Newsgroups: rec.arts.int-fiction
Path: nntp.gmd.de!news.ruhr-uni-bochum.de!news.rwth-aachen.de!newsserver.rrzn.uni-hannover.de!hrz-ws11.hrz.uni-kassel.de!news.th-darmstadt.de!nntp.zit.th-darmstadt.de!voskovec.radio.cz!news.cesnet.cz!01-newsfeed.univie.ac.at!swidir.switch.ch!in2p3.fr!univ-lyon1.fr!jussieu.fr!math.ohio-state.edu!uwm.edu!spool.mu.edu!newspump.sol.net!www.nntp.primenet.com!nntp.primenet.com!netcom.com!erkyrath
From: erkyrath@netcom.com (Andrew Plotkin)
Subject: Re: Hugo vs. TADS/Inform
Message-ID: <erkyrathDy1Gw0.27F@netcom.com>
Organization: NETCOM On-line Communication Services (408 261-4700 guest)
X-Newsreader: TIN [version 1.2 PL1]
References: <51i60m$li0@ecuador.earthlink.net> <51ihl6$2n6@newsbf02.news.aol.com> <51kvnr$k2n@argentina.earthlink.net> <51mp3t$ql3@news.dx.net> <Dy07y7.3Bq.0.bloor@torfree.net>
Date: Fri, 20 Sep 1996 16:09:36 GMT
Lines: 50
Sender: erkyrath@netcom.netcom.com

Kent Tessman (by723@torfree.net) wrote:
> 4.  Could the Hugo Compiler be modified to generated Z-code?

> Well, I guess.  But Graham Nelson would probably tell you that writing a 
> Z-code compiler (the guts, I mean) is anything but a trivial task.  For 
> the sake of efficiency, economy, and performance, the function of the 
> Hugo Engine and the syntax of the language are fairly closely tied. This 
> also begs the question:  why would you want to?  There are an awful lot 
> of reasons why you would _not_ want to run Z-code--Inform is one of the 
> good arguments in favor of it.

I posted a middlin-short article about why you would want to; you've 
probably read it already. And the limitations can be avoided.

> Which ultimately means, of course, that we're talking about languages, 
> anyway.  It doesn't really matter which language you're working in so 
> long as the finished product measurs up.  You don't know if your home 
> computer applications were written in C, Pascal, C++, assembler, or some 
> combination of these or others.  And likely you don't care, so long as 
> the finished product works for you.  Which means that what matters to 
> this debate--is it a debate?--is how well the language works for the 
> programmer.

I *strongly* disagree. The first principle of programming for real people
is: It is better for the programmer to suffer for an hour than for the
user to suffer for a minute. (And it is better for the compiler-writer to
suffer for a week than for the programmer to suffer for an hour. And so
on.) Langauge differences only matter when all other things are equal 
*for the user*. 

In your example, C, Pascal, C++, and assembler *are* all equal to the 
user -- because they all generate the same binary format, which the user 
runs in the same way on the same set of machines. This situation is so 
common that it's easy to ignore user-impact entirely in selecting a 
language. But it's a mistake. (I am just getting into Java, and so far I 
like it a lot more than C++. But that doesn't mean I can switch my next 
C++ project to Java without asking "What does this mean for the user?" At 
the moment, it means a hell of a lot. Maybe in three years every desktop 
machine will run Java, and then the answer will be different.)

Obviously, this is precisely why I brought up the question of compiling to
Z-machine format. 

--Z


-- 

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