W3C home > Mailing lists > Public > public-rdfa-wg@w3.org > June 2011

Re: Comments on the RDFa Core Default Profile: dcterms prefix

From: Ivan Herman <ivan@w3.org>
Date: Wed, 15 Jun 2011 11:22:46 +0200
Cc: Niklas Lindstr?m <lindstream@gmail.com>, Sebastian Germesin <sebastian.germesin@dfki.de>, St?phane Corlosquet <scorlosquet@gmail.com>, Martin Hepp <martin.hepp@ebusiness-unibw.org>, Gregg Kellogg <gregg@kellogg-assoc.com>, W3C RDFWA WG <public-rdfa-wg@w3.org>, Thomas Baker <tbaker@tbaker.de>, Dan Brickley <danbrickley@gmail.com>, Shane P McCarron <shane@aptest.com>
Message-Id: <3C614058-FDC0-4AF5-ADA8-C4651188537E@w3.org>
To: Thomas Baker <tom@tombaker.org>
Tom,

On Jun 14, 2011, at 23:28 , Thomas Baker wrote:

> Dear all,
> 
> Swoogle is still unavailable, so I cannot confirm the data from June 2010
> [1].  


Well, the swoogle results (huge preference for dcterms) is in line with the Yahoo crawl results:

http://www.w3.org/2010/02/rdfa/profile/data/yahoo/

The Sindice resutlts

http://www.w3.org/2010/02/rdfa/profile/data/sindice/

also give a clear preference for dcterms although it does include dc on the top, whereas the yahoo crawl does not. Note that one of the main difference between Sindice and Yahoo is that the former crawls the SW in general, which also includes a fair number of older RDF data, whereas Yahoo looks at RDFa data only. Ie, in this respect, it is more relevant.

(That being said, I must admit I was surprised by the Yahoo results... but, well, that is what the facts tell us!)

> FWIW, data from BTC2010 shows http://purl.org/dc/elements/1.1/ being used
> in more than twice as many triples as http://purl.org/dc/terms/.  On the other
> hand, I have heard informally that /terms/ is used in RDFa data much more than
> /elements/1.1/.

Yep, see above.

> 
> In the past, DCMI has expressed a preference for:
> 
>    dc:       -> http://purl.org/dc/elements/1.1/
>    dcterms:  -> http://purl.org/dc/terms/ 
> 
> The prefix "dct:" is also used in place of "dcterms:".  As pointed out below,
> the following was also considered reasonable [3,4]:
> 
>    dc:       -> http://purl.org/dc/terms/
>    dce:      -> http://purl.org/dc/elements/1.1/
> 
> In [3], Mikael's point was that using dc: and dcterms: for 
> /terms/ could be confusing _if_ /elements/1.1/ were _also_ being
> used.

I agree

> 
> My preference, then:
> 
> -- If both /terms/ and /elements/1.1/ are being used, pick one of the two
>   mappings above, with a slight preference for the first.  (Or, if "dcterms:" is
>   simply too long, use "dct:".)

Based on the overall results in

http://www.w3.org/2010/02/rdfa/profile/data/

this is not an option any more...

> 
> -- If /terms/ *only* is being promoted, go ahead and use "dc:" for /terms/.
> 
> DCMI's position on this question is not strong either way.  Ultimately, the
> goal should be to minimize the risk of erroneous usage, such as using the
> prefix "dc:" both for /elements/1.1/ *and* for /terms/.  How likely would this
> error be, given the widespread use of dc: for /elements/1.1/?  If this group
> prefers to promote *only* /terms/ in this context, DCMI will have no objections
> and will point people to the decision.  

So: is it then a consensus (this question goes to everyone, not only to Tom) that we should use the 'dc' prefix for http://purl.org/dc/terms/ in the default RDFa profile? As far as I am concerned, I do not have any objections to that as long as the DCMI is fine with this.

Tom (and Shane, explicitly): there is a corresponding issue, though. The RDFa document:

http://www.w3.org/TR/rdfa-core/

uses the dcterms prefix with http://purl.org/dc/terms/. For consistency's sake we should probably change that to 'dc' everywhere. Would that be fine with you?

Thanks!

