Péter's Digital Reference Shelf
November 2002

Title: Ulrich's On Disc (New Version)
Publisher: R. R. Bowker, LLC
Cost: About $1,000 for single user
Tested: October 20-25, 2002 (preliminary version)

Bowker is now part of the Cambridge Information Group, one of the most dynamically developing and innovative players in the league of professional information services. This review looks at the beta version of Ulrich's's on Disc, released in August. The final version is expected shortly.

The Content

Ulrich's got a new and improved interface for its 70th birthday and some of the wrinkles in its content were fixed. Still, it has some significant content deficiencies that the help file does not warn you about, and librarians just don't know about.

Almost true to its advertising blurb, Ulrich's indeed added 9,516 records to last year's 248,340, making this current edition 257,856 records. Bowker still modestly refers to it as "nearly 250,000 records," or the copywriter just forgot to change that line in the blurb.

In the past, one good thing about Ulrich's was that it corrected the errors of commission that I pointed out in my reviews and conference presentations, or least removed the absurd data, such as the more than 7-million subscribers at $35 a pop to the US Surf magazine. Lately, I don't see this happening consistently.

Bowker still claims in the record about its own directory that CODEN is a bibliographic classification code along with LC and DDC. It is akin to a biology teacher claiming in his biography that his favorite bear is the koala.

And while some of the factual errors have been corrected, Bowker still believes that a subscription to the Elder Law Advisory of the West Law Group costs $13,095. I did not expect them to believe me, but I've shown them the publisher's Web site so they could find the current price -- 10 percent more than a year ago, but still a reasonable $270 -- not the obviously absurd price that Bowker keeps quoting. These errors of commission are less dangerous than the errors of omission that by definition are not visible.

For example, the information about the coverage of abstracting/indexing publication remained woefully inadequate. In the area of energy science and technology, Ulrich's claims that the printed version of Energy Research Abstracts covered 41 journals before it ceased publication in 1995. In actuality, it covered many more. Likewise, it pegs the coverage of Energy Abstracts at six journals and its price at $975. This is very unlikely data, and so is the information about journal coverage by A/I services.

Errors of omission

You may say that Bowker does not care too much about the CODEN error, as it is still available for only about eight percent of the records. And in the past year only two more records were enhanced with CODEN.

It is a much more serious limitation that less than half of the records have circulation information and even fewer have searchable price information (which is in U.S. dollars). These are much touted points of the digital edition and many librarians happily go about searching by such criteria, not realizing that either of these searches automatically and silently exclude more than half of the records. If you combine these two criteria, you limit your domain to 52,784 records. Don't even think about searching by LC classification as it is available in only 17.9 percent of the records.

If you combine LC classification with the circulation and price data, you are searching in a 17,055 title domain, possibly without knowing. Academic librarians should definitely know about these limitations, but any librarian must bear in mind that a simple query, like finding all the titles that cost less than $1,000, is not an easy task. True, it is not easy for Bowker to get that data, as there are many book series whose prices vary from volume to volume. Also, there are print periodicals, such as H. W. Wilson's Book Review Digest, whose subscription price is based on the library's budget and the anticipated use of the publication (also known as service-based pricing). The point is that Bowker should explain the serious limitations of price searching instead of just saying that "not all records include all fields." That is a gross understatement.

The same is true when you want to find the journals with the highest circulation in a specific discipline. This is exactly what I would have needed in a modest research of correlating the impact factor of journals with their circulation. For example, Elsevier is by far the leading publisher in energy science and technology. No wonder that one-third of the top-ranked 66 journals in the Energy and Fuel category in the latest edition of the Journal Citation Reports is Elsevier property. Too bad that many of them have no circulation data in Ulrich's, leaving you helpless. The help file for this field says that "searching on the circulation category allows you to limit the results of a search to serials that have a certain level of circulation." I am afraid that you will limit your search much more than you wanted if you ever search by circulation.

As for the language in the help file, this is the way they taught us to put it in the law school, but librarians deserve better. Ulrich's contains tens of thousands of journals that have a substantial level of circulation, but neglects to provide supporting information. Nor do they set a minimal limit as a prerequisite for including this data.

