comp.lang.ada
 help / color / mirror / Atom feed
From: cis.ohio-state.edu!pacific.mps.ohio-state.edu!math.ohio-state.edu!usc!elr
Subject: LOOKING FOR FORTRAN TO AD
Date: 17 Sep 93 02:14:00 GMT	[thread overview]
Message-ID: <7.2106.238.0N5502FA@cld9.sccsi.com> (raw)

On 09-07-93, Gerben D. Blom posted to All:

GDB| Does anybody know where to find a FORTRAN to ADA translator? We will not
   | use this for porting program structures as the performance will probably
   | be insufficient. It would help us a lot in porting the large pieces of
   | math calculations. Our goal is to port a aircraft tracker written in
   | FORTRAN to ADA, such that it can be used for further development. We
   | expect that the program structures will therefore have to be changed
   | considerably. The calculation algorithms will however remain unchanged.

Gerben:  I had the pleasure of sitting through a rather lengthly presentation
made by a company (whose name now just escaped me).  They had developed a
proprietary technology to do exactly what you want done.  Although they were
thinking of putting it into a product form, the one thing which prevented it
was the vast amount of "exceptional" Fortran code.

My client had gobs of CDC Fortran code that used CDC extensions throughout
and this would have required the tool to be tuned to these extensions.  The
promise made by the company was that they would guarantee that the Ada code
produced by the process would produce exactly the same output, given the 
same input, as the original Fortran code.  Note that this does not mean that
the resultant code would be correct; it would contain exactly the same errors
as the original!  However, it would work just the same and be interchangable
with the original...

The vendor further stipulated that while the resultant code would be fully
compilable by any Ada compiler, the code would not look like well structured
Ada code.  It would, of course, be cleaned up and made pretty.  It is tough
however, to intuit structure where none existed before.

In your situation, where there are long runs of straight code, this might 
not be as obvious a problem...  

One feature of the technology described by the vendor which tickled my
fancy was its parallelism.  Since the processing of each subroutine was a
separate function, they were prepared to bring in an 18-wheeler filled with
PCs, all processing in parallel to undertake the conversion....

Unfortunately, the client chose to pass on the opportunity....  Now if I 
could only remember the vendor's name (I seem to recall that they were out
of Florida)...

Regards, Mikey <michael.hagerty@nitelog.com>
---
 * MR/2 1.54 #63 * Learn Japanese at home, cultivate a nejibiki...
 * R109U:* Usenet * Nitelog BBS * Monterey CA * 408-655-1096       

             reply	other threads:[~1993-09-17  2:14 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
1993-09-17  2:14 cis.ohio-state.edu!pacific.mps.ohio-state.edu!math.ohio-state.edu!usc!elr [this message]
  -- strict thread matches above, loose matches on Subject: below --
1993-09-17 15:43 LOOKING FOR FORTRAN TO AD cis.ohio-state.edu!pacific.mps.ohio-state.edu!math.ohio-state.edu!darwin.sura.net!dtix.dt.navy.mil!cs.umd.edu!not-for-mail
replies disabled

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox