James Fuller writes:

> On Nov 28, 2007 8:46 PM, chromatic <[EMAIL PROTECTED]> wrote:
> > On Wednesday 28 November 2007 10:59:30 James Fuller wrote:
> >
> > > I do not nec. agree with 'a particular grammer is not' part of the
> > > core ... if that grammar is so common to every problem (like regex
> > > is) then why not include it?
> >
> > The criterion for including a module in the core is "Is this
> > necessary to get and install other modules?" not "Why not include
> > this module?"
> feels like we are externalizing what I would call build artifacts of a
> language .... e.g. a distro of Perl 6 should be easy to adopt and easy
> to use immediately

chromatic said "the core"; you said "a distro of Perl 6".  They are not
the same thing.

> ..... I would like to see some basic level of XML support in this
> distro.

I'm sure you would.

Personally I would like to see all the modules I often use in the
distro.  It's very rare for me to use XML, but there are lots of other
things I do use often.  How about we make the core the union of the
modules you want and those that I want?  Or should we let anybody else
in this game too.

> I understand that there can be different distros customized to certain
> problem domains, but as explained I see XML as common to all those
> problem domains.

Currently you use XML all over the place.  I can believe that.  But for
your argument to make sense you'd need to show that:

* XML is widespread.
* XML will continue to be widespread for the life of Perl 6.
* Whatever XML module we design now will be the best XML module for
  people to use throughout the desing of Perl 6, without a better one
  coming along.

Perl 5 has lots of modules in its core.  Some of them, with hindsight,
might not be the best possible choices:

* Text::Soundex -- how much is that used?
* CGI -- is that really the best CGI-parsing or HTML-generating module
  still?  (That so many others have been written suggests not.)

So, to make a claim for any 'domain-specific' functionality to be added
to the core, you would have to be better at making predictions than
those who added modules to Perl 5.  _Are_ you better at such
predictions?  What evidence have you got for that?

When XHTML1 launched did you correctly predict that XHTML2 would turn
into an ignored project that no web-browser vendors are interested in
implementing, and that instead they would be implementing HTML5, a
language based on HTML4 that encourages authors not to bother with XML?

If you didn't predict that, how does it fit in with your claim that you
"can only see more XML for all of us"?


Reply via email to