Ivan

> 
> Tom
> 
> [1] http://answers.semanticweb.com/questions/640/dct-or-dcterms
> [2] http://gromgull.net/2010/10/btc/explore.html
> [3] https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind0909&L=dc-architecture&D=0&P=26639
> [4] https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=dc-architecture;6b83a7bf.0909
> 
> On Fri, Jun 10, 2011 at 06:40:29AM -0700, Ivan Herman wrote:
>> I think the appropriate thing to do is to use the prefix that Thomas Baker, ie, DCMI, prefers... 
>> 
>> Ivan
>> 
>> On Jun 10, 2011, at 03:34 , Niklas Lindström wrote:
>> 
>>> Hello!
>>> 
>>> I've seen and always been using 'dct' for DC Terms, though I'm not
>>> sure how common it is in general. I would certainly prefer a short
>>> name. Ideally 'dc' (since "dc11" is legacy), or 'dct' otherwise.
>>> 
>>> Best regards,
>>> Niklas
>>> 
>>> 
>>> 
>>> On Fri, Jun 10, 2011 at 10:45 AM, Sebastian Germesin
>>> <sebastian.germesin@dfki.de> wrote:
>>>> Hi,
>>>> 
>>>> I just did a quick check on Google and found [1].
>>>> In their answer, they search on Swoogle [2] for the most used prefixes
>>>> and according to their numbers, 'dcterms' is used around 100 times more than
>>>> 'dc'.
>>>> Though I - personally - prefer shorter prefixes, I would adopt the "most
>>>> used" prefix.
>>>> 
>>>> Unfortunately, Swoogle is down, so I had no chance to verify the numbers
>>>> given in that post.
>>>> 
>>>> 
>>>> Best,
>>>> 
>>>> Sebastian
>>>> 
>>>> [1] http://answers.semanticweb.com/questions/640/dct-or-dcterms
>>>> [2]
>>>> http://swoogle.umbc.edu/index.php?option=com_frontpage&service=digest&queryType=digest_ns&searchString=http://purl.org/dc/terms/
>>>> 
>>>> 
>>>> 
>>>> Am 09.06.2011 um 23:42 schrieb Stéphane Corlosquet:
>>>> 
>>>>> 
>>>>> 
>>>>> On Thu, Jun 9, 2011 at 5:31 PM, Martin Hepp
>>>>> <martin.hepp@ebusiness-unibw.org> wrote:
>>>>> Hi all:
>>>>> What about dct: for dcterms?
>>>>> 
>>>>> I think that using dc: for dcterms may cause more confusion than good.
>>>>> 
>>>>> It will confuse those who know about the two namespaces. Newcomers will
>>>>> not know anything about these namespaces. 'dc' is such a common prefix, it
>>>>> would be a pity to not include it in the default profile, and have to
>>>>> declare it in the list of prefixes. afaik, all terms from the old namespace
>>>>> map well to the new one, but Thomas should confirm this.
>>>>> 
>>>>> Steph.
>>>>> 
>>>>> 
>>>>> Just my 2 cents.
>>>>> Martin Hepp
>>>>> 
>>>>> 
>>>>> On Jun 9, 2011, at 9:13 PM, Gregg Kellogg wrote:
>>>>> 
>>>>>> On Jun 9, 2011, at 9:12 AM, Stéphane Corlosquet wrote:
>>>>>> 
>>>>>>> If these prefixes are going to become widely used in the future, would
>>>>>>> it be possible to use the 'dc' prefix instead of 'dcterms' for
>>>>>>> http://purl.org/dc/terms/? 'dc' is shorter and will look more familiar to
>>>>>>> newcomers. This was discussed and advised in the DC-ARCHITECTURE during the
>>>>>>> development of RDFa in Drupal 7, see the rationale by Dan Brickley [2] and
>>>>>>> the responses from Thomas Baker and Mikael Nilsson [3].
>>>>>> 
>>>>>> +1
>>>>>> 
>>>>>>> regards,
>>>>>>> Steph.
>>>>>>> 
>>>>>>> [1] http://www.w3.org/profile/rdfa-1.1#crawl
>>>>>>> [2]
>>>>>>> https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=dc-architecture;6b83a7bf.0909
>>>>>>> [3]
>>>>>>> https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind0909&L=dc-architecture&D=0&P=26639
>>>>>> 
>>>>>> Gregg
>>>>> 
>>>>> 
>>>> 
>>>> --
>>>> M.Sc. Sebastian Germesin
>>>> Deutsches Forschungszentrum fuer Kuenstliche Intelligenz GmbH (DFKI)
>>>> 
>>>> Stuhlsatzenhausweg 3
>>>> 66123 Saarbruecken
>>>> Germany
>>>> 
>>>> Tel.: +49.681.85775.5079
>>>> Fax.: +49.681.85775.5021
>>>> 
>>>> email:   sebastian.germesin@dfki.de
>>>> GPG:     http://www.dfki.de/~germesin/gpg/germesin_dfki.gpgkey
>>>> web:     http://www.dfki.de/~germesin
>>>> skype:   neogermi1337
>>>> twitter: neogermi
>>>> github:  http://github.com/neogermi
>>>> 
>>>> ###
>>>> #
>>>> #
>>>> 
>>>> 
>>>> 
>>> 
>> 
>> 
>> ----
>> Ivan Herman, W3C Semantic Web Activity Lead
>> Home: http://www.w3.org/People/Ivan/
>> mobile: +31-641044153
>> PGP Key: http://www.ivan-herman.net/pgpkey.html
>> FOAF: http://www.ivan-herman.net/foaf.rdf
>> 
>> 
>> 
>> 
> 
> -- 
> Tom Baker <tom@tombaker.org>


----
Ivan Herman, W3C Semantic Web Activity Lead
Home: http://www.w3.org/People/Ivan/
mobile: +31-641044153
PGP Key: http://www.ivan-herman.net/pgpkey.html
FOAF: http://www.ivan-herman.net/foaf.rdf
Received on Wednesday, 15 June 2011 09:20:06 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 27 April 2012 04:55:17 GMT