Re: Building a translator. Want to know if this is feasible...

Yes, its possible and practical to do this, *if* (collective) You
still have knowledge of the syntax and semantics of the language.
(Of course, you have to know about *everything* the
application actually uses. )


We've done essentially this for JOVIAL to C translations
(350K SLOC systems in one translation step).
We had "one" reference manual that doesn't quite match
what the compiler does, but we also had access to
a team of developers that are still actively developing(!)
in this language.


If you have access to the information, it is simply a little
painful to build the translator, because you get to go
through a number of trial translator iterations before
you decide you have it. We did about 8 iterations.


Having good infrastructure is the next major hurdle. (We naturally
used our DMS for this). That really helps you get past "minimal
translator development personnel resources"


Building the translator is hard. Convincing yourself it is right
(or right enough) is equally hard. You need to either test
the translator itself heavily (we did some of that),
and/or have good application tests that can be reapplied
after translation (thank heaven this customer had that).
Translator infrastructure doesn't help you with this part.


Bottom line: there's more work here than you think.
But it isn't driven by the formality of the language definition documents..


--
Ira D. Baxter, Ph.D., CTO 512-250-1018
Semantic Designs, Inc. www.semdesigns.com


"socrates" <socrates171@yahoo.com> wrote in message
> Now my "pending task" will be to make a translator that converts this
> dead language to an object-oriented language like java.

Return to the comp.compilers page.
Search the comp.compilers archives again.