zhou@brazil.psych.purdue.edu (Albert Zhou) (01/05/91)
I am intending to expand the current matrix operation interface I've designed
into a full scale language on PC. Here is what I am considering:
(1) I don't want it to become a special language like mathematica,
gauss, imsl and so on. Instead, I want it to be a clone of a popular
language on PC with full matrix operation capacity. Since it has to be
interpretive, the the ideal candidate would be BASIC. I would call it
MATRIX BASIC.
(2) Besides matrix operations, it will contain also other frequently
used math functions like normal distribution.
(3) Programs written in MATRIX will be made a stand-along .EXE files.
between interpretive BASIC and compiled BASIC. It is not a compiled
program; instead, it is a combination of the interpreter and the
preprocessed MATRIX BASIC code.
Any comments, suggestions or objections? I would especially want to know
how big the market for MATRIX BASIC will be. Also, if anyone would suggest
another language to clone, I am listening.
M
[I'd expect the market to be close to zero. Matrix lovers have a wide variety
of inexpensive possibilities already ranging from MATLAB, both the old free
and newer cheap versions, low-priced versions of APL, and all sorts of C
and Fortran library packages. -John]
--
Send compilers articles to compilers@iecc.cambridge.ma.us or
{ima | spdcc | world}!iecc!compilers. Meta-mail to compilers-request.corbett@road.Eng.Sun.COM (Robert Corbett) (01/05/91)
In article <11651@j.cc.purdue.edu> zhou@brazil.psych.purdue.edu (Albert Zhou) writes: >I am intending to expand the current matrix operation interface I've designed >into a full scale language on PC. Here is what I am considering: > (1) I don't want it to become a special language like mathematica, >gauss, imsl and so on. Instead, I want it to be a clone of a popular >language on PC with full matrix operation capacity. Since it has to be >interpretive, the the ideal candidate would be BASIC. I would call it >MATRIX BASIC. ANSI BASIC, X3.113-1987, provides matrix operations including addition, subtraction, multiplication, inverse, determinant, and scalar product. I presume that BASIC implementations that claim conformance with the standard, such as True BASIC, implement those features (they are not optional). If your package only includes simple matrix operations, I doubt you will find much of a market. If your package includes support for complex matrices and analytic functions over them (not at all easy to implement), your sales might increase by dozens. Yours truly, Bob Corbett -- Send compilers articles to compilers@iecc.cambridge.ma.us or {ima | spdcc | world}!iecc!compilers. Meta-mail to compilers-request.
zhou@gatech.edu (Albert Zhou) (01/06/91)
In article <5326@exodus.Eng.Sun.COM> corbett@road.Eng.Sun.COM (Robert Corbett) writes: >ANSI BASIC, X3.113-1987, provides matrix operations including addition, >subtraction, multiplication, inverse, determinant, and scalar product. I >presume that BASIC implementations that claim conformance with the standard, >such as True BASIC, implement those features (they are not optional). So how is this BASIC doing these days? Does it have a lot of market and is it sold separately by vendors? >If your package only includes simple matrix operations, I doubt you will find >much of a market. If your package includes support for complex matrices and >analytic functions over them (not at all easy to implement), your sales might >increase by dozens. MATRIX BASIC of course will be more powerful than that. It will introduce a few new operators, such as ~ (horizontal concatenation), | (vertical concatenation), ` (transpose), \ (elementwise division), # (elementwise multiplication, ... Also, the original package has been designed for statistic processing. So MATRIX BASIC could have many built-in functions such as sum(x), ss(x) (sum of square), mean(x), var(x) (variance), ... Unfortunately we haven't implemented complex matrix since there hasn't been such need. [True Basic is a commercial package whose authors include the original authors of Dartmouth Basic. I don't know how it's doing, but observe that I don't see anything about it in the trade press. -John] -- Send compilers articles to compilers@iecc.cambridge.ma.us or {ima | spdcc | world}!iecc!compilers. Meta-mail to compilers-request.
khb@Eng.Sun.COM (Keith Bierman fpgroup) (01/08/91)
In article <11651@j.cc.purdue.edu> zhou@brazil.psych.purdue.edu (Albert Zhou) writes: > I am intending to expand the current matrix operation interface I've > designed into a full scale language on PC. Here is what I am considering: such features were quite common on the mainframe BASICs of yesteryear (e.g. an MBASIC on the JPL UNIVAC, back in '77). I don't recall these dialects taking over the world; though my HP 71b does support matrix operations. As noted by Corbett, the current ANSI standard for BASIC includes such features. -- Keith H. Bierman kbierman@Eng.Sun.COM | khb@chiba.Eng.Sun.COM SMI 2550 Garcia 12-33 | (415 336 2648) Mountain View, CA 94043 [So can you get any kind of Basic on a Sun? -John] -- Send compilers articles to compilers@iecc.cambridge.ma.us or {ima | spdcc | world}!iecc!compilers. Meta-mail to compilers-request.
bliss@sp64.csrd.uiuc.edu (Brian Bliss) (01/10/91)
|> and all sorts of C and Fortran library packages. -John]
They're already there, and highly optimized, why not use them? All you
need is an interpreter for a C or Fortran-like language that uses the same
calling convention as compiled C or Fortran, and use it to call the
packages. Add type checking and automatically include the appropriate
array bounds to the packages to make things bulletproof.
bb
--
Send compilers articles to compilers@iecc.cambridge.ma.us or
{ima | spdcc | world}!iecc!compilers. Meta-mail to compilers-request.