Newsgroups: comp.lang.apl
Path: watmath!watserv2.uwaterloo.ca!watserv1!70530.1226@compuserve.com
From: Mike Kent <70530.1226@CompuServe.COM>
Subject: Re:  Kaizen etc.
Message-ID: <921029061855_70530.1226_DHP28-1@CompuServe.COM>
Sender: daemon@watserv1.uwaterloo.ca
Organization: University of Waterloo
Date: Thu, 29 Oct 1992 06:18:56 GMT

 A couple of things Dick Holt didn't mention: 

 [1]   The 386-based APL*PLUS printer support includes a PostScript      
       driver and a serviceable Adobe type 3 font (in the commercial    
       version).

 [2]   Dyadic Systems "Dyalog APL" product line.  As far as I know,
       there's no inexpensive version -- the commercial product is
       priced comparably to the commercial APL*PLUS products.  I don't
       have any experience doing commercial development with these 
       products, but have seen some of the results that have been 
       acchieved by people who do (at two large NYC finacial firms),
       and the overall quality of the system seems to be on a par with
       APL*PLUS and APL2.  But there is one place where the Dyalog
       systems *clearly* outpace the Manugistics and IBM current
       offerings:  integration with host windowing environments.
       For PC/DOS platforms, Dyalog/W is a true MS Windows program
       and has a Visual-BASIC style interface builder WS; for UN*X
       platforms, Dyalog/X does XWindows very nicely.  If your
       appplication requirements include a spiffy, modern, GUI-based
       user interface, this may be the interpreter line for you.
       The documentation is fine for someone who knows APL and wants
       to learn about this implementation, but (IMHO) would be a bit
       of a struggle for a newcomer to APL (not impossible, but 
       a serious effort).
  
 [3]   IBM's documentation for APL2/PC is excellent; it includes:
	
	 --  the APL2 language reference, perhaps the best programming-
             language documentation IBM has ever produced.  THE definitive
             source for how nested-array APLs behave (at least until the
             ISO Extended-APL standard becomes official).

         --  an APL2/PC-specific reference documenting the differences 
	     between the mainframe version (what the APL2 Language
	     Reference documents), and the additional facilities which
             don't exist on the mainframe; thorough and comprehensive
	     if not quite up to the Language Reference standards.

	 --  "APL2 at a Glance" which is a fine and extensive            
             introduction and tutorial.  Less depth than the Pakin &
	     Polivka or the Gilman & Rose classics on "flat" APL,
             but there's always the Languge Reference (again).

 [4]   The "editor 3" editor supplied with APL2 is almost pretty good,
       but (IMHO) consistently falls just a little short of what
       you really want.  On the other hand, you can use ANY editor that
       doesn't muck around with your video adapter -- I use the T editor
       which is the basis for ED3 (free via IBM's Employee Written S/W  
       program) and I've seen KEDIT used (with Personal REXX macros
       for stuff like keeping labels in L1, L2, ... order and keeping
       the local-variables list alphabetized, no less), so this is not
       a terrible problem.  Unfortunately (as Dick mentioned) none of
       these facilities are available in the TryAPL2 package.

 [5]   True component files are indeed available for APL2/PC but are
       not part of IBM's package (though APL2/PC has support for files
       you can keep arrays in, you have to write some cover functions
       to make them behave like component files, and the emulation has 
       some limitations).  For $175, though, you can get a full-featured
       and solid implementation of component files from Interprocess    
       Systems.  Now if they'd just port their IEDIT mainframe editor  
       and debugging environment ... and the subset flat-APL to FORTRAN
       translator ...

 [6]   What about J (not that J hasn't been discussed from time to time
       here, but it's cheap, APL-like [though the notion that J is a   
       "new dialect of APL" strikes me as about as accurate as the notion
       that Italian is "a new dialect of Latin"], and learnable)?  And 
       what about NIAL?


       Mike Kent                              70530.1226@CompuServe.COM


