Re: alternating tags in a list?

Subject: Re: alternating tags in a list?
From: Paul Prescod <paul@xxxxxxxxxxx>
Date: Tue, 15 Dec 1998 09:21:52 -0600
Guy_Murphy@xxxxxxxxxx wrote:
> I agree whole heartedly. Extrapolating the arguement of we don't need to
> impliment that in XSL because you can do it with DOM elsewhere would
> quickly paint XSL into a corner considering that we can do everything that
> XSL will do with the DOM elsewhere. In which case why do we need XSL.

Extrapolating any argument beyond common sense takes you to a silly place.
That's why we must apply common sense. 

I can do the same thing with the opposite argument. Some formatting tasks
depend on data from a relational database so XSL must incorporate SQL.

> "We can do it with the DOM elsewhere" isn't isn't a valid arguement for not
> doing something in XSL. The point of XSL should surely be that it be an
> ideal language for transforming and rendering XML, if you point to
> something else to use instead then you've shot XSL in the head.

First, I'm not sure we need that particular metaphor, but I acknowledge
your right to use whatever one you want. Second, "there is a better place
in the information system to put this functionality" is indeed a good
argument. Third, XSL is not to be the ideal language for transforming XML:
it is to be a good language for transforming XML. XML is not the ideal
markup language, HTTP is not the ideal transmission protocol. The Web is
built on the idea that sometimes we constrain functionality in order to
either get something more widely implemented or increase the rigour of the
system: in this case we can do both.

> I start to get concerned with arguements of purity as I don't see issue
> such as these as either or... I tend to think why not both? Just because I
> won't use one doesn't mean you shouldn't.

I have watched in various places where a single extra feature in a
language can severely restrict the number of implementations and, more
worrisome, actually restrict the number of things that can be done with
that language. It takes a long time, looking closely at a single language
to recognize when this has occurred, which is why most people do not
notice it.

For instance, there will never be a popular XML DTD editor. That isn't
because schema languages are going to come along and cut off DTDs at the
knees. Schemas could never come along and there would STILL never be a
popular DTD editor. A single feature in XML prevents that from happening.

There will be very few XML editors that handle the full range of attribute
values in a reasonable way. I can predict that right now. Attributes seem
like a simple feature to implement, but there is one little corner of the
XML language that makes them actually pretty tricky.

There will never be a program that can take an arbitrary TeX document and
extract the words from it reliably. Never, ever, ever. Not just unlikely
-- impossible.

Because of one characteristic of EPS files, I cannot load them into most
vector drawing programs and manipulate them. Because of the same
characteristic, they often cannot be previewed in Word and moved reliably
between the Macintosh and Windows. In fact, this characteristic has caused
me no end of misery over the past two weeks.

If you want more information, the W3C domain lead of XML development at
the W3C wrote about this as far back as 1994:

 Paul Prescod

 XSL-List info and archive:

Current Thread