Erin Morgan wrote:
> Yes, the idea of "extended search" is more achievable but I also see
> this as a short-term solution. Search an index (say, the catalog).
> Get a list of results. Behind the scenes reformulate the search to
> query another index (say, OAIster). Search. Get a list of results.
> Repeat. Such an implementation is a step in the right direction, but
> as the number of your "information silos" grows so does the tedium
> experienced by the user. "Which silo should I choose this time?"
> Again, metasearch won't really work here because the content is in
> different locations using different protocols, etc.
This is why I mentioned that this could be improved through the use of XLINKS and Web Services. Instead of selecting specific databases, the user could select special, selected "themes" such as, Archaeological resources, Art resources, etc. and in the background, multiple queries could be directed specifically for each database. With Web Services especially, all the protocols could be dealt with, and the librarian maintains some control over the search. I can't do this yet.
> I don't necessarily agree about loosing control of a catalog if
> masses of uncertain authority control content is integrated into it.
> Yes, your descriptions will be less consistent with each other.
This is unfortunate because the whole point of a catalog is to provide reliable access to the materials in your collection, however you define "collection." The only reason that reference librarians can help regular users is because there is a system (consistency) to it. Therefore, when someone asks, "What do you have from the Budapest Museum of Art?" within certain known parameters, the reference librarian can help and answer this truthfully, because the catalog is designed to provide reliable results through consistent use of authorities.
Of course, this is an "expert system" and the normal user would have a lot of problems with it. There has not been nearly enough interest previously on making our expert system much more user friendly, and this needs to be corrected as much as possible, but if we take away the consistency by merely mashing everything in together, when someone asks the question, "What do you have from the Budapest Museum of Art?" the reference librarian can only reply, "I haven't got the slightest idea and the system doesn't allow me to find out. Let's founder around a bit until we get a result that makes you happy and you won't complain," thus emulating a Google result.
> On
> the other hand, libraries, especially libraries with larger
> collections can not afford to catalog (describe) all of their
> materials in great detail. It is too expensive.
This is a sentiment that I have seen written over and over, but I haven't seen any evidence to support it: cataloging is "too expensive." In my experience, cataloging is not among the most highly paid of professions, and when you take the cost of original cataloging of an item (say, an average of 1/2 hour work--say $30-$50, but in many cases less), and compare it to what the amount the author cost to actually research and write the book (plus additional expenses), the cost of editing, layout, printing (when it is done), plus all the secretaries and others who help, I don't think that is such an expensive price. If cataloging systems were improved through importing records in other formats, there could be a tremendous time and cost saving.
So, I must ask: if this expense is too much, how much would the price have to drop to not be too expensive? What if it cost $20 per record? Or $10? I think library managers would think it was all too expensive and would only want $0.
But the real price continues to drop when you add onto this the possibilities for sharing records, this $30-$50 is shared among all the institutions who can use that record. So, if the Library of Congress catalogs a book and I take the record, it costs me zero. They can take a record of mine (thereby cutting in half our costs), and so on.
Today, the possibility fo
r cooperation in "metadata creation" are greater than ever. Publishers create metadata, international organizations, educational institutions and many others. Libraries and especially catalog divisions need to open their eyes to the new vistas of cooperation. In view of these considerations, I honestly do not think it is a matter of cost. It is a matter of people changing some work practices, and really cooperating. I've written quite a bit on this in AUTOCAT.
> There is too much
> stuff, and from the user's point of view, the payoff is too small. As
> more and more free text becomes available (think Google Books), the
> critical need for authority control and controlled vocabulary decreases.
There is a lot of stuff, but as I wrote above, when you add everyone involved in metadata creation right now (a lot of people around the world), it could be done so long as people were willing to cooperate and change. The real problem, in my opinion, would be selection. But it must be admitted that the technology for such a system could be built today.
Concerning the need for authority control decreasing, that would have to be demonstrated rather conclusively. This is not my experience at all--in fact, just the opposite. Just for one example off the top of my head, in Google Books (which I absolutely love because I can get very expensive books for nothing!), just try finding all the volumes of a multi-volume set. I haven't been able to do it yet. There is a large set of the works of Lawrence Sterne, and I can only find a couple of volumes of it although I know the others are in there. This should be the simplest task of any "aid to finding materials."
Also, known item searches almost disintegrates in Google Books. I can search for specific items and get a zillion hits, and it turns out I am getting the advertisements at the backs of the books or in bibliographies!
Other things I don't know if our users would stand for: if someone were researching the history of gays in the 17th-18th century, how would they find it in full-text without controlled vocabulary? You certainly don't look up "gays" or "homosexuals." You would have to actively come up with all kinds of words which would be the most slanderous today. How would a reference librarian handle that one tactfully?
At least with Google Books, everything should be fixable because almost everything is in a library, so there is already a record somewhere. All they need is to be linked up.
But perhaps you think that people no longer need or want to find materials by their authors, titles, and subjects, and they don't want help in determining which version of a work to choose? That users don't care that a search for WWI misses--by definition--all primary sources and is OK?
As I mentioned before, it is my opinion that our users do not understand that this is happening, and whenever I show people examples like this, their reaction is not, "Well, I don't need that anyway." Their reaction is, "Of course I want those things. Why does Google do that?" Then it becomes a complex explanation, but at least I have their attention for a few moments.
> Finally, providing access to content, IMHO, is only part of the issue
> when it comes to a "catalog". People have more access to information
> than they ever had. We are still drinking from the proverbial fire
> hose. Moreover, as people are able to carry around larger and large
> amounts of content on portable storage devices, the idea of library
> as container of information is becoming less relevant. In order to
> distinguish ourselves from other information providers, I advocate
> librarianship work to enable people to use the content they find in
> our libraries and out one the 'Net. In addition to providing access
> to materials also provide the users with tools making it easier for
> people to get their jobs done.
I guess I figure that if it ever comes to a time when people can carry the entire contents of the LC library and the Internet Archive around on their flashdrives, I don't think they will b
e able to find much on them without help. Information does not organize itself, it takes a human mind to do it and finding that information is going to be difficult, too.
I want people to use the net as well, as my experiment in my catalog demonstrates. But I also don't want to give up too soon on the very purpose of a catalog. At least keeping catalogs of different standards discrete gives the reference librarian a chance to do his/her work (LCNAF headings here; German headings there; controlled descriptors there; nothing controlled here), otherwise once it's all mashed together, nobody stands a chance.
Regards,
Jim Weinheimer
Received on Fri Aug 24 2007 - 03:09:28 EDT