On Thu, Nov 12, 2009 at 1:52 PM, Weinheimer Jim <j.weinheimer_at_aur.edu> wrote:
> I understand this. I was (and am) looking at the matter as a web designer and seeing that because it is a "dead page" the id.loc.gov reference is essentially useless as it now stands. (I realize that this may not be a very popular thing to point out), while the dbpedia page is highly useful. Why? Because of the links that the one provides, both internal and external, while the other winds up in a dead-end.
Jim, your argument honestly doesn't make any sense.
1) What does looking at this page as a web designer matter? It's not
intended for human consumption.
2) Why would you send a human to DBpedia over wikipedia or Freebase?
3) id.loc.gov/authorities is a thesaurus, not an encyclopedia. For
that matter, neither is dbpedia -- it's intended to provide
identifiers to represent the non-information resouces described by
Wikipedia articles and categories.
What you're saying is exactly as if I criticized www.biblio.tu-bs.de
for not giving me anything useful to incorporate into
http://purl.org/NET/lccn/36029351#i. Or that your OPAC won't give me
back results of local bookstores.
>
> One of the main reasons the id.loc.gov version is so dead is that it takes conflates 3 separate parts, thereby turning this into a strange type of RDF-ized "textual string."
> Shakespeare, William, 1564-1616--Authorship--Oxford theory actually *is* http://id.loc.gov/authorities/sh85120839
>
> when it should be (and easily could be) broken up semantically, so that at least the internal links will work. As I consider this some more, I cannot believe that such a treatment is even correct in theory. Shakespeare as a topic should not be completely divorced from Shakespeare, William, 1564-1616--Authorship--Oxford theory, but it is an integral part of it. Still, it is separate in id.loc.gov, while it is not done this way in dbpedia, as there is a link to Shakespeare.
>
Ed tried to start a discussion on this mailing list about how to model
coordination in SKOS.
https://listserv.nd.edu/cgi-bin/wa?A2=ind0911&L=NGC4LIB&T=0&F=&S=&P=6304
This is non-trivial. I also don't know of any interface where you can
break a subject heading up into its component parts. Even
authorities.loc.gov doesn't have any capacity to break a heading up.
Nor does http://www.biblio.tu-bs.de/db/lcsh/page.php?urG=LCS&urA=18&urS=_shakespeare,+william,+1564-1616+--+authorship+--+collaboration
In your original mention of this link you talk about removing parts of
the string to find other things. You can do that in id.loc.gov, too:
http://id.loc.gov/authorities/label/Shakespeare,%20William,%201564-1616--Authorship
Besides, Shakespeare, William isn't a topic, it's a personal name and
the names authority hasn't been added yet. Once the NAF is there, I'm
pretty sure you'll see linkages between the two.
> The lack of flexibility in this unwieldy string makes me think that once again, we are replicating the card catalog online, forcing people to go to the beginning of the string before things can make sense. Very, very unfortunate.
>
> The dbpedia page for Amharic also links into some other useful, similar pages:
> http://sw.opencyc.org/concept/Mx4rvVjlN5wpEbGdrcN5Y29ycA
> http://rdf.freebase.com/ns/guid.9202a8c04000641f80000000000a43f2
>
> Faced with these choices, why would any non-library developer choose to work with id.loc.gov, except perhaps in only a theoretical way? And if people won't work with it, who will use it as a reference? So, certainly we can put these things on the web just like anyone else, but if our tools are dead-ends, and theirs "have links" so that people can actually use them, then it's a no-brainer, as it seems to me. Still, yes, if everybody used http://id.loc.gov/authorities/sh85004449#concept to code their pages for Amharic language, matters would be better. Using http://id.loc.gov/authorities/sh85120839 for Shakespeare, William, 1564-1616--Authorship--Oxford theory is less satisfactory. Still, for the moment, I think in both cases, I would bet my money on dbpedia.
You're still missing the point. id.loc.gov has nothing to do with
_pages_. It has to do with _data_. It's purpose isn't to add as a
link to your webpage, it's to add a link to your 600 field so we're
not doing weak text string queries across systems.
I suppose I don't understand your motivations in this critique of
id.loc.gov: first you start a thread that had over 100 something
replies about how we need to get our data "out there", even if it's a
text delimited file and then you slam id.loc.gov (which, incidentally,
gets the data out there in exactly the way that Tim Berners-Lee would
prefer) for not being 100% perfect.
-Ross.
Received on Thu Nov 12 2009 - 15:15:38 EST