I read Arms' paper with some interest. Ultimately I'm getting from it
that he is of the opinion that such automation appears to be inevitable,
and that that's a good thing. In general, I find myself agreeing, but
it's worth considering a couple of things on our headlong flight into
automation, one of which Arms touches on and the other, related, that I
fear might blow up in his face if he's not careful.
First, personalised service - there is no substitute for the immense
amount of knowledge and experience humans bring to a task, even if not
in their own area of expertise. Librarians are skilled negotiators,
searchers, collators of information, deliverers of useful knowledge, and
often the glue that sticks communities of researchers together. An very
worthwhile example of this is given in Nardi and O'Day  in their
chapter concerning a case study on human librarians. Despite the fact I
work in the area, I can't see agents or intelligent interfaces replacing
this for some years to come yet, no matter how good they get.
Second, universal access to information. It's quite possibly an
oxymoron, and a dangerous one at that. Perhaps I read too much into
Arms' statement that "automated digital libraries combined with open
access information on the Internet offer to provide the Model T Ford of
information. Nobody would claim that the Model T Ford was a peer to the
handcrafted cars of its generation, and automated digital libraries
cannot approach the personal service available to the faculty of a well
endowed university. But few people could afford a hand-built car, and
few people have easy access to a major research library. The low cost of
automated digital libraries is already bringing scientific, scholarly,
medical and legal information to new audiences." While this is
superficially true, it's also true that the information gap is growing
rather than shrinking - whole new audiences are in fact being shut off
from this information. It's arguable, perhaps, that they never had it in
the first place, but I beg to differ - most, if not all, moderately
sized towns in this country (Canada) have at least one library (and the
same was true of when I lived in the UK). I doubt that the US is any
less endowed with such institutions.
Granted, people mostly use libraries at very superficial levels, but
even in the smallest public library the librarians can be of immense
help in accessing information, and don't forget the inter-library loan
service quite effectively removes barriers or geography (many of the
works I request come from the British Library, a mere 3000 miles
away...). What the Internet has done is put a lot of information out
there, much of it suspect, into the hands of people who don't
necessarily know the difference between good and bad advice or
information. It's hard to know, therefore, if the electronic information
haves are better or worse off than their have not counterparts, but
that's a debate I leave to another day.
Bear in mind, however, that we risk finding ourselves in a worse
situation than existed before the explosive growth of the Internet,
perhaps reminiscent of the Industrial Revolution, with a large
underclass who don't have all our perceived advantages. To assume the
systems Arms proposes will solve that problem is naive at best.
One other thought strikes me - librarians are the very people we need to
help us design these systems - they know their job, they know how to
organise the information properly, how to talk to their clients and get
the best results for them, and they have several hundred years of
experience to draw on. Don't discount them so easily - I certainly
National Research Council Canada
Institute for Information Technology
(The opinions expressed herein are the author's own and are not representative of those of the National Research Council of Canada or the Government of Canada.)
Nardi, B. and O'Day, V. "Information Ecologies: Using Technology with
Heart", MIT Press, April 2000.
To the Editor, D-Lib Magazine:
I would like to make a few comments on the article by William Arms, Automated Digital Libraries: How Effectively Can Computers Be Used for the Skilled Tasks of Professional Librarianship? in the July/August 2000 issue.
First, I want to make clear that I agree with Mr. Arms in his assertion that libraries need to change -- in fact, they have been changing for some time at an almost breathtaking pace in some areas, as novelties in automation are introduced alongside major changes in management practices and staffing levels. I believe that many of these changes are desirable and necessary, and I am an enthusiastic supporter of change in libraries. These changes must be made in an effective and thoughtful manner however, and with complete knowledge of what the goals of the library are and what librarians currently do.
Non-librarians rarely have an effective knowledge of the goals of libraries or of the tasks of the librarians, themselves. Mr. Arms' essay is a fine example of these misperceptions about the library that he shares with many others. Perhaps Mr. Arms' final conclusions are correct, or perhaps they are not--I do not want to get into that here, but his ideas of libraries are misinformed. While there are many places where this is apparent in his article, I shall restrict my criticisms only to those points I feel are most important.
