Re: Managing vocabularies at W3C

Hi Phil,

The problems you outline below are of a different provenance in the US.  DCAT does not need "fixing" as a specification (that is the good news), but the predecessor American tradition needs some context.

This famous cartoon by Benjamin Franklin ...
http://www.loc.gov/pictures/item/2002695523/
demonstrates the point.

Caution here my Brit friend ...
At the time of publication (1754) the American Revolution was many years in the future.  Dr. Franklin was a bona fide genius, the inventor of Electricity.  And, although his doctorate degree was from Harvard, he would not let himself be addressed as "Dr. Franklin" until after being awarded a degree, causa honoris, from Oxford.  As a publisher he knew well how pesky Investigative Journalists can be about official records.

In the context of W3C name spaces, by "Join or Die", Dr. Franklin meant that on a frontier there was a certain locational safety coupled with a certain positional safety and together they determined the probability of persistence and survival.  Although there were only seven English Colonies at the time and not thirteen, a safe place (today called a "Safe Harbor" in legalese) had all seven component names (URN's) or none at all.  Things could always go sideways if you were caught in the wrong place at the wrong time.

Americans continue this Linked Data tradition by normalizing names and official alpha and numeric encodings with customary frame buffer widths.

Although I can neither make the paper deadline nor risk the local entertainment cost overruns a visit to Amsterdam entails, I am willing to provide the following background materials for the "Ben Franklin Standard":

A fully row linked GOVT_UNITS_20160601.txt table (USGS Board on Geographic Names) in CSV (With Column Headers), TXT (Headless), a  plain jane SQL (without mySQL extensions), a PERL filter to adjust the SQL processing artifacts, and of course an Indexed SQL Schema.  This is all about indexes and the ability of a relational data base to make and use indexes which in effect have a zero-width joiner between fields.  These indexes are "Machine Readable".  Not.

This "Standard" + DCAT = Cookbook Recipe Zero.

Let me know.

--Gannon





--------------------------------------------
On Wed, 10/5/16, Phil Archer <phila@w3.org> wrote:

 Subject: Managing vocabularies at W3C
 To: "W3C Web Schemas Task Force" <public-vocabs@w3.org>
 Date: Wednesday, October 5, 2016, 5:00 AM
 
 Dear all,
 
 A perennial topic of discussion at W3C is how we can better
 support 
 vocabulary development and management. The way schema.org is
 managed is 
 an example that clearly works and has wide support. Is it
 right for 
 other vocabularies? (more yes than no I assume). What value
 does a 
 w3.org namespace offer? What are the appropriate rules for
 adding new 
 terms to existing vocabularies? What's the appropriate
 relationship 
 between a document created by a Working Group or Interest
 Group and 
 published in w3.org/TR space and the associated namespace?
 
 And when you have a vocabulary, how do you publish a profile
 of it, one 
 that you can validate instance data against? How can you
 request data 
 following a specified profile?
 
 These are the kind of questions we have been wrestling with
 for too long 
 and we'd really like to get it sorted. The workshop we're
 organising at 
 the end of November in Amsterdam is a big part of that. The
 headline 
 vocabulary in question is DCAT [1], a vocabulary that needs
 some 
 attention, but the underlying process issues are just as
 important.
 
 If you have views on this, please come and share them in
 Amsterdam 30 
 November - 1 December. Details, including the awkward fact
 that the 
 deadline for paper submissions is this weekend, are at
 https://www.w3.org/2016/11/sdsvoc/
 
 Thanks
 
 Phil.
 
 
 [1] https://www.w3.org/TR/vocab-dcat/
 -- 
 
 
 Phil Archer
 W3C Data Activity Lead
 http://www.w3.org/2013/data/
 
 http://philarcher.org.
 +44 (0)7887 767755
 @philarcher1

Received on Wednesday, 5 October 2016 22:51:09 UTC