I'm an indexer (although not representing the big commercial companies).
I will add support for Nordic Webindex (http://nwi.ub2.lu.se) in some near
future. We are already collecting metadata for statistical purposes (to see
what we need to support). The EWI (http://www.ub.lu.se/desire/) support will
most likely be more advanced.
On Fri, 29 Nov 1996, John Kunze wrote:
> > From [log in to unmask] Thu Nov 28 05:58 PST 1996
> > On Wed, 27 Nov 1996, John Kunze wrote:
> > > The needs of Indexers will be different from those of Searchers and those
> > > of the Providers.
> >
The searchers will only get what we, the indexers feel that we can support.
To be honest (and cynical), a majority of the searchers only need what they
know exists... Those that advertise only care about the majorities.
> > I'd like to see some reasoning as to why these groups should have
> > different needs from DC. To my mind, the Indexers are providing a service
> > to the Searchers and so they'd better make sure that their use of metadata
> > data ties into the needs of the Searchers (ie is the same if possible).
>
> You're right about the Indexers (Altavista, Lycos, etc.) providing a
> service, but because they're not better represented in the meta2 group,
> we have to sell DC to them as if they were an audience. I'll put in a
> plug now to support Stuart in getting them to the next meeting.
>
> I'm especially concerned about the Indexer "audience", because they
> have non-obvious needs that we know little about. At the Boston W3C
> workshop, for example, we learned that Lycos was unlikely to be able
> to guarantee preservation of element order within records; this has
I can do that. But I have other needs. First, I currently need to
transform any DC record to a GILS one (because I use Z39.50 with that
profile). In some not too distant future I will need to transform
it to zdsr/zstarts instead. Other indexers will have to do other
transformations. Second, I need to support other types of metadata than
DC (which is virtually absent in HTML documents). In particular
I need to support attributes that are fairly close to an IAFA template.
Also, since you in this group are (1) "rocking the boat", and (2) since I've
got an important would-be client in the public sector that want a fairly high
quality database of records describing various governmental databases and
publication it might be that I add support for embedded GILS (please note
that the GILS community doesn't rock the boat as much as you do) before DC.
> a profound impact on any of our plans to assign significance to element
> order. Similarly, Benjamin Franz's message (from the robots mailing list,
> posted to meta2 Oct 15), even if somewhat extreme, indicates that there
> is some sentiment among Indexers that external schemas in DC won't work.
I've made that point in this list before. I don't like the idea to spawn
nsgmls or something similar for parsing external metadata record. What an
indexer need is a foolproof way of determining which "document like object"
is described by which metadata object. And this connection between
metadata and the described object should be obvious from their respective
HTTP headers. (And ideally NOT using HTTP-EQUIV).
> We ignore Indexers at our peril. I don't like admitting it, but we
> need them more than they need us.
True
>
>
Yours,
Sigfrid
________________
Sigfrid Lundberg, Ph.D. [log in to unmask]
Lund University library
Netlab, PO Box 3, S-221 00 Lund +46-222 36 83
|