Re: [xtm-wg] rdf & topic maps

["Bernard Vatant" <universimmedia@wanadoo.fr>]

Bryan, Michel

As a former editor at Open Directory, I think I have a few things to say
about that.
When I left ODP last year, I thought also about other mapping(s) of the
Directory.
(it was before I knew the very existence of Topic Maps concept at all !)
Now I think such projects have important shortcomings :

1. ODP uses, as far as I know, some version of RDF syntax which is not the
last one. Am I correct ?
2. ODP RDF dump is a huge, not to say monstruous, thing, which will be more
and more heavy to handle, and that you must take as a whole : there is no
way to use only a part of it, by the very structure of links.
3. ODP is a very unstable structure, with a chaotic behaviour.
Reorganizations are bound to occur at any moment at any scale : since no
single-hierarchy ontology is able to represent the complexity of the Web it
is supposed to map, at any moment, the existing structure at any hierarchy
level is the image of the present view of the world of the editor(s) having
power at this level. These human parameters are bound to provoke
imprevisible catastrophic changes.
At the bottom of this issue in the very incapacity of a basically
single-hierarchy structure to represent complexity. Try to find out the
representation of  some complex knowledge fields in ODP, like environment,
sustainable development, fair trade ... and you'll find it's completely
exploded in many distinct, overlapping and/or redundant categories, often
ignoring each other, with no way to have the synthetic vision Topic Maps
can achieve. Attempts to solve this problem are the main cause of permanent
reorganizations, every new editor thinking he/she has THE solution, and
will apply it as soon as he/she has the power to do so. (ODP is really a
very exciting power game, just try it if you can spare the time, it's more
thrilling and dangerous than Quake 3). That's the main reason, among
others, why I left ODP. And I don't see how a TM representation of ODP can
make more than reproduce or show evidence of this fundamental shortcoming.
4. ... last but not least : the links in ODP are not "translatable" IMO in
equivalent TM associations, except the class/subclass relationship. I mean
by that cross-links (non-hierarchical) between categories : "see also"
"related category" are left to the local editor's good or bad will, and
given the loose collaborative spirit going on, despite fine formal
guidelines about it, many cross-links are one-way only. It seems that some
progress have been made on that point lately anyway.
Moreover, these cross-links have no semantic definition, which means the
roles of two categories in the association representing a cross-link
(provided it's a two-ways link) are only defined in the editor's mind that
made the link ... and it's non-adressable :))

[For information of the list] There are presently two ways to use ODP :
either download/update RDF every week/month/ ... anything you can afford to
manage - 2,362,178 sites - 34,199 editors - 341,152 categories are present
figures - or to use it "live", which some software do already, like
anaconda www.anaconda.net

All that being said : ODP "chaotic ontology" is de facto a reference, given
it's the biggest web directory, and its structure can be used for free - if
you can afford managing it. It should be fine to show that XTM can handle
that, show its shortcomings ... and eventually propose a new directory
paradigm, keeping the fine "humans do it better" principle, but enabling an
effective management of complexity.

Good luck Bryan :))
Keep in touch, I'm very interested anyway in whatever you can get to with
that project.


----- Message d'origine -----
De : Bryan Thompson <bryan@globalwisdom.org>
À : Michel Biezunski <mb@infoloom.com>
Cc : Jason Markos <jsm@empolis.co.uk>; <xtm-wg@yahoogroups.com>
Envoyé : vendredi 9 février 2001 05:37
Objet : [xtm-wg] rdf & topic maps


> Michel,
>
> I plan to publish a translator from the open directory, which is exported
> as RDF, to XTM 1.0.  My intention was to do this as part of the
> topicmaps.org
> site.  I see such code examples are a necessary aspect of marketing XTM.
>
> Is there a particular reason why you are compiling such things for the
> infoloom
> site?  I've been talking some with Jason about XTM marketing and it seems
> that
> the "marketing group" should draft a soliciation of examples, code, etc.
for
> topicmaps.org.
>
> Your thoughts?
>
> -bryan
>
>
>
> To Post a message, send it to:   xtm-wg@eGroups.com
>
> To Unsubscribe, send a blank message to: xtm-wg-unsubscribe@eGroups.com
>


------------------------ Yahoo! Groups Sponsor ---------------------~-~>
eGroups is now Yahoo! Groups
Click here for more details
http://click.egroups.com/1/11231/0/_/337252/_/981715850/
---------------------------------------------------------------------_->

To Post a message, send it to:   xtm-wg@eGroups.com

To Unsubscribe, send a blank message to: xtm-wg-unsubscribe@eGroups.com