Newsgroups: rec.arts.int-fiction
Path: nntp.gmd.de!news.ruhr-uni-bochum.de!news.rhrz.uni-bonn.de!RRZ.Uni-Koeln.DE!news-koe1.dfn.de!news-kar1.dfn.de!news.nacamar.de!www.nntp.primenet.com!nntp.primenet.com!netcom.com!erkyrath
From: erkyrath@netcom.com (Andrew Plotkin)
Subject: Re: Good Language/Bad Compiler
Message-ID: <erkyrathE2Bq2J.Czt@netcom.com>
Organization: NETCOM On-line Communication Services (408 261-4700 guest)
X-Newsreader: TIN [version 1.2 PL1]
References: <32AFE4EA.147F@geocities.com>
Date: Thu, 12 Dec 1996 23:18:19 GMT
Lines: 33
Sender: erkyrath@netcom13.netcom.com

Bif (bifurcate@geocities.com) wrote:
> First, let me say that Inform is a great language. It's innovative and
> extremely useful. The documentation is very good; I like the writing
> style. The author also wrote two big games, which is a morale booster
> for those of us slogging through our programming.

> Secondly, let me say that the compiler sucks. It prints out five times
> as many error messages as there are errors. It doesn't handle nested ifs
> right.

Sometimes and recently. :)

> Can you change the program so that it prints only one error message and
> halts?

Gack, no no no. No compiler should ever do this (and I don't know any 
compiler that does.) It is often possible for the compiler to recover 
from discovered errors, which means that it can *correctly* find errors 
later in the program. I would much rather ignore a lot of spurious errors 
than have to recompile six times to find six real errors.

> And I think else's should always match the last if that doesn't
> already have its own else; this makes some complicated routines more
> readable (less unreadable).

Inform 6 does this. 

--Z

-- 

"And Aholibamah bare Jeush, and Jaalam, and Korah: these were the
borogoves..."
