Re: xmlTree

[Ian Davis <id@xxxxxxxx.xxxx]

Hi James,

On Wednesday, July 21, 1999, 8:59:41 PM, james wrote:

> From: "james@xmlTree.com" <james@xmltree.com>

> Dear Syndication List members,

> I have opened up xmlTree's weblogs and news channels to the outside world in a machine
> readable format (OCS) at http://www.xmltree.com/export/ocs.cfm.  This is a complete list -
> some of the channels listed were entered by me after contacting channel content providers,
> some by the providers themselves, and some from My.Userland and RssMaker (both after
> contacting the respective owners).

Firstly, this is a great use of OCS and a real world application liek
this can only help to shape the specification as we start to use it
in anger. We now have one client (Headline Viewer) and one server
application. Great!

I'd also like to say that I now consider the XML Tree OCS file to be a
superset of the RSSMaker OCS file, so if you're writing a channel
viewer you should use the list at XML Tree rather than parse two
files. I will ensure that the XML Tree listing contains an
up to date mirror of the RSSMaker listing.

> What does xmlTree add?  Only keywords and categorisation (using Dewey).  A sample entry
> looks like:

> <rdf:description about="http://www.betanews.com/";>
>     <dc:title>Beta News</dc:title>
>     <dc:creator>BetaNews</dc:creator>
>     <dc:description>Beta News</dc:description>
>     <dc:subject>
>         <rdf:Description>
>             <dcq:subjectScheme>DDC Local</dcq:subjectScheme>
>             <rdf:value>005.1</rdf:value>
>         </rdf:Description>
>     </dc:subject>
>     <dc:subject>beta, betanews, daily, e-zine, headlines, news, software,
> world</dc:subject>
>     <rdf:description about="http://betanews.com/mnn.php3";>
>         <ocs:language>en</ocs:language>
>         <ocs:format>http://my.netscape.com/rdf/simple/0.9/</ocs:format>
>     </rdf:description>
> </rdf:description>


I like the use of the Dublic Core Qualifiers to denote the Dewey
scheme. If anyone needs more information see
http://www.ukoln.ac.uk/metadata/resources/dc/datamodel/WD-dc-rdf/WD-dc-rdf-19990701.html,
specifically example 9. I'd like to add the scheme to OCS v0.4. Any
other feedback would be appreciated.

> There was some negative feedback about using Dewey on Dave Winer's discuss.userland.com
> site at http://discuss.userland.com/msgReader$8578,
> but I think that the alternative suggested (Library of Congress) does not offer as
> intuitive a scheme.  I'm agnostic about my selection of categorisation, however, and will
> change if someone can show that an alternative is as rigorous, updated frequently so that
> new topics are better covered, is extensible (so that xxx is a superset of xxxx) etc etc.
> I agree that this type of categorisation should not be the only one, which is why I also
> include keywords (see http://www.xmltree.com/metadata/search.cfm) for an example keyword
> search tool.
If I had to make a choice of a formal categorisation scheme it would
definately be Dewey, but I have big reservations about it's
applicability in this arena. I think that one of the comments on
userland was that Dewey was way out of date and that topics will get
skewed towards the general computing section. The changes to XMLTree
have changed my mind a little on this - there seems to be a fairly
even spread of topics (although there's a big clump around
Shakespeare!)

I like keywords because they're easy and anyone can classify by
keywords - Dewey takes time and skill to classify with. A mix of the
two schemes seems a good compromise.



.id.

-- 
weblog - http://alchemy.openjava.org/
me - http://www.fdc.co.uk/people/iand/ 
email - iand@fdc.co.uk | icq - 4423828