Hi Zack!
20-Oct-95 14:37:00, Zack Jones wrote to Brian Mclaughlin
          Subject: What's Wrong with This?
 >> Or, they are in the PB Forum Library on CIS, as PBFILE.ZIP and
 >> MAX12A.ZIP (to allow for the 6.3 limitations on CIS).
 ZJ> I'll grab 'em from CI$ the next time I call in there.
 ZJ> Take Care, Zack Jones
 ZJ> zack@hom.net
    Congratulations on your promotion to co-moderator of our humble
    echo.  I think you'll find that there is less trafic here, but
    that the trafic you do get is somewhat more meaningful than the
    trafic in QUIK_BAS, since we're not talking about a dead language
    here.
    I thought I'd bring you up to speed as to what I've been up to
    in PowerBASIC.
    I've been working for over a year on the Fermat project.  If you
    remember from the QUIK_BAS days, Fermat is the name Calvin French
    gave to the theoretical object oriented BASIC I discussed in an
    echo-article called "OOB, Posibility or Pipedream?"
    The first step to writing an OOB to structured BASIC preprocessor
    was to design a pattern matching language that is robust enough
    to parse source code for syntax.  I meant to write a grep-like
    pattern matcher when I started out, possibily a BASIC lex or yacc,
    but it went much further than that and I ended up writing a complete
    language within a language called LPM (Laleh's Pattern Matcher).
    LPM is now at inhouse version 4.0 and is about 50 times more powerful
    than I had planned it to be, but in computing there is no such thing
    as too fast, too robust, or too expressive.  The first two versions
    of LPM were written in VBDOS.  When I ported over to PowerBASIC,
    I ended up adding some features to LPM that I hadn't even considered
    under VBDOS for performance reasons.
    The next step of the Fermat project will be to use the LPM tool
    to write the preprocessor itself.  I wrote a prototype preprocessor
    back in February that did the job, but demanded absolute syntactical
    adherence or it would crash.  There's no use having an OOL if you
    can't make a mistake without choking the system.  LPM is more
    robust than a hardcoded translator.
    The $60,000 process of developing a language within a language has
    given me a look into PowerBASIC, let me tell you!   
    
    It has also produced some useful code that I will soon be releasing 
    as a commercial PB library called "MondoLib for PB."  MondoLib will 
    NOT be a huge Hanlin-like collection of routines.  Instead, it will 
    be about 20 HLL powerhorses, some which I've already released into 
    the public domain, others like LPM Lite and JAMPLAY, which plays 
    PLAY statements over an AdLib.  The library also has a fast 
    ARRAY_SCAN replacement, a math expression parser, a full ANSI 
    emulator, and an INSTR replacement that scans up to 5 times faster 
    than INSTR under certain conditions.  It will also have mouse 
    routines and EMS routines.  In short, anything I've had to develop 
    for Fermat will be in the library.  In some cases, I've had to 
    strip out some of the more powerful features, such as only having 
    LPM Lite rather than the full language.
    I've also been considering writing a book on advanced PB programming,
    but I don't know how big a base there is with PB, so I don't know
    that that would be profitable.  
    Now, how about you?  What have YOU been up to?  Do you still work
    for the TELIX people?  
 Khodafez,
 Jamshid Khoshrangi
___ Terminate 3.00/Pro
* TerMail/QWK * AhuraMazda(tm) Software [ams@stage.lionsgate.com]
--- Maximus/2 2.01wb
---------------