24 Apr
2006
24 Apr
'06
7:22 a.m.
On Mon, Apr 24, 2006 at 11:11:39 +0100, Damon Chaplin wrote:
On Sat, 2006-04-22 at 21:32 +0100, Steve Harris
wrote:
The current system is RDF/XML, not vanlla XML, you cant parse it usefully
with stock XML parsers.
XML is not very easy to extend without breaking other peoples tools.
On Sat, Apr 22, 2006 at 03:05:18PM -0400, Dave
Robillard wrote:
Speaking as a general developer, I'd much rather you just used plain
XML. (Pretty much as you have now, in fact.) nonono :) I think metadata outside the plugin is
without a doubt the
right way to go. I meant I'm just not a huge fan of the particular
syntax of this Turtle stuff (as opposed to normal well-formed XML).
Mostly because it means we need special tools and who knows what
libraries to deal with it.
you cant usefully read RDF/XML with just an XML parser anyway. It's quite
a lot of work to transform from the XML tree to the RDF graph.
But yes, Turtle support is less widespeard than RDF/XML, but there are
still Free/Open parsers for every language I can think of (C(++), perl,
java, python, ruby, javascript, etc.) Turtle isn't a standard, is aimed at much more
general problem areas,
and will just force people to read more docs and install more packages.
It might be a bit nicer for the person writing the plugin file, but that
isn't that much work. (Or does turtle include some essential feature?)
Turtle only missed being a standard months ago becasue of a wierd patent
problem with a well-known company that shall be nameless. It will be a
standard very soon.
Future changes should not have to break back-compatibility. We only need
to do it now to remove some old stuff that doesn't work very well.
- Steve