elwell@osu-eddie.UUCP (Clayton M. Elwell) (11/15/85)
Note: I have no connection with Mystic Canyon Software. This review is based
solely on my impressions gained by using the product.
Now that that's over with, on with the review.
A week or so ago, someone posted a message to net.micro.pc that claimed, in
effect, that Mystic Pascal was a scam. This had me a little worried,
because I had just ordered it. Well, it arrived, two weeks to the day
after I mailed my order.
I opened it to find a manual and a disk containing Mystic Pascal Version 1.14.
This is a pre-release; the first production version will be 1.5, which they
claim will be out "in a few weeks". There are some features missing, but
since they were honest about it I will not include them in this review. When
I receive version 1.5 I will post a revised review.
The manual is complete but minimal. I hope they flesh it out with more
examples and (for professional software developers like me) some more
technical details, but it does explain everything you need to know about
using the system. It is 74 pages, most of that devoted to standard
procedures and functions (1 per page).
If you've seen the advertisements, you may have been skeptical of whether
any product could do what they claim for $64.00. The answer is yes.
Someone has put a LOT of work into it. Operation is extremely smooth. With
a few days more practice, I expect it to be effortless.
Statistics: ISO Standard Pascal Level 0 (Version 1.5)
Compiles to 8086 machine code
Two levels of optimization
Full-screen Editor (WordStar/Turbo Style)
Immediate execution mode
Help windows on ISO Pascal, Editor, & Extensions
Multi-tasking support with named message queues and
arbitrary messages
Program up to 1Mb, Data up to 1Mb
Compiler/Editor Size: 67K
Help file Size: 25K
One word describes this system in general: SPEED.
Mystic Pascal is an incremental compiler. You change five lines in a 2000
line program, it only recompiles those lines. It compiles the 7K sample
program supplied with the package in 0.03 seconds. That is not a typo.
247 lines in 0.03 seconds. This evening I will benchmark execution
speed and code size against MS-PASCAL and Turbo Pascal and post the results.
The product has an extremely professional feel. Even as a prerelease, it
is well worth the $64.00 purchase price just as a tutorial on using
multi-tasking with message passing and as an example of the Right Way to
write a commercial program for the IBM PC. If you develop software for
the PC, you should buy this program, even if you hate Pascal (I did until
now...).
Stay tuned for benchmarks...
--
-- Clayton Elwell
Elwell@Ohio-State.CSNET
Elwell%Ohio-State@CSNET-RELAY.ARPA
...!cbosgd!osu-eddie!elwell
-----------------
"Roads? Where we're going, we don't need roads..."mikes@3comvax.UUCP (Mike Shannon) (11/20/85)
Clayton Elwell in the cited article writes: > One word describes this system in general: SPEED. > > Mystic Pascal is an incremental compiler. You change five lines in a 2000 > line program, it only recompiles those lines. It compiles the 7K sample > program supplied with the package in 0.03 seconds. That is not a typo. > 247 lines in 0.03 seconds. This evening I will benchmark execution > speed and code size against MS-PASCAL and Turbo Pascal and post the results. ------- Uh, the way I figure it, .03sec/7kbytes comes out to about 4 and a half microseconds per character. This amounts to only a couple of machine instructions per byte. A program can't even decide if the character is an 'a' or not in this amount of time, much less do some real compiling. Maybe you're only off by a couple of orders of magnitude. Who's kidding who (or is it whom?)? Eagerly awaiting more benchmark humor, -- Michael Shannon {ihnp4,hplabs}!oliveb!3comvax!mikes
larry@geowhiz.UUCP (Larry McVoy) (11/23/85)
In article <286@3comvax.UUCP> mikes@3comvax.UUCP (Mike Shannon) writes: >Clayton Elwell in the cited article writes: >> Mystic Pascal is an incremental compiler. You change five lines in a 2000 >> line program, it only recompiles those lines. It compiles the 7K sample >> program supplied with the package in 0.03 seconds. That is not a typo. >> 247 lines in 0.03 seconds. This evening I will benchmark execution >> speed and code size against MS-PASCAL and Turbo Pascal and post the results. >------- > Uh, the way I figure it, .03sec/7kbytes comes out to about 4 and >a half microseconds per character. This amounts to only a couple of >machine instructions per byte. Well, you're both right. I think how works is this, it compiles the code when you load the program. So, with turbo, you get fast load, slow compile, and the other way around with Mystic. Have this guy give times for: START C> turbo sample.pas C (i think you type C for compile) [line #'s flashing by] Done. STOP START C> mystic sample.pas C (or whatever) Done. STOP I think you'll find them comparable. -- Larry McVoy +----------------+ | Slower traffic | Arpa: mcvoy@rsch.wisc.edu | keep right | Uucp: {seismo, ihnp4}!uwvax!geowhiz!geophiz!larry +----------------+
elwell@osu-eddie.UUCP (Clayton M. Elwell) (11/25/85)
The figure I quoted is the time between me hitting the keystroke for "compile" and the time it's done. This is what I call 'compile time,' not the time from when it "starts parsing". I don't much care about how many machine cycles it takes, it doesn't take much *real* time... --Clayton -- -- Clayton Elwell Elwell@Ohio-State.CSNET Elwell%Ohio-State@CSNET-RELAY.ARPA ...!cbosgd!osu-eddie!elwell ----------------- "Roads? Where we're going, we don't need roads..."