W3C home > Mailing lists > Public > public-semweb-lifesci@w3.org > August 2007

Re: Fleshed out "HTTP URIs are not Without Expense"

From: Bijan Parsia <bparsia@cs.man.ac.uk>
Date: Wed, 22 Aug 2007 10:58:32 +0100
Message-Id: <E4E7EE41-1FF6-4859-8E39-D04E92AB483A@cs.man.ac.uk>
Cc: public-semweb-lifesci@w3.org
To: "Chimezie Ogbuji" <chimezie@gmail.com>

On 22 Aug 2007, at 00:58, Chimezie Ogbuji wrote:

> Since this dialog is playing out on several fronts and I would like
> the dissenting view well-articulated, I've taken the liberty to flesh
> out the (previously empty) "HTTP URIs are not Without Expense" Wiki
> (http://esw.w3.org/topic/HCLS/HCLS_URI_matrix/HttpUrisAreExpensive).
> I've moved it to a top level location (instead of /topic/HCLS):

I have no preferences about organization, just as long as there's a  
list I can find :)

> http://esw.w3.org/topic/HttpUrisAreExpensive
>
> I also included links (in context) to the URN registries finding and
> the AWWW document.

Thanks Chimezie! Lots of good stuff in there.

I replaced the list in the generics tool page a pointer to this page  
and deleted the backlink.

As it reads now, however, it's clearly an advocacy page (though  
advocacy for what I'm not sure; probably for "not recommending that  
people only use http uris" which I've supported). But moving  
disputation from the mailing list to the Wiki doesn't really change  
the dynamic! (Though it can be helpful by providing a canonical spot  
for an argument.)

So, what is in dispute:
	1) I think there is dispute about whether HTTP uris are expensive,  
or whether their expense is (typically) a problem
For example, David raised this point:
	<http://www.w3.org/mid/ 
EBBD956B8A9002479B0C9CE9FE14A6C203113B6B@tayexc19.americas.cpqcorp.net>
	" Yes, some may view the cost of minting dereferenceable URIs as too  
high in
some situations, but I think the ubiquity of the web shows that very
many will view the cost of publishing dereferenceable URIs as low enough
-- and probably still falling -- to be outweighed by the benefits."

	2) I think there is dispute about how to handle cases where the  
expense is burdensome
	And the expense can fall on different people. For example, it can be  
expensive for a data provider to migrate to an HTTP uri scheme but  
that might be very helpful to tool builder or data users.

	Providing a "public trust" publishing and (http) URI minting site  
can help lower the costs for a data provider (such as Jane)  
(biordf.org comes to mind; as do PURLs).

	So one way to handle the issue in the recommendation is to say  
"Consider using foo or bar service, at least for providing a mapping,  
when the burden of providing your own HTTP uri space is too high."
	If there is an automagical, coordinated mapping from URNs to HTTP  
uris (either as replacement or by description) then one might get the  
best of both worlds (at the cost of having a trusted mediator)

So the first is a dispute about whether there are cases where http  
uris are significantly more expensive than the alternative (or for  
which cases this is true). The second is a dispute about various  
proposals (e.g., recommend that such people use such and such an URN  
scheme, etc.)

If I were refactoring that issue discussion page, I would present 1  
as a matter of research and present 2 as a series of proposals (with  
pros and cons).

There's a related issue, esp if it is recommended that people only  
use HTTP uris. Michel asked:
	<http://www.w3.org/mid/ 
AB349814F1ECB143A5D4CD29C7A645690192D831@CCSEXB10.CUNET.CARLETON.CA>

" How do I make statements about these
resources, without taking the responsibility of serving it up in my own
namespace [4], which might ultimately not integrate with content from
another 3rd party content provider."
	
This is a question about legacy identifier schemes where there is an  
owner, or at least a popular source, of data using the identifer.  
Should one lobby the data provider to switch over? To provide  
mappings? To use a resolver/mapping/mediator service? Can one just do  
this on one's own? As a third party?



Cheers.
Bijan.
Received on Wednesday, 22 August 2007 09:57:16 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 18:00:49 GMT