[Foundation-l] Universal Library

[David Goodman]

I have been re-reading their documentation, and they have it well in
hand.  We would do very well to confine ourselves to matching up the
entries in the WMF projects alone. Some of the data in WMF is more
accurate than some of the OL data, but I would not  say this to be a
general rule. Far from it: the proportion of incomplete or inaccurate
entires in enWP is probably well over 50% for books. (for journal
articles it is better, because of a project to link to the pubmed
information)  The accuracy  & adequacy -- let alone completeness-- of
the bibliographic information in WS is close to zero, except where
there is a IA scan of the cover and title page, from which full
bibliographic information might be derived, but cannot necessarily be
taken at face value.

The unification of editions is non-trivial, as using the algorithm you
suggest, you will also have all works related to Verne, and
additionally a combination of general and partial translations,
children's books, comic adaptation, and whatever.
Modern library metadata provides for this to a certain limited
extent--unfortunately most of the entries in current online catalogs
do not show full modern data--many catalogs never had more than
minimal records;  Dublin core is probably not generally considered to
be fully up to the problem either, at least in any current
implementation.

Those working on the OL side are fully aware of this. They have made
the decision to work towards inclusion of all usable & obtainable data
sets, rather than only the ones that can be immediately fully
harmonized. This was very wise decision, as the way in which the
information is to be combined & related is not fully developed, and ,
if they were to wait for that, nothing would be entered. There will
therefore be the problem of upgrading the records and the record
structure in place--a problem that no large bibliographic system has
ever fully handled properly--not that this incarnation of OL is likely
to either. Bibliographers work for their time, not for all time to
come.


David Goodman, Ph.D, M.L.S.
http://en.wikipedia.org/wiki/User_talk:DGG



On Thu, Sep 3, 2009 at 6:38 AM, Yann Forget<yann at forget-me.net> wrote:
> David Goodman wrote:
>> I have read your proposal. I continue to be of the opinion that we are
>> not competent to do this. Since the proposal  says, that "this project
>> requires as much database management knowledge as librarian
>> knowledge," it confirms my opinion. You will never merge the data
>> properly if you do not understand it.
>
> That's all the point that it needs to be join project: database gurus
> with librarians. What I see is that OpenLibrary lacks some basic
> features that Wikimedia projects have since a long time (in Internet
> scale): easy redirects, interwikis, mergings, deletion process, etc.
> Some of these are planned for the next version of their software, but I
> still feel that sometimes they try to reinvent the wheel we already have.
>
> OL claims to have 23 million book and author entries. However many
> entries are duplicates of the same edition, not to mention the same
> book, so the real number of unique entries is much lower. I also see
> that Wikisource has data which are not included in their database (and
> certainly also Wikipedia, but I didn't really check).
>
>> You suggest 3 practical steps
>> 1. an extension for finding a book in OL is certainly doable--and it
>> has been done, see
>> [http://en.wikipedia.org/wiki/Wikipedia:Book_sources].
>> 2. an OL  field,  link to WP -- as you say, this is already present.
>> 3. An OL field, link to Wikisource. A very good project. It will be
>> they who need to do it.
>
> Yes, but I think we should fo further than that. OpenLibrary has an API
> which would allow any relevant wiki article to be dynamically linked to
> their data, or that an entry could be created every time new relevant
> data is added to a Wikipedia projects. This is all about avoiding
> duplicate work between Wikimedia and OpenLibrary. It could also increase
> accuracy by double checking facts (dates, name and title spelling, etc.)
> between our projects.
>
>> Agreed we need translation information--I think this is a very
>> important priority.   It's not that hard to do a list or to add links
>> that will be helpful, though not  exact enough to be relied on in
>> further work.  That's probably a reasonable project, but it is very
>> far from "a database of all books ever published"
>>
>> But some of this is being done--see the frWP page for Moby Dick:
>> http://fr.wikipedia.org/wiki/Moby_Dick
>> (though it omits a number of the translations listed in the French Union
>> Catalog, http://corail.sudoc.abes.fr/xslt/DB=2.1/CMD?ACT=SRCHA&IKT=8063&SRT=RLV&TRM=Moby+Dick]
>> I would however not warrant without seeing the items in hand, or
>> reading an authoritative review, that they are all complete
>> translations.
>> The English page on the novel lists no translations;  perhaps we could
>> in practice assume that the interwiki links are sufficient. Perhaps
>> that could be assumed in Wiksource also?
>
> That's another possible benefit: automatic list of
> works/editions/translations in a Wikipedia article.
>
> You could add {{OpenLibrary|author=Jules Verne|lang=English}} and you
> have a list of English translations of Jules Verne's works directly
> imported from their database. The problem is that, right now, Wikimedia
> projects have often more accurate and more detailed information than
> OpenLibrary.
>
>> David Goodman, Ph.D, M.L.S.
>> http://en.wikipedia.org/wiki/User_talk:DGG
>
> Regards,
>
> Yann
> --
> http://www.non-violence.org/ | Site collaboratif sur la non-violence
> http://www.forget-me.net/ | Alternatives sur le Net
> http://fr.wikisource.org/ | Bibliothèque libre
> http://wikilivres.info | Documents libres
>
> _______________________________________________
> foundation-l mailing list
> foundation-l at lists.wikimedia.org
> Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/foundation-l
>