[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference turris::languages

Title:Languages
Notice:Speaking In Tongues
Moderator:TLE::TOKLAS::FELDMAN
Created:Sat Jan 25 1986
Last Modified:Wed May 21 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:394
Total number of notes:2683

135.0. "Intermediate language ?" by LSNCSC::TAVERNON (Alain Tavernon, CSC Switzerland) Wed Apr 15 1987 06:11

    I have a customer who is writing a MODULA-2 compilier. He asked
    me following question:
    
    Is that true that every VAX compiler is producing an intermediate
    language, who is the same for all compilers, in a first pass of
    compilation. 
    If this is true is it possible to get the specifications of such
    an "intermediate language".
    
    As I am complete neophyte in compilers could answer to this question.
    
    Alain 
     
T.RTitleUserPersonal
Name
DateLines
135.1TLE::BRETTWed Apr 15 1987 09:478
    Is it true
    - No
    
    Is it possible to get such a spec for any of our compilers 
    - Not that I'm aware of
    
    
    /Bevin
135.2Engineering a CompilerTLE::FAIMANNeil FaimanWed Apr 15 1987 10:3714
    Digital's compiler technology -- our algorithms, data structures,
    intermediate langauages, symbol table designs, etc. -- is
    proprietary technical information, important to maintaining the
    competitiveness of our products.  Any documentation on these is
    clearly for Digital internal use only. 
    
    However, it is no secret that several of the VAX/VMS compilers share
    a common back end, the VCG (VAX Code Generator).  The book
    _Engineering_a_Compiler_, by Anklam, Cutler, Heinen, and MacLaren
    (available from Digital Press) describes this code generator in some
    detail (including an overview of its intermediate language).  You
    might want to recommend this book to your customer. 
    
	-Neil
135.3Co-operation?CHOVAX::YOUNGBack from the Shadows Again,Wed Apr 15 1987 22:445
    As we do not currently have a (marketed) Modula-2 compiler perhaps
    you could pursue some kind of joint venture?  (I don't know how
    these things work.)
    
    --  Barry
135.4Cooperation ,why not ?LSNCSC::TAVERNONAlain Tavernon, CSC SwitzerlandThu Apr 16 1987 05:2210
    Thanks for the answers, the customer will buy the book but is a
    little be frustrated.
    
    Barry ,for cooperation I think that this customer could be very 
    interested but I don't know how to proceed and here we have no
    resources or contact for that (I work in a local support
    organization).  Anybody interested to go further with 
    cooperation ? If yes contact me (I am out from 21 APR to 4 MAY).
    
    Alain 
135.5We already have a Modula compilerTOKLAS::FELDMANPDS, our next successThu Apr 16 1987 11:179
    You should know that Digital already owns a Modula compiler (Ultrix, I
    believe), done at DEC Western Research Labs.  So far, the people who
    decide such things have decided not to go into the Modula compiler
    business (although our compiler is available to educational
    institutions).  This information makes me doubt that DEC would be
    interested in a cooperative effort, though you should still try to get
    an answer from someone in authority.
    
       Gary
135.6our modula compilerOPHION::HAYNESCharles HaynesThu Apr 16 1987 18:046
    Since I don't think this is the place to talk about product plans,
    talk to Joel McCormack about this.
    
    (SONORA::JOEL)
    
    	-- Charles
135.7License might be possible if they *add value*THE780::PEIRCEMichael Peirce -- Santa ClaraThu Apr 16 1987 20:337
    A company I used to work for (Sentry/Schlumberger) built their Pascal
    based test language (Pascal w/ statement like X := 5 omhs;) on DEC's Pascal
    compiler.  I believe they licenced the sources and went from there.
    I also believe this was expense for them, but less expensive than
    developing their own compiler.
    
    -- michael
135.8TLE::NOLANFri Apr 17 1987 13:1614
    
    	There is Modula compiler available for VAX/VMS but not from
    DEC.  It is available from Logitech, I believe.  I works just like
    any other VAX  language, conforms to the VAX calling standard (as
    far as I can tell), presents itself to the debugger as PASCAL (which
    is reasonable).
    
    	It is, however, a mutli pass compiler built from the PDP-11
    Modula compiler and so compiler times are not good.  Generated code
    is OK but not great - it was written at a University in Germany,
    where the language was more important that how fast it ran
    
    chris.
    
135.9no more source licensesTLE::SPENCERJoyce K. SpencerFri Apr 17 1987 17:0413
RE: < Note 135.7 by THE780::PEIRCE "Michael Peirce -- Santa Clara" >

    ...
    A company I used to work for (Sentry/Schlumberger) built their Pascal
    based test language (Pascal w/ statement like X := 5 omhs;) on DEC's Pascal
    compiler.  I believe they licenced the sources and went from there.
    ...

    At one time, Digital did sell a source license for the VAX
    PASCAL compiler, but not for quite some time now (since
    approximately the VAX PASCAL V2.1 timeframe).

    					Joyce Spencer
135.10I'm not sure Licensing from DEC works wellTHE780::PEIRCEMichael Peirce -- Santa ClaraMon Apr 20 1987 14:3810
    RE: .9
    
    Sounds about right.  They licesed the V1 compiler and got burned
    with from the incompatabilities betweent V1 & V2 of Pascal.  Of
    course these people had a history of getting burned - they licensed
    the VMS V3 debugger (to add commands to access the tester hardware
    - they had a 730 embedded in their tester) and this caused a few
    problems when V4 came along.
    
    -- michael
135.11Clarification of .2TLE::FAIMANNeil FaimanTue Apr 21 1987 12:1314
    It has been suggested that I provide the following clarification to
    my reply .2: 
    
    	The VAX Code Generator is not a product, and is not available
        to customers.  The (public) information that I provided about
        it is meant purely to clarify the response to the question
        about a "common Digital intermediate language".  
        
        I have suggested the _Engineering_a_Compiler_ book purely
        as a textbook which happens to contain more information about
        the VCG.  It is certainly not the best general textbook on
        compiler design and implementation.
        
    -Neil    
135.12WRL Modula-2 compiler for VMS existsCOOKIE::DOUCETTEChuck Doucette - Database A/D @CXOTue Jun 23 1987 14:2728
	There is a version of the DEC WRL (Western Research Lab) Modula-2
compiler that exists for VMS (it works - but it doesn't offer the level of
support that other compilers do like interfacing with VAX DEBUG [yet anyway]).
However, I don't think it is available to customers. Once again, send mail to
Joel McCormack for details (SONORA::JOEL).

	As you may have inferred from previous notes - it does generate an
intermediate language similar to p-code. I do believe the document that Mike
Powell wrote describing this intermediate language is public domain. So, if
you are interested in a specification of this intermediate language - Joel may
be able to help you. Also, as others have just stated, other DEC compilers do
not generate this intermediate language (if any) - so you can't draw any
conclusions about being compatible with other DEC compilers. Mike Powell has
also written other papers about why and how he built this compiler. Those may
also be available.

	I have started a conference to describe this compiler, the language it
accepts, its compilation speed, and the speed of code it generates (which I
might add is quite exceptional - see the topic on "VAX compilers speed" for
details). If you are interested in reading this conference (COOKIE::MODULA2)
please send me mail and I will add you as a member (browsers are welcome).

	By the way, if you or your customer are disappointed not to find any
DEC compiler product for Modula-2, perhaps you could note your concern with
the appropriate people (someone in TLE I expect).

Chuck