyway, but some people still expect their
>variables to be initialized by angels!

Do you seriously think this is a major consideration in buying a compiler
or do you just have nothing else to add?  Also why did you feel compelled
to offer your help to this guy when you had no specific information to
offer?

>John L. Bradberry        |Georgia Tech Research Inst|001100110011001100110011

Barbara Vaughan

quan@sol.surv.utas.oz (Stephen Quan) (01/21/91)

gt4512c@prism.gatech.EDU (BRADBERRY,JOHN L) writes:
>In article <12277@pucc.Princeton.EDU> BVAUGHAN@pucc.Princeton.EDU writes:
>>>Lahey is one of the least expensive at allowing you to use memory beyond
>>>the dreaded 640k boundary
>>You've got to be kidding. The only Lahey version that allows you to break
>>640K is the F77L-EM/32, which is $895 ($995 as of April). It also
>>requires the purchase of QEMM or something on top of that..

1)  Lahey Personal Fortran. (???)    ($99)   - Imposes a 64K limit.
2)  Lahey Fortran.          (F77L)   ($800)  - Raises the limit to 640K.
3)  Extended Memory.        (F77LEM) ($1600) - Uses memory overlays(?).

The original poster wanted to know if there was a cheap personal Fortran
compiler.  It really depends on what he wanted to design with it.  Well,
whatever, I believe 1) or 2) would be the most appropriate.

>Your pricing and configuration information is faulty!
>If you had FULLY understood what you THOUGHT you read in this passage, I
>was indicating that a LAHEY version (not the $99 version) did perfom
>this task. I noticed YOU didn't offer any suggestions here...

Mentioning 'least expensive' and the breaking of the 640K limit they
way you had, did make it sound like the $99 version broke that limit.
The F77LEM manages to break the 640K limit by using what they call
'Overlays'.  Overlays swap/pages memory so that more than 640K can
be addressed.

>John L. Bradberry        |Georgia Tech Research Inst|001100110011001100110011
>Scientific Concepts Inc. |Microwaves and Antenna Lab|Int : gt4512c@prism
>2359 Windy Hill Rd. 201-J|404 528-5325 (GTRI)       |GTRI:jbrad@msd.gatech.
>Marietta, Ga. 30067      |404 438-4181 (SCI)        |'...is this thing on..?'   
Stephen Quan,
University of Tasmania.
"..Who will be getting a Lahey Fortran compiler, ..someday."

16012_3045@uwovax.uwo.ca (Paul Gomme) (01/22/91)

In article <12277@pucc.Princeton.EDU>, BVAUGHAN@pucc.Princeton.EDU (Barbara Vaughan) writes:
> In article <19886@hydra.gatech.EDU>, gt4512c@prism.gatech.EDU (BRADBERRY,JOHN L) writes:
> 
>>Lahey is one of the least expensive at allowing you to use memory beyond
>>the dreaded 640k boundary
> 
> You've got to be kidding. The only Lahey version that allows you to break
> 640K is the F77L-EM/32, which is $895 ($995 as of April). It also
> requires the purchase of QEMM or something on top of that.  The $99
> Personal Fortran is strictly a 640K compiler.

While I won't contest the pricing information (since I don't know what the
prices are), there are two factual errors here:

(1) I've used the extended memory version of the Lahey compiler; it does _not_
    require QEMM or any other expanded memory manager.  In fact, I could not
    get either the compiler or compiled code to run when I had QEMM loaded.

(2) the Personal Fortran from Lahey is restricted, as I recall, to 64K of user
    code, and 64K of data.  You need the "full" F77L to use 640K.

>>John L. Bradberry        |Georgia Tech Research Inst|001100110011001100110011
> 
> Barbara Vaughan
--------------------------------------------------------------------------
Bitnet: gomme@uwovax.bitnet   gomme@uwovax.uwo.ca   Internet: gomme@uwo.ca

BVAUGHAN@pucc.Princeton.EDU (Barbara Vaughan) (01/23/91)

In article <1991Jan21.133951.8347@uwovax.uwo.ca>, 16012_3045@uwovax.uwo.ca (Paul Gomme) writes:

>In article <12277@pucc.Princeton.EDU>, BVAUGHAN@pucc.Princeton.EDU (Barbara Vaughan) writes:

>> You've got to be kidding. The only Lahey version that allows you to break
>> 640K is the F77L-EM/32, which is $895 ($995 as of April). It also
>> requires the purchase of QEMM or something on top of that.  The $99
>> Personal Fortran is strictly a 640K compiler.

>While I won't contest the pricing information (since I don't know what the
>prices are), there are two factual errors here:

>(1) I've used the extended memory version of the Lahey compiler; it does _not_
>    require QEMM or any other expanded memory manager.

I wasn't sure it was QEMM (that's why I said 'or something'.  It DOES
require either Lahey Ergo OS/386 ($395) or the Ergo Developer's Kit
($695).

>(2) the Personal Fortran from Lahey is restricted, as I recall, to 64K of user
>    code, and 64K of data.  You need the "full" F77L to use 640K.

This is correct, at least as far as the code size.  Sorry for my mistake.
Obviously, by '64K of data' you don't mean the size of external data
files, which your program couldn't care less about. Data read into arrays
will increase the size of your compiled code, so in that sense the 64K
is a limit.  In addition there is a 64K limit on stack size and saved
local variables in Lahey Personal Fortran.

Barbara Vaughan