Li,
it would be not difficult to define a syntax and rules so that exclusive
ownership (standardization model) and
shared ownership (social wikipedia model) can exist simultaneously. At this the
standardization model can be
designed very efficiently (decentrally defined templates for human readable
representation in many languages
and metrics for similarity search are at once accessible, cf. my fist email of
yesterday).
These advantages would be also possible for the social model, because shared
ownership could be integrated
into the standardization model if only one socially good organized domain name
owner (e.g. wikipedia.org)
allows this.
Of course this is only possible, if there is (official) support for a concrete
design variant of the
standardization model. If there are questions concerning the concrete design in
detail, please don't hesitate
to ask me.
Wolfgang
----- Original Message -----
From: "Li Ding" <[email protected]>
To: "Wolfgang Orthuber" <[email protected]>
Cc: "Kingsley Idehen" <[email protected]>; "Linked Data community"
<[email protected]>;
<[email protected]>
Sent: Monday, May 18, 2009 6:57 PM
Subject: Re: Owning URIs (Was: Yet Another LOD cloud browser)
If there is no explicit official standard and recommendation for such global
task sharing, there is the
danger that those who need a special vocabulary develop many incompatible
standards for exchange of machine
readable data.
The earlier such a standard for well defined task sharing is recommended by
W3C, the easier it can be
introduced.
well, we can learn from the past:
(i) biomedical terminologies are systematically maintained and
probably globalized - I would say such terms are maintained via
standardization processes. recent advance of DOI allows use to assign
an official URI for a publication. so does DNS.
(ii) English is evolving by daily life usage (that just reminded me an
earlier mail by Jeremy on "live meaning and dead languages") - this is
case for a social vocabulary development
As mentioned in the previous email, standardization and social
evolution are useful and somehow complementary. However, they are more
or less an approach to the goal, which has been discussed in previous
email:
* given an URI in browser, a user really need to get it resolved (i.e.
be able to fetch its definition) , whether using the namespace of the
URI or use a search engine is just an option to the browser.
* knowing the ownership will be helpful for users to trust the
description of URI. The coexistence of exclusive ownership
(standardization model) or shared ownership (social wikipedia model)
are complementary solutions.
As David said, social evolution could be "dirty", but it also looks good:
* it is really not an easy job for one to maintain a comprehensive
description of URI for a long time (hard drive may crash, money may
run out, interests may switch, language may change). In addition to
the exclusive ownership for important concepts, open source style
social development should help too. so why not allow multiple
ownership for one URI
* the web is open, how can we help end users to pick one from the
multiple URIs referring to the same concept. so why not use social
ranking (where ownership may play an important role) to enable the
Darwin evolution.
--
Li Ding
http://www.cs.rpi.edu/~dingl/
