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

[pandeng@xxxxxxxxxxxx (Steve Schafer)]

On Sat, 16 Oct 1999 17:14:18 -0700, you wrote:

>That's not portability.  It's what TeX calls "device independence".

I'm sorry, but that sounds like saying, "That's not an apple, it's a
Golden Delicious." What I'm talking about isn't the _only_ kind of
portability by any means, but it is most certainly _one_ kind of
portability.

In any case, I'm not talking about device independence (which is a
problem that I believe has been solved reasonably well). I'm talking
about implementation independence. If I have to use a single XSL
processor implementation in order to ensure consistent results, then I
have gained nothing by using XSL instead of some proprietary format.

>The purpose of typography is to best communicate the information
>contained in content.  When paper was the only medium, designing the
>best possible representation in a fixed way was the best way to
>communicate the information.  But now that information can be
>presented to myriad users in myriad ways, the best way to communicate
>the information is to describe a series of optimal constraints, not
>to focus on the best possible picture of the information.

I'm not suggesting that detailed control over the formatting process
is appropriate for _every_ occasion. But it _is_ appropriate for some
occasions, and I shouldn't be tied to one vendor's product in order to
achieve it. If what you said in your paragraph above were really true,
then XSL doesn't even need to support arbitrary fonts. Just a handful
of generic typefaces would suffice. Of course, if you did that, many
potential users would be up in arms.

>You chose your page breaks for a reason.  Encapsulate those reasons in the
>stylesheet, not the breaks.  An ideal stylesheet language will include
>widow and orphan control, weighted keep-with values, different rules for
>ending recto and verso pages, etc.

With an acceptably powerful formatting language, I would be able to
handle _most_ formatting rules in that language without "manual"
intervention. But I have yet to come across an even moderately complex
document in which _no_ manual tweaking was necessary. Of course, I
could write more rules to handle those special cases, but that
degenerates into a rather silly exercise if I have to write a separate
rule for each individual case.

>Everything.  If you send me a picture of an 8-1/2x11 page with 9-pt type,
>and I'm using a 640x480 monitor, I have to enlarge the page and then scroll
>left and right for every line.  It gets real tired real fast.  Give me the
>constraints you used to create that page, and let me apply them to my
>narrower page and bigger font, and I'll be very happy.

If I send you a document, meant for online viewing, and I didn't take
into account the characteristics of the viewing medium, then I should
be tarred and feathered, no question about it. It would be entirely my
fault. But to forbid my access to powerful tools because I _might_
make such a mistake is just as egregious an error.

-Steve Schafer


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