I tested this new version with an actual project on my plate. I checked the top 140 journals in terms of indexing coverage by the Energy Science and Technology database in 2002. Ulrich's only had circulation information for 75 of them. The ratio was far worse for the most important journals, the ones with the highest impact factors. Elsevier had 22 journals on my list, six had circulation information. Its imprint Pergamon Press fared somewhat better with seven journals having circulation data from 16 of its titles. North-Holland, another Elsevier imprint, had three titles, none of them having circulation data. John Wiley had 11 journals, five of them with circulation data.

The Magic Links

The best thing in Ulrich's is that it provides links to many of the journals. Again, my actual project served as a testbed. The good news is that for 95 percent of the journals a record could be found, if not through the ISSN then through title words. Sixty-six percent of the records had good links directly to the journals. An additional 15 journals had links, but these led not to the journal, but to the publisher's or sponsor's home page, requiring searching and clicking to find the journal.

Probably the worst of those links was the one for Physica Scripta which took me to the home page of the Kungl. Vetenskapsakademien. I don't chicken out seeing a foreign term, but this was pure Swedish, and until I found the English button, I was not happy. Even when landing at the English home page I needed some orientation to get to the journal's page through a few clicks of trial and error. The funny thing is that the URL for the journal's home page is as simple as it gets: www.physica.org and that's exactly what Ulrich's should use, tack så mycket.

Finding nuclear fusion through the link which takes you to the home page of the International Atomic Energy Agency may trigger some confusion, but it is better than the completely incorrect links, such as the one for Our Planet which was a dead end. The correct URL could not be more simple: ourplanet. Decent browsers will even add the "http://www" prefix and the ".com" suffix and zip you there. Luckily, in my sample only six records had wrong URLs. Overall, this aspect, in spite of glitches, is the best improvement in Ulrich's content. It is another question that Bowker should use these links to get current and accurate information about journals that are described by the publishers on the Web. I could easily find the ISSN, circulation and pricing information about journals where Ulrich's just drew a blank.

The Software

The software is very fast (except in re-sorting long result lists), and quite intuitive -- much better than the previous version. Probably its most appealing feature is that while you are typing a word in the query cell, the program looks up the term in the title, publisher, subject, etc. index -- depending on which cell you are typing the search term in. For example, when I typed "library" in the subject cell of the query template, as soon as I entered the first two letter the software showed the appropriate subject heading. This shadowing is perfect to lead users into looking up the indexes that otherwise they would not do, and picking the term from the index and search for it in one fell swoop, rather than monkeying with the spelling of some difficult words.

It would be important to allow users to add another term from the index to the same cell, but the new term always overwrites the incumbent word. The user should be able to edit queries after their execution, unfortunately the software deletes the search template after it finishes searching. As a result, by simply adding truncation to the end of the name of the publisher, I had to retype the entire query combination, an irritatingly inefficient step. While it is fine that you can store and reload queries, they are automatically executed, not giving you a chance to make a change.

Batch searching has been enhanced. Beyond searching for a batch of ISSNs, you can now search from a simple text file that may have hundreds of journal names or publishers' names that you downloaded from some Web site. I liked this feature even when it was restricted to ISSN (which is still the most likely application simply because of the standard format of ISSNs), so I like the enhanced version even more.

There are other customization options, which is always welcome, such as defining your own result list format, which can be very important for those who do bibliometric research and can add selected data from Ulrich's by downloading a result list of hundreds or thousands of journals with their choice of data elements, then import them into a spreadsheet for further processing. That was exactly what I did to add some Ulrich's data to my spreadsheet of other data collected for the journals uniquely identifiable through ISSNs.

What I said in my earlier review is still mostly true, and unless Bowker systematically cleans up the database and adds LC classification codes to the most important records of active academic journals, it will face increasingly dangerous competition from groups of smart librarians who are developing free journal directories of very high quality, and exuding the breeze of innovation, like jake, which is my other reviewed product of the month.