 |
 |
Jonathan -- I suspect a message sent to the developers network mailing list would have the greatest possibility of hitting the most right people. (Perhaps the only higher action-to-frustration route than posting it here on code4lib itself.) Peter On Feb 23, 2012, at 5:52 PM, "Jonathan Rochkind" <[log in to unmask]> wrote: > On 2/23/2012 5:35 PM, Stephen Hearn wrote: >> But there's a catch--when WorldCat redirects a search to the selected >> local library catalog, it targets the OCLC record number. If the >> holding library has included the OCLC record number in its indexed >> data, the user goes right to the desired record. If not, the user is >> left wondering why the title of interest turned into some mysterious >> number and the search failed. > > I've been wishing OCLC would change this for a while. > > When specifying WorldCat's redirects for your local catalog, it's > already possible to NOT specify an OCLCnum based search, but only > specify an ISBN, ISSN, etc search. If you do this, and the record HAS > an (eg) ISBN, it'll redirect to an ISBN search in your catalog. But if > the record doesn't have an ISBN, ISSN, etc, I think it'll just redirect > to your catalog home page. > > So WorldCat is already capable of redirecting to an ISBN search. But if > you config the OCLCnum search, it seems it'll always use it instead. > > I wish WorldCat instead would do the ISBN search if there is an ISBN, do > an ISSN search if there's an ISSN, and only resort to the OCLCnum search > if there's no ISBN or ISSN to search on. Or at least that could be a > configurable option. Would result in a greater proportion of succesful > 'hits' when redirecting to local catalog, which may not have an OCLCnum > in it for every single record that it possibly could. (For that matter, > what about when there are multiple OCLCnums, multiple records, for the > same manifestation? For instance, a German language cataloging record > and an English language cataloging record, for the exact same > manifestation, have a different OCLCnum. Will OCLC ever send the German > language cataloging record OCLCnum and miss becuase you had the English > language one? I dunno). > > Anyhow, I've tried making this suggestion before to relevant OCLC > people, but it's possible I never found the relevant OCLC people. It's > kind of hard to figure out how to make such feature suggestions to OCLC > in a way that won't just be dropped on the floor (not sure it's > possible, in fact). > > Jonathan > > >> Stephen >> >> On Thu, Feb 23, 2012 at 4:11 PM, David Friggens<[log in to unmask]> wrote: >>>>>> why local library catalog records do not show up in search results? >>> Basically, most OPACs are crap. :-) There are still some that that >>> don't provide persistent links to record pages, and most are designed >>> so that the user has a "session" and gets kicked out after 10 minutes >>> or so. >>> >>> These issues were part of Tim Spalding's message that as well as >>> joining web 2.0, libraries also need to join web 1.0. >>> http://vimeo.com/user2734401 >>> >>>>> We don't allow crawlers because it has caused serious performance issues in the past. >>> Specifically (in our case at least), each request creates a new >>> session on the server which doesn't time out for about 10 minutes, >>> thus a crawler would fill up the system's RAM pretty quickly. >>> >>>> You can use Crawl-delay: >>>> http://en.wikipedia.org/wiki/Robots_exclusion_standard#Crawl-delay_directive >>>> >>>> You can set Google's crawl rate in Webmaster Tools as well. >>> I've had this suggested before and thought about it, but never had it >>> high up enough in my list to test it out. Has anyone actually used the >>> above to get a similar OPAC crawled successfully and not brought down >>> on its knees? >>> >>> David >> >>