Re: Nostradamus (was Re: FO. lists as tables)

[pandeng@xxxxxxxxxxxx (Steve Schafer)]

On Fri, 15 Oct 1999 23:45:41 +0100 (BST), you wrote:

>No. I see portability as meaning use of the same stylesheet with
>different systems. A notion of portability that means everyone is using
>the same system (baring dull issues like efficiency or host operating
>system) is not what I had in mind at all. 

But if I have to do all my work on a single system in order to ensure
consistent results, then where is the portability? I may as well use a
proprietary document composition application. My data is still held
hostage.

>the user interface is not much of an issue if you are just batch
>formatting FO's being churned out by an XSL engine, is it? You are
>not using the system to author anything, just reading it in and driving
>a printer or screen or voice synthesiser or whatever.

I used "user interface" as merely one example of the kinds of things
that can be modified without altering the results of the formatter.
There are many other things in the "meat" of the formatter that could
be changed and improved without affecting the end result.

>It is not an unspecified free-for-all

As far as I'm concerned, it's awfully close. I'll refer once again to
HTML as an example: How many thousands (millions?) of man-hours do you
think have been wasted by people doing nothing but testing and
modifying their HTML so that it comes out right in all the different
versions of all the different browsers? Unless XSL can provide a
_substantially_ higher level of "identicality" (and from the wording
in the current Requirements Summary I can't see that it will), people
will continue to have to do the same kinds of things. Only this time
it will be even worse, because they'll be working with XSLT instead of
XSL fo's themselves, and will therefore be one more step removed from
the end product and consequently have less control over the details of
the generated fo's.

>No tweaked, discarded. Are you really saying that (given the code)
>getting son-of-tex, son-of-word, son-of-frame, and daughter-of-quark
>to all produce identical line and page breaks on a given input would
>just be a matter of some `tweaks'.

In my code, the line breaker is a separate routine, independent of
everything else. It receives as input a line length in absolute units,
and an ordered list of objects representing glyphs, spaces, kerns,
discretionary breaks, and penalties (it is a minor variation of the
TeX algorithm). The only aspect of line-breaking that isn't in that
one routine is orphan/widow consideration. Yes, modifying that code
_would_ be a "tweak." Even if I had to replace it wholesale, it's just
one routine.

-Steve Schafer


 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list