W3C home > Mailing lists > Public > public-xg-webid@w3.org > June 2011

Re: [foaf-protocols] WebID test suite

From: Henry Story <henry.story@bblfish.net>
Date: Wed, 29 Jun 2011 23:17:16 +0200
Cc: public-xg-webid@w3.org
Message-Id: <E45C4FFB-77F1-4382-9171-3C56A07A07EB@bblfish.net>
To: Mo McRoberts <mo.mcroberts@bbc.co.uk>

On 29 Jun 2011, at 22:54, Mo McRoberts wrote:

> On 29 Jun 2011, at 21:14, Henry Story wrote:
> 
>> Are there mechanisms in place that help synchronise the way these OIDs get used? The meaning of words (and so OIDs)  depends according to David Lewis "... on the conventions [the user] is party to. And these conventions are regularities in behaviour, sustained in an interest in co-ordination and an expectation that others will do their part" ( in "Convention: A Philosophical Analysis") Words do not get their meaning unless their is co-ordination. OIDs are words with global meaning, so there needs to be global coordination for them to be able to stabilise on a meaning. For this to work there has to be games that drive this co-ordination towards a stable point. If the main interchange format for the meaning of these words does not publish the OIDS, then how could there be a coordination problem that aligned the meaning of these OIDs with one another? 
> 
> By convention, existing schema are reused just as they are in RDF land.

Yes, but conventions don't just happen like that. There has to be coordination games for them to happen. There have to be reasons for them to come to an equilibrium. There does not in fact have to be a meeting of people for this to happen because otherwise there would be an infinite regress. How did people come up with language? "By comming together to convene on the meaning of terms" cannot be the right answer of course as David Lewis points out. 

We all drive on the right hand side (in France) because most other people do drive on the right hand side, and there is very good reason to keep driving on that side when one knows that others are doing that. It's a convention, because things could have been otherwise - eg: driving on the left as in the UK and Japan.

> See /etc/openldap/schema/* on pretty much any Linux or Mac OS X box for a selection of common schemata.

Ok, thanks for pointing that out. I did not realise I had that on my computer. Interesting :-)

> 
>>> A fun project for somebody one day would be to run a web service on a well-known URI which spat out RDFS/OWL/whatever for the objectClasses and attributes at a particular OID. Would be a nice extension to Harald Alvestrand’s OID browser… [e.g., http://www.alvestrand.no/objectid/2.5.html]
>> 
>> yes, that would be very interesting. Two further studies that would be needed in addition to this are:
>> - are these OIDs really used the same way, across enterprises? If so, what forced people to using them the same way
> 
> Nothing -forces- people, just as nothing forces people to use FOAF or Dublin Core. Crafting new attributes and classes is expensive and increases the time spent configuring or customising software which talks to those directory services, so inertia dictates people tend towards

ok, so there is a bit of synergy there in configuration space.

In the semantic web there is a much stronger force for coordination: the URI of any relation can be clicked on, and its documentation found immediately. Unless you had pointed it out I would never have known by looking at an OID that where I would find the documentation for it. This means that it can only evolve a lot more slowly. On the other hand it has been around for some time, so that compensates....

> 
>> - what is the proportion of these OIDs that do have such a widespread meaning?
> 
> Most of the ones you see “in the wild” do — browse around Harald’s site, or oid-info.com and you'll see there's an awful lot of “well-known” OIDs, but there are plenty of essentially private arcs too. Most large companies will have an arc or two under which they put private attributes, internal SNMP traps and so on.

ok.

> Naturally, every critical extension in an X.509 cert is identified by an OID too :)

yes, I knew that :-)

So there are these OIDs but there, and those that are one Harald Alvestrand site could certainly be semwebized. That would indeed be a useful project.

Now that still leaves the issue that the Ldap Interchange Format does not use OIDS, but LDIF is the key format for exchanging this information. It's a bit of a pitty that this got lost, because now the databases have ids that are global but they don't use it to communicate....

Still this does mean that writing ldap to semweb transformers could be a lot easier than transforming a normal SQL database, as those don't have GUID.

Henry


> 
> M.
> 
> -- 
> Mo McRoberts - Data Analyst - Digital Public Space,
> Zone 1.08, BBC Scotland, 40 Pacific Quay, Glasgow G51 1DA,
> Room 7066, BBC Television Centre, London W12 7RJ,
> 0141 422 6036 (Internal: 01-26036) - PGP key 0x663E2B4A
> 

Social Web Architect
http://bblfish.net/
Received on Wednesday, 29 June 2011 21:18:00 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:06:24 UTC