1) "The libraries at Harvard employ a thousand people and the Library of Congress more than four thousand."
These are the only examples Mr. Arms gives of staffing levels, but they are certainly both unique cases in the library community. Harvard has a unique, decentralized library system which in effect, contains several mini-libraries, each with its own administrative organization, catalog, human resources, etc. The Library of Congress is another entity truly unique unto itself. Academic libraries have far fewer staff, normally numbering less than 90 professionals. When support staff are added, the average increases to 240, and adding in occasional student help goes to 313. At Harvard, when student help is added, the total number of staff rises to 1240. (Taken from: Association of Research Libraries Statistics available at: <http://fisher.lib.Virginia.EDU/newarl/>) These numbers have already declined substantially in the last decade or so. When considering that the library must serve an entire campus community of diverse interests, many physical locations, along with visiting scholars of discriminating interests, current staffing levels do not seem quite so outrageous.
2) "Google is more up to date than Inspec, its coverage is broader and its indexing records are good enough for me to find what I am looking for."
Mr. Arms commends Google searches as opposed to the library catalog. This misapprehension of the question is often encountered among non-librarians.
One must be aware of the purpose of the library catalog. Its purpose is to allow users to find what is in the collection when the author, the title, or the subject is known. We must keep in mind that this is a completely different task from a web search engine. It means that when I do a subject search in the library catalog on "dogs," it is designed to help me find everything that is in the collection about dogs, along with some very nice cross-references to other interesting terms I may have never even considered. Or, I can find everything in the collection by Leo Tolstoy.
Of course, this is the goal of the catalog, and it is rarely--if ever--achieved perfectly. Libraries are long-lived institutions (Princeton has been around for 250 years) and the records made today should mesh well with records created from long ago. Important factors that make this a difficult goal to reach are: historic changes in staffing levels, the introduction of new practices and new formats, changes in rules, staff morale, and so on.
I'm sure that Mr. Arms does not believe that when he searches for his topics in Google, he is finding everything on that topic. Google, and other web search engines aim for customer satisfaction (these results are good enough!) instead of showing the fullness of what is available to the user. It is difficult to determine if something is good enough if one has no idea of the whole. Finally, items in Google have been created only in the last few years, whereas in the library, a user is not limited to such a small fraction of materials.
3) "Google ranks web pages by how many other pages link to them. It gives greater weight to links from higher-ranking pages."
Basing the importance of documents on the number of times an item is cited is a compelling, yet obviously highly dangerous argument. Taken to an extreme, the conclusion would be: #1 is best, #Last is worst, and if this is so, who would even consider looking at #Last? Most people would prefer to have selection performed by an expert, and not by the world in general.
"As a result, Google is remarkably successful in presenting a user with the most important page on a topic..."
The most important page should not be confused with the most popular page. What follows is a case in point. Everyone can point to examples of lazy authors citing errors that tend to replicate themselves. One of my favorite examples is the pseudo-character "Nicolai Lenin," the revolutionary who created the Soviet state. In reality, Vladimir Ilich Lenin never went by the name Nicolai Lenin. Some early journalist probably used "Nicolai" in error and it was repeated in many school books, that I know of, until the 1980's, if not continuing to the present day.
If this were the situation in Google now, a search for Nicolai Lenin would retrieve only the poorest information, while the scholarly information would be completely unknown to the user. A search in the library catalog, however, shows the reference:
Lenin, Nicolai, 1870-1924
Use: Lenin, Vladimir Ilich, 1870-1924
When users see this, they have learned something in addition to getting the information they want.
4) "Attempts to catalog and collect web materials using skilled librarians and archivists have floundered on the scale of effort needed to do even a rudimentary job."
Libraries have never collected everything ever printed on any topic--even exhaustive collections decide not to include some items. Some people consider the process of selection to be a waste of time--if not a positive evil--while others consider it to be a valuable service.
I agree that modern cataloging is a time-consuming effort [which has undergone much simplification] but we should not discount that lack of control also demands effort that consumes a lot of time. Of course, this is the time spent by the users plowing through masses of information they don't want or need, while they miss things of great importance.
