[rdfweb-dev] Serious question about silly attributes

[Martin L Poulter]

I'd like clarification on a point about how "democratic" the RDFweb
would allow us to be with attributes. I spoke to danbri last night but
didn't get a chance to hear his face-to-face explanation- sorry Dan.

Julian Bond suggested that 
>> - Webpage and IM Chat look to me like primitives that have multiple
>> types. Rather than thisWebpage and thatWebpage, shouldn't it be
	webpage
>> type=""? and instead of thisIMChat and thatIMchat, <im service=""
>> rdf:resource="URI">

Morten Frederiksen replies:
>'tis not the RDF way, which is triples, not quadruples.
...and gives references.

I accept this, but the implications seem contrary to the intended spirit
of making data (well, definitions of data) democratic. 3 examples which
bear on the point:

1) Whenever I install a web-based discussion board system, I create a
local instant messaging service. I'm free to represent the IDs in RDF
and merge my RDF document with FOAF, but direct inclusion in the FOAF
spec seems to privilege Microsoft's, AOL's and a few other IM systems
over others. Okay, they are more popular at the moment amongst the
aggregate of Internet users, but this will change as fads come and go,
and will be irrelevant to certain communities.

2) The inclusion of geekCode in the spec has been described as a "joke"
but I wouldn't call it that: people find all sorts of _silly_ or
_frivolous_ uses for the Internet, but an attempt to enable people to
do those things is not a joke.
(BTW, Slashdot Karma is much more relevant to the present-day geek
community than Geek Code, but then if you include Slashdot you have to
include lots of other boards.)
Let's imagine we keep geekCode in the spec because it enables people to
say things that some people want to say about themselves. What then
about variations that users come up with themselves, like the Wednesday
code? 
http://groups.google.com/groups?hl=en&lr=&ie=UTF-8&oe=UTF-8&selm=6386i8%24ajp%40huitzilo.tezcat.com
Okay, it's a joke, but in a truly people-owned semantic web, people
should be able to do this. And in principle they can, but how much
difference will not being in the FOAF spec make?

Now if there were _some way_ for the FOAF spec to only describe generic
IM services, discussion board karma ratings or personality descriptions
and leave the specification of these to external documents
maintained by the relevant service or community, that would seem to make
the RDF Web very democratic and would require minimal maintenance for
the spec. I'm ready to trust that this is not possible and/or
desirable, but I'd like more explanation of why.

3) A final example: images that depict a person could do with an
attribute for AmIHotOrNot rating [*]. More use would be made of this
than of, say Myers-Briggs personality type. But once you create an
attribute for that, you realise there are hundreds of these services:
http://dmoz.org/Recreation/Picture_Ratings/
Given that not all these can or should have FOAF attributes, it is up to
the communities (Goths, Norwich men...) to create their own vocabs
for their pictures and integrate the data with their FOAF data (let me
just interject the word "semantic" here ;-)  ). Will this be easy to do?
If not, which picture-rating vocabularies get the advantage of inclusion
in FOAF? If none, why exclude this form of self-description while
including Geek Code?


[*] I'm wary of my contributions sounding seedy, but again I'm just
thinking of real people's online lives, and it is easy to find
academically respectable but similarly structured applications.

PS: I've thought of another seedy example: purity ratings. Lots of
people are going to want to put Purity Test scores in their online
self-descriptions. There are lots of different online PTs but each has a
different URL so each attribute can in principle be given its own URI.
Would attaching semantics to each attribute be easy? Why have Geek Code
and not purity rating?
http://www.armory.com/tests/purity.html
-- 
Dr Martin L Poulter    Senior Technical Researcher, ILRT, Bristol, UK
Research interests: Philosophy of belief and Bayesian inductive logic

Home Page: http://www.weird.co.uk/martin/                                 
for Cult Concern FAQ + WEIRD (not WIRED) + "Bob" in the UK + Automated Love
+ Scientology Criticism + Sexual Politics + Helena Kobrin's Legal "Ethics".

Community blog: http://www.weird.co.uk/blog/