Till Kinstler wrote:
> Edward M. Corrado schrieb:
>
>> Yes, the ability to create your on interfaces using the API is great.
>
> Is there some documentation or whatever available on that API somewhere?
I don't know.
>
>> While there certainly are technical changes of provided > 400,000
>> items (most in full text) for people to re-index on there own,
>
> The Dartmouth College Beta at
> http://dartmouth.summon.serialssolutions.com/ has 170,508,797(!)
> records
> (http://dartmouth.summon.serialssolutions.com/search/results?spellcheck=true&q=),
> though I have the feeling only part of it with fulltext indexed (if any).
Opps, my bad, I left off three zeros, I meant > 400 million. The
170,508,797 is only want Dartmouth has decided they have access to [1].
If you click the "Add results beyond your library's collection" box you
will see they are actually searching against 420,351,451 records. I am
pretty sure they index full text when it is available, but they don't
display it for license reasons. However, I may be wrong about that and I
couldn't find a definitive answer on their Web site in 3 minutes of
looking. Does anyone know?
> You can handle that (Summon is using Solr as index software,
> correct?), but you'll need some hardware, I guess...
Yes Summon is using Solr. Not sure how much hardware you need but I'm
sure I could get my hands on it if needed with today's prices. It's
amazing what size datasets you can deal with these days.
> Throw away your OPAC, even if it can technically handle that amount of
> data, you won't find anything with typical OPAC search and sort
> capabilities.
>
I don't see Serials Solutions as going for that level, at least not yet,
with Summon. I see that they are more looking at strictly discovery.
They don't even do any delivery, just send you off to a link resolver.
>> bigger issue is getting the publishers who are providing Serials
>> Solutions the content to agree to this.
>
> We are getting some data from commercial providers as well. There
> seems to be a wide range of attitudes among providers what one is
> allowed to do with that data. That is an issue, sure. And it gets
> really tricky when you have a data pool with different licenses
> attached...
Yea, different licenses make it really difficult. It also goes down to
who is negotiating what. We get databases and other electronic content
through many different funding sources with many different players
involved that their is probably no way we would be able to reasonably
get enough data to do what Summon aims to do.
>
> Regards,
> Till
>
Edward
[1] I say decided they have access to, because some stuff the have
access too, like the Code4Lib Journal is indexed by Summon but do not
show up as being something that you can get to via Dartmouth.
Received on Thu Jun 18 2009 - 15:19:04 EDT