If the labor of a cataloger is seen as valuable, and the labor required to create consistent, easy-to-find access is, say, 30 minutes, and this record can be used instantly by millions of people around the world to easily and quickly find relevant items, one must ask: is this an inefficient use of labor? If this record were to be useful for decades or even longer, I ask again--is this an inefficient use of labor?
5) Mr. Arms example of the reference librarian helping him find certain data he needed is indeed very heartening. Still, it must be understood that, as he states,
"There was nothing magical about the methods that she used. "
True, but she also represents only the tip of the iceberg. She could do her job perfectly, but if the rest of the people hiding behind the scenes failed to do their jobs properly, all of her techniques would founder. The reference librarian represents the last link in a chain--and if that chain should break anywhere, she could not do her job properly.
"She simply had more expertise in the idiosyncrasies of the information available and how to navigate through it."
One must avoid the fallacy of believing that if the library did not exist or do its work, these "idiosyncracies" would disappear. The user would discover the millions of other idiosyncracies discovered by expert librarians who smooth these over for the user. Those who believe that searching Google or another web search engine is any less "idiosyncratic" than searching the library catalog are only fooling themselves.
6) "... the tools available to the user are sufficiently good that most searches can now be carried out directly by the user."
This has been the case for some time. Librarians have always used the same tools as users have, with the exception of some highly-expensive database searching. Now, many of these databases have become more generally available, even such databases as Lexis-Nexis and Medline. The need for expertise in searching has not disappeared merely because anyone is allowed to search these tools without training. In fact, many of the "power searches" used by expert librarians often are not available in the web versions (in a market-driven economy, there is less reason to cater to a very small minority of customers). [For an interesting discussion of these issues, see: Design wise : a guide for evaluating the interface design of information resources / Alison Head. -- Medford NJ : Information Today, 1999.]
My comments against web search engines are not meant to argue against their existence--on the contrary, they are one of many tools that can be used for information retrieval. Sometimes the results can be excellent (look at the number of URLs I cite in this letter!), but more often the results are spotty and semi-coherent at best. It also guarantees that users will waste enormous quantities of time examining items of little or no interest, while they remain sadly unaware of the scope of the resources available to them.
I agree with Mr. Arms that libraries must create more access to their collections and catalogs. They must find ways for users to more efficiently, and more independently, use the tools so carefully created by the libraries. Today with internet search engines, there is no doubt that the vast majority of web users will not ask questions of a reference librarian if they encounter a problem--they'll just click on something else. There is a danger that if libraries refuse to change their traditional methods of information access, they will be completely ignored in the information revolution.
Change is imperative, and if it is fated that librarians must go the way of the trilobyte, as Mr. Arms suggests, then so be it. I believe the library profession is adaptable enough that it can evolve to other--perhaps higher--levels if need be.
This should not occur without full congnizance and appreciation of what libraries and librarians do now, however. Search engine results that are designed to make us happy by finding something successfully (This is good enough!), with displays that silently lead us to the most popular materials may ultimately be found satisfactory for the rigors of academic research.
Or perhaps not. Perhaps people will demand some kind of search engine that gives the user an idea of what is in the collection based on the user's interest, and that gives the users more power over what they see.
It is interesting to note that companies are coming to the conclusion that full-text web searches and single forms of arrangement are not enough. There has been a great deal of interest in this subject shown in a new ASIS special interest group for Information Architecture, where I moderate its listserv, sigia-l.
If you're interested in another side of this debate, you can join the list at:
For those who are interested, I have suggested a system that I believe retains the strengths of librarianship, but allows for the tremendous changes introduced by the information revolution. It was published in Vine magazine, How to keep the practice of librarianship relevant in the age of the Internet [http://agent.sbu.ac.uk/publications/vine/116/article216.html], while an online version of my article is at: <http://www.princeton.edu/~jamesw/metadata.pdf>.
James L. Weinheimer
Aug. 31, 2000
Letters concerning stories will be forwarded to the authors, and the authors' responses will be published with the letters whenever possible. If you prefer, you may request that your letter not be published.