Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site looking.UUCP
Path: utzoo!watmath!looking!brad
From: brad@looking.UUCP (Brad Templeton)
Newsgroups: net.lang.pascal
Subject: Re: The case for Borland's Turbo Pa
Message-ID: <693@looking.UUCP>
Date: Thu, 23-Oct-86 12:41:31 EDT
Article-I.D.: looking.693
Posted: Thu Oct 23 12:41:31 1986
Date-Received: Fri, 24-Oct-86 02:04:39 EDT
References: <691@looking.UUCP> <72100001@prism>
Reply-To: brad@looking.UUCP (Brad Templeton)
Organization: Looking Glass Software Ltd. Waterloo, Ontario
Lines: 33
Summary: 

In article <72100001@prism> brian@prism.UUCP writes:
>
>
>Oh, and if you think that other pascal compilers are comparable to turbo
>as far as compile speed if you compile from disk, I think you are wrong.

That's because Turbo Pascal
	doesn't have to load in the compiler (or multiple passes of it).

	doesn't have to write out the object code. (or does it in a single
	  block write)

	loads in the source with a single multi-block read into ram.
	  (You would be surprised how much faster this can be than a tpical
	  buffered getc)

	Has no link phase or libraries. (one of the biggest wins, but one of the
	  biggest costs.

Now you may say, "this cleverness is exactly what's right about TP," and you
are probably right.  My point was that the compiler itself is not
supernaturally fast.  Anybody could have done the same with their compiler.
If they had realized the market, they would have.
>
>That is what the development process is sometimes all about -- trying new
>stuff as quickly as possible.  Then use microsoft's compiler to generate the
>production code (if you think it is really necessary).

If your program is not CPU bound itself, use an interpreter.  That lets you
try stuff even faster and gives you superior debugging.
>Brian K. Moran  brian@mirror.TMC.COM	
-- 
Brad Templeton, Looking Glass Software Ltd. - Waterloo, Ontario 519/884-7473