[microformats-discuss] Video Pop-up Link Maker

[Tantek Ç elik]

On 10/12/05 4:45 PM, "Stephen Downes" <stephen at downes.ca> wrote:

> Danny Ayers wrote:
> 
>> On 10/12/05, Tantek Çelik <tantek at cs.stanford.edu> wrote:
>> 
>> Even Dublin Core is more
>>  
>> 
>>> theory than practice on the Web,
>>>    
>>> 
>> 
>> How did you measure that?
>> 
>> I tried Google with "purl.org/dc" which produces 298,000 hits, but
>> those are all *about* DC, which is still consistent with theory over
>> practice.
>> 
> 
> Hm. A small subset of extant DC documents is aggregated here:
> http://oaister.umdl.umich.edu/o/oaister/
> 
> 5,914,431 records from  536 institutions
> (updated 30 September 2005)
>
> 
> No, not merely theory.


Ok, let's walk that path.

I went to the URL you gave:

http://oaister.umdl.umich.edu/o/oaister/

NO MENTION of Dublin Core on that page.

View source.

NO USE of Dublin Core on that page.

Ergo that page IS NOT an example of Dublin Core use in the wild.


Let's dig deeper though.

Click "Go to search now..." button.

NO MENTION of Dublin Core on that page.

View source.

NO USE of Dublin Core on that page.

Ergo that page IS NOT an example of Dublin Core use in the wild.


However, I can't help but notice the correlation between the search field
names and Dublin Core properties: Title, Author/Creator, Subject, etc.

Then again, such generic field names were in physical card catalogs and
library electronic card catalogs LONG before Dublin Core was conceived.  In
fact, Dublin Core was based on those non-Web sources AFAIK.  Thus the
existence of similar search field names is not sufficient to assume/infer
existence of Dublin Core, since an earlier source could be responsible.


Let's try a search anyway.

Enter "Smith" in the Author field.


First result is:

http://www.aclweb.org/anthology/P97-1036

NO MENTION of Dublin Core on that page.

View source.

NO USE of Dublin Core on that page.


At this point I'm willing to give up on this example and pronounce it as not
providing ANY examples of Dublin Core on the *Web*.


The statement stands (*emphasis* added):


  Dublin Core is more theory than practice on the *Web*.


If you disagree, please provide a good sampling of *actual* *http* URLs
which clearly illustrate use of Dublin Core, preferably with *visible* data
(rather than invisible metadata).

And even then, it is going to be difficult to disprove "more theory than
practice", as Danny pointed out:

>> I tried Google with "purl.org/dc" which produces 298,000 hits, but
>> those are all *about* DC, which is still consistent with theory over
>> practice.

Thanks,

Tantek