Path: utzoo!attcan!uunet!bu.edu!rpi!zaphod.mps.ohio-state.edu!wuarchive!cs.utexas.edu!sun-barr!rutgers!cmcl2!kramden.acf.nyu.edu!brnstnd
From: brnstnd@kramden.acf.nyu.edu (Dan Bernstein)
Newsgroups: comp.lang.misc
Subject: Re: Answers, Chapter 1: TeX
Message-ID: <2047:Nov607:10:1690@kramden.acf.nyu.edu>
Date: 6 Nov 90 07:10:16 GMT
References: <1990Oct29.191321.3202@maths.nott.ac.uk> <4349@lanl.gov>
Organization: IR
Lines: 24

In article <4349@lanl.gov> jlg@lanl.gov (Jim Giles) writes:
> It is because of this problem of inheriting data through several levels
> of the call chain that the test in the loader is most important.  The
> loader _can_ perform this test reliably and quickly.

It is inappropriate to give the loader sufficient knowledge of your
language to perform these tests. It is also rather stupid to delay
checks until load time, since some packages (such as libraries) may not
be linked until much later.

The .h-.c (package spec-package body) model is sufficient to detect
procedure call interface errors at compile time.

> I don't know any natural ways of using pointers.  Pointers are one of
> the most unnatural data structuring tools that I've ever encountered.

Knuth says that all data structures in the real world seem to be quite
appropriately modelled as a set of structures, some fields of which
contain pointers to other structures. There are many, many, many people
who agree with him that pointers are a natural way to express data
structures. I am not saying that you're wrong, but your statement is not
objective.

---Dan