Re: Using third-party vocabularies

>At 01:22 PM 12/5/02 +0000, Brian McBride wrote:
>>At 12:14 05/12/2002 +0000, Graham Klyne wrote:
>>
>>[...]
>>
>>>What the previous text does not say, and concerning which there 
>>>was a comment on the Concepts document,
>>
>>Reference please.  What was the comment?
>
>It's logged, with links, at:
>   http://www.ninebynine.org/wip/DocIssues/RDF-Concepts/022-SocialMeaning.html
>
>The comment being:
>[[
>Similarly, the tying of the meaning of a URI to the ill-specified
>intent of some organization poses a giant bar to the deployment of
>RDF.  Under these circumstances how can any organization use an URI
>that they do not own?  The owning organization might, after all,
>choose to change the meaning of any URI they own at any time.  This
>seems to me to be a bar to any communication between organizations
>using RDF.
>]]
>
>>>  is that even though third-party vocabularies are generally 
>>>unconstrained by opthers who may use them, there may yet be some 
>>>that are sufficiently well-trusted for serious use.  If you don't 
>>>want to go into legal territory, the final sentence might be pared 
>>>down to, say:
>>>
>>>[[
>>>For important documents this may mean that use of third-party 
>>>vocabulary is restricted to terms defined by reputable 
>>>organizations (e.g. recognized standards bodies), or that 
>>>otherwise have socially well-established meanings.
>>>]]
>>
>>Right, that avoids the pitfall I mentioned, but I'm still wondering 
>>why a normative spec would be saying anything of the form  "There 
>>might be ..."
>
>OK, let's try again:
>
>[[
>For important documents, the use of third-party vocabulary should be 
>restricted to terms defined by trustworthy parties (e.g. recognized 
>standards bodies or reputable organizations), or that otherwise have 
>socially well-established meanings.
>]]
>

Seems to me that we losing the essential point here. Of course anyone 
CAN use any RDF they find to draw conclusions, and the conclusions 
will be valid consequences of the RDF they use. We shouldn't go on 
record as seeming to be saying that using RDF not authenticated by a 
socially reputable source is technically dangerous or illegal or 
somehow frowned upon, or that using RDF necessarily requires one to 
consult lawyers.

All this comes from Tim wanting us to be clear that RDF, when 
asserted, really is asserted, is all.  And that this means 'asserted' 
in the ordinary sense, not in some narrow technical sense: one uses 
RDF to actually say things
.
We might say something like this:

---
Publishing some RDF should be understood as making a public assertion 
in the same sense that publishing some natural language makes an 
assertion, and the full meaning of this assertion should be 
understood as encompassing all the usual social senses of meaning in 
the usual way that this term might be understood to apply to any 
other publication of a definition, claim, assertion of fact or 
statement of opinion in any other medium; except that in the case of 
RDF, these notions of meaning should be understood to carry over to 
any formal consequences of the published RDF which can be validly 
inferred from the published RDF by using correct RDF reasoning. The 
intended meaning conveyed by publishing some RDF in this sense may be 
more than is captured by the formal semantics - it may for example be 
conveyed partially by English comments or other non-machine-readable 
forms of expression in the original document - but it is required to 
be an extension or strengthening of the formal meaning, and hence to 
be preserved under valid formal inference processes.

When RDF published by several different sources is put together, the 
combined assertions may have valid entailments whose full meaning can 
be determined only by examining the intended meanings of all of the 
sources, and in such cases the question of which source or sources 
are responsible for the veracity of the conclusion may be complex, 
and require some detailed analysis. In extreme cases this could be 
subject to legal or contractual interpretation. For example, [[insert 
clown example and discussion here.]]

While we anticipate that RDF content will be used primarily for less 
contentious forms of expression, the general point remains that users 
are ultimately responsible for checking that the meanings of the 
urirefs in any RDF they publish correspond to the intended meanings 
of such 'third-part' urirefs as specified in the originating 
documents, since readers of their RDF are free to combine it with RDF 
published by others which also use those same urirefs, and to draw 
any formally valid conclusions.

---

I think this style - a kind of vague general warning that there is a 
real issue here, but short of an outright MUST or SHOULD - is about 
the right one to adopt.

Pat
-- 
---------------------------------------------------------------------
IHMC					(850)434 8903   home
40 South Alcaniz St.			(850)202 4416   office
Pensacola              			(850)202 4440   fax
FL 32501           				(850)291 0667    cell
phayes@ai.uwf.edu	          http://www.coginst.uwf.edu/~phayes
s.pam@ai.uwf.edu   for spam

Received on Thursday, 5 December 2002 19:44:35 UTC