Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!usc!ucla-cs!maui!ryan
From: ryan@maui.cs.ucla.edu (Ryan R. Ramsey)
Newsgroups: comp.sys.mac.hypercard
Subject: Re: Has anyone tried ColorCard or CompileIt?
Message-ID: <26662@shemp.CS.UCLA.EDU>
Date: 23 Aug 89 18:26:07 GMT
References: <762woodl@yvax.byu.edu>
Sender: news@CS.UCLA.EDU
Reply-To: ryan@cs.ucla.edu (Ryan R. Ramsey)
Organization: UCLA Computer Science Department
Lines: 35

In article <762woodl@yvax.byu.edu> woodl@yvax.byu.edu writes:
>
>     I just recently purchased CompileIt, but haven't used it extensively yet.
>I did run their demo which is a HT function that does square roots.  You
>compile it.  It worked much as they claim (e.g., the function in HT takes 2-3
>seconds on SQRT(1000), whereas when it is compiled, it is almost instantaneous.
>It appears to have become an XFCN, but I didn't actually check with ResEdit.
>The original function is only about a dozen lines and it took 13 minutes to
>compile, so you will spend some time waiting.
>
>Larry Wood, Brigham Young University

I have TRIED to use CompileIT many times but it seems to be more of a hassle
than it is worth. The observations I have made are:

	1> It does not compile the stack but only blocks of code. If you have
	   a big stack with 100 subroutines, CompileIt will make 100 different
	   XCMDs.

	2> The XCMDS work very fast, but unless you compile everything, the  
	   difference in speed is hardly noticeable.

	3> CompileIT is slow. I tried to compile a 20 line subroutine and
	   it took almost 2 hours.

	4> CompileIT does not understand a lot of hypertalk structures thus
	   makeing many scripts uncompileable.


+----------------+----------------------------+
| Ryan R. Ramsey | UCLA Computer Science Dept.|
+----------------+----------------------------+
| ryan@CS.UCLA.EDU                            |
| ...!{uunet,rutgers,ucbvax}!cs.ucla.edu!ryan |
+---------------------------------------------+