Re: Dons flame resistant (3 hours) interface about Linked Data URIs

I answer to Toby just becouse its handy to do so but i just want to
make a general statement.

Toby is stating the classical view, clean knowledge representation, 0%
dealing with ambiguity.

Hugh is hinting at is that the complexity of the "clean solution" is
overwhelming since it is counterintuitive in a lot of case and instead
of more hundreds of pages of tutorials and scripts we all adopt a more
pragmatic and tolerant view - if web effect is to be reached.

The solution to Toby's problem, quite simple, is not to use dc:created
for your birthday, use a foaf bdaty (or something) property. If its dc
creator its a document, simple. (if dc doenst have ranges, drop DC
alltogether :-) )

Application who care about people will simple ignore dc:created and
make their own assumptions and data cleanup exactly like the HTML
browser make assumptions and cleanups.

Other applications which care about documents will make other processing.

How do people know they should not use DC for the markup? They learn
when they see applications making mistakes :-)  (so you got togive
them some SW application for them to care about SW.. surprise!)

The truth is that today we have seen no attempt on the "pure sw" to
really write software that is tolerant with respect to the different
way markup is used

We seem to think that we need to build a SW where the result is
perfect has no errors or noise "or else it will all blow up because no
human is in the loop". (see a previous message)

This view ignores that a great number of applications will still have
the user at the end of the chain.: SW applications can be simply
supporting your ranking to be smarter, or propose a user a set of
alternatives thta the user can still detect as ultimately wrong or
right..

hopefully right 95% of the time when the cleanup works properly.

Other applications/scenarios/use cases which require 99 or 99.99 of
certainty of results from data harvested from the open web  will come
later.. encourged by the successes of the earlier. we should NOT plan
for those at this pont.. or else we get stuck in 10 more years of
discussions.. (no, not really, by next year google will have
thousands0 of well defined ontological terms to use for people's
pages, probably game over)

example? take the most interesting semantic web application that
exists, Google Social Graph, API and it mixes people with URLs ... by
design. :-) (plus internal task specific cleanups as they see fit)

cheers
Giovanni

2009/7/10 Toby Inkster <tai@g5n.co.uk>:
> On Fri, 2009-07-10 at 10:12 +0100, Richard Light wrote:
>> so the final decision is: which of these URLs
>> should be the one which represents the subject of discourse?  And the
>> answer has to be:
>
> Neither.
>
> If I use the URL of my machine readable data to identity myself (myself
> being the subject of discourse) then it immediately creates ambiguities.
>
> What would it mean for the file to have a dc:created property? Would the
> value of that property be my date of birth, or would it be the date I
> first uploaded my data?
>
> The classic example is that if I use the same URL to represent myself
> and my web page, then how can I state that I am the creator of my web
> page without also asserting that I'm my own father.
>
> The URL of the file and the URL of the subject of discourse must differ
> so that we can make unambiguous statements about each, and make
> statements about the relationship between the two. An incredibly easy
> way to do this is to just add a fragment to the URL of the subject of
> discourse. Examples include:
>
>        <http://tobyinkster.co.uk/> represents my web page
>        <http://tobyinkster.co.uk/#i> represents me
>
>        <http://id.loc.gov/authorities/sh85082139> is a web page
>        <http://id.loc.gov/authorities/sh85082139#concept> is the
>           abstract concept of Mathematics.
>
> --
> Toby A Inkster
> <mailto:mail@tobyinkster.co.uk>
> <http://tobyinkster.co.uk>
>
>
>

Received on Friday, 10 July 2009 19:39:26 UTC