Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!ames!lll-winken!uunet!portal!cup.portal.com!mmm
From: mmm@cup.portal.com (Mark Robert Thorson)
Newsgroups: comp.editors
Subject: Re: Editing a Macintosh File (never the same way twice)
Message-ID: <17716@cup.portal.com>
Date: 28 Apr 89 22:44:16 GMT
References: <17605@cup.portal.com> <39727@think.UUCP>
Organization: The Portal System (TM)
Lines: 75

> Is it Apple's fault that your word processor forces you to solve these
> problems using manual overrides?  A good text formatter should allow
> you to specify minimum widow sizes and control hyphenation.

That depends on the level of quality the user is striving for.  In the book
design I am using, figure references, such as "see Figure 10" have the
"Figure 10" part in bold face.  To make things look nice, I want to keep this
part from being split across a line break.  If my editor anticipated this,
it would be doing wrong, because in other book designs it might be preferable
to allow the split.

To say that an editor should be able to automatically handle all formatting
decisions is like saying everybody should write programs purely in high-level
languages without GOTO's.

> So, decisions about whether changes should be made should be based
> upon whether Mark Robert Thorson can tell the difference?  There's a
> whole science of typography and font design, which spends much of its
> time worrying about hundredth-of-an-inch differences in the lines that
> makes up printed characters, so obviously some people can.  I think
> Apple should be commended for caring enough.

I am well aware of typography and book design issues.  You are correct
in pointing out that some of these issues concern 1/100th of an inch of
white space.  It is precisely these kinds of tweaks which Apple changes
in every OS release, which screws everything up.  The document I just 
finished expanded by about 1% because of their tweaks.

I carefully examined the tweaks they made in their previous revision of the
OS by printing the same file under the old and new system, then comparing
them on a light table.  The changes they made were completely imperceptible
without this kind of comparison.

The question isn't so much whether the tweaks are needed or not, but why
the ******* tweaks can't remain stable.  Why weren't the tweaks in the last
release okay?  Or the release before that?  You can bet dollars to doughnuts
that they will tweak 'em again in the next release and the one after that.

If a spy took a job with Apple for the purpose of wrecking the company,
this would be a most clever, subtle, and effective way to do it.

> So they should never fix bugs, in case some users happen to be relying
> on the incorrect behavior?  What about the people who DO care about
> the precise spacing, and complained to Apple that their documents were
> coming out wrong?  Unfortunately, someone has to lose.  And it's
> usually the person relying on the bug.

I would be astounded if Apple received even one request from a user asking
for the tweaks they have put in.  I am more inclined to believe that Apple
typographers are trying to justify their reason for existence (at Apple).

> Suppose Apple had a compiler that incorrectly compiled multiplication
> statements, always producing a result one higher than the correct
> result (don't laugh -- early versions of a popular microprocessor chip
> had a bug in a multiply instruction).  A user of this compiler might
> write
> 
> 	A = B * C - 1;
> 
> in order to multiply B and C.  Then a new version comes out, which
> fixes the bug.  Should the user complain that this broke his programs?
> If so, what is Apple supposed to do?

A microprocessor manufacturer who has this kind of bug should introduce a new
opcode for the corrected form of the instruction.  There is no good reason
to screw the existing customer base IF there is any way to avoid it.

> 
> Barry Margolin
> Thinking Machines Corp.
> 
> barmar@think.com
> {uunet,harvard}!think!barmar

Mark Thorson, somewhere in California.