Re: "C" AND Ada

[blackje%sungod.tcpip@GE-CRD.ARPA]

Received: by sungod.steinmetz (3.2/1.1x Steinmetz)
	id AA02000; Tue, 25 Aug 87 10:18:39 EDT
Date: Tue, 25 Aug 87 10:18:39 EDT
From: emmett black <blackje@sungod>
Posted-Date: Tue, 25 Aug 87 10:18:39 EDT
Message-Id: <8708251418.AA02000@sungod.steinmetz>
To: info-ada@ada20.isi.edu
Subject: Re: "C" AND Ada

concerning translating other languages to Ada --

I do sort of agree with Kent (the man from xanth) that 
converted programs should be "rethought" rather than just
automagically and blindly translated..

And I also agree that those Cobol conversions he mentions 
create code that is less maintainable than the original...

However, I do not quite agree that we should extend that observation
into Ada...  If you JUST blindly translate from some language
into Ada and stop there -- then I still agree -- it's a mistake.
But we did a FORTRAN to Ada translator -- which I thought was a 
mistake at first -- and did use it to translate a buncha stuff...
and you DO get "BAD Ada" -- (GIGO, remember?) -- but DONT 
STOP THERE -- that bad Ada turned out to be significantly 
more useful in spotting all the nasties the FORTRAN programmers
had slipped into the code and in providing a means for "reverse
engineering" the FORTRAN -- making it EASIER to RETHINK and 
rewrite the old program -- because it was easier to get a grasp
on what the original program was supposed to do in the first place.

If somebody tried to turn in that first pass automagically translated
into Ada program -- they should get a really unfriendly reception...
but if they use that as a means to accelerate the re-engineering,
then they did OK.

--Emmett
	BlackJE@GE-CRD.ARPA