Hi Tom,

The only way I know is to go to each Work and remove the "bad"
subjects, adding or leaving the "good" ones.

That process may be performed by a bot like (my) VacuumBot, but (1)
subjects are in a list (making it impossible(?) to look for works with
a specific subject via the query.json API) and  (2) I found it
trickier to identify the meaning of subjects based on the words and
"correcting" them on that basis (changing a format "hardcvoer" into
"Hardcover" is less tricky). Therefore I haven't tried to make
VacuumBot do this.
One could use a data dump to find records with "bad" subjects of
course, to counter the first problem. Setting rules for what should be
changed could help a bot programmer to make bots do the hard work.
(I'm busy however, so please don't count on me ;).)

Ben

P.S. It looks like subject used to be a first class record object in
the OL Infobase, as in the dumps I've counted 91400 records of type
"/type/subject". There is only a name and administrative metadata in
those records, so no apparent relationships among the records.

On 25 November 2012 17:47, Tom Morris <[email protected]> wrote:
> When I look at the subjects listed here:
>
>  http://openlibrary.org/search/subjects?q=history
>
> I see 9 or 10 different variants of the simple one word subject "History"
> including a variant which apparently is a "Place."
>
> Is there a way to merge these the way one merges books/works?
>
> Tom
>
> _______________________________________________
> Ol-discuss mailing list
> [email protected]
> http://mail.archive.org/cgi-bin/mailman/listinfo/ol-discuss
> To unsubscribe from this mailing list, send email to
> [email protected]
>
_______________________________________________
Ol-discuss mailing list
[email protected]
http://mail.archive.org/cgi-bin/mailman/listinfo/ol-discuss
To unsubscribe from this mailing list, send email to 
[email protected]

Reply via email to