Re: [xep-support] XEP Line Breaking Strategy

From: David Tolpin (
Date: Sat Nov 16 2002 - 14:08:01 PST

> I'm wondering if it's possible for XEP to fall back to the XSL behavior
> when the kerning goes negative (that is, the characters overprint each
> other)? That would provide the best of both implementations--the best
> result when just a little squeezing will fix it, a readable result when
> it won't.
> I realize that having this type of situation is fundamentally user
> error, but it's also the case that the XSL Formatter behavior does
> ensure that the result is always readable, even if it's not always what
> was desired.


1. It is possible to implement this behaviour.

2. I am not completely convinced it is a good behaviour, since it hides bugs from stylesheets'
creators. Any good hyphenation algorithm attempts to hyphenate as little as possible, thus
if one can accept hyphenation somewhere, it should be allowed there.

3. For me, the two alternatives are not 'to hyphenate or not to hyphenate', but whether to
truncate a word when it becomes unreadably compressed or let it compress infinitely
and show the user, through that infinite compression, that there is a problem in the stylesheet.

I think that a formatter should hide as few errors as possible. Since the simple hyphenation
approach described in XSL FO Recommendation allows to specify minimum length of a word eligible
for hyphenation, users are encouraged to use explicit hyphenation along with setting limits
to get the desired effect.

Since there is a clear way to say 'break only if a word is really long' I doubt there should
be a special mode to handle words which are too long. That would require additional heuristics inside
a formatter, while the standard-based approach is completely configurable by the formatter.

(*) To unsubscribe, send a message with words 'unsubscribe xep-support'
in the body of the message to from the address
you are subscribed from.
(*) By using the Service, you expressly agree to these Terms of Service

This archive was generated by hypermail 2.1.5 : Wed Dec 18 2002 - 08:41:28 PST