Re: 2.6 patterns: let's try variations on the XML syntax

>Scott Lawton wrote:
>>
>> Please don't stop halfway.  Complaints about the syntax being verbose are
>> opportunities for articulating the benefits of XML.

Paul Prescod replied:
>I've been working with SGML and XML for several years now, and I hear this
>mantra repeated every so often. Nobody has yet agreed to take it to its
>limit, however.

First, let me say that I quite value your perspective (based on reading
various list archives over the past few months).  XML needs to address a
much wider audience than SGML; let's see if we can combine the wisdom of
the "old dogs" with the needs of the "new dogs" (if you'll pardon the
analogy).

So, in the spirit of constructive discussion:


>Should we abandon Java and other programming languages
>that do not use XML syntax?

Well, XSL did express programming constructs in XML; something that (to me)
is much more awkward than expressing a template match in XML.  I don't find
it all that readable (compared to traditional syntax) -- but I shrugged off
the extra effort since I agree with the goal of expressing in XML.

For example, compare

<xsl:define-macro name="numbered-block">
  <xsl:macro-arg name="format" default="1. "/>
  <xsl:number format="{arg(format)}"/>
  <fo:block/>
    <xsl:contents/>
  </fo:block>
</xsl:define-macro>

to one's favorite variation of:

define numbered-block(contents, format: "1. ")
  <xsl:number format="{arg(format)}"/>
  <fo:block/>
    {arg(contents)}
  </fo:block>
end

In addition to being recognizable as a macro/subroutine/function/procedure,
that also has the advantage of showing that "contents" is essentially a
parameter/argument.

Up next: invoking macros, for-each, choose/when, etc.

But, what's the goal?  To pick the best syntax for macros, programming
constructs and template match, or to express as much of XSL in XML as is
reasonable?


>XML already goes incredibly far in using
>XSL for everything. Pushing it into the pattern is a bad idea not only
>because of the verbosity, but because the verbosity and character set
>problems will prevent the language from being used in other contexts, such
>as in queries from attribute values in XML documents or in query
>languages meant to be typed on a command line or from a programming
>language (like SQL).

Quick reaction: if a separate working group is formed to create a query
language, it should define a set of requirements, outline its scope, and
tackle the problem.  Maybe that's a good idea; I don't know enough about
the big picture to comment.  But in the absence of that, I think it's best
to apply the considerable expertise of the working group (and others) to
come up a good XML syntax.

Scott



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