Péter's Digital Reference Shelf
November 2003

Title: Xrefer Plus
URL: http://www.xrefer.com
Publisher: Xrefer
Cost: Dependent on the number of titles chosen and the type and size of the subscribing institute.
Tested: Oct. 8-12, 2003

Xrefer has come a long way since its debut in 1999, at least in the United Kingdom where many academic libraries subscribe to the Xrefer Plus service. In the U.S. it is still not a household term, but it's certainly better known than it was in 2000. Now the British digital publisher of mostly British reference works has launched a campaign in the U.S. market for its subscription-based service, and has among its clients the City University of New York, Simmons College, the University of Chicago and the New York State Library.

Two years ago in this column I highly praised the free version of Xrefer, which had more than 50 reference works. (Oddly, the home page of Xrefer Plus incorrectly claims that the free version had only 30 titles.) The free version was expected to be temporary (lasting for nearly three years) in order to demonstrate the power of cross-referencing entries across the entire collection (more about that later).

THE CONTENT

Xrefer Plus can have many incarnations depending on which sources the subscriber chooses beyond the core collection. There is, for example, Xrefer100 and Xrefer125, alluding to the number of sources subscribed. As the number of sources increases, there certainly will be additional packages and, hopefully, there will also be a smaller core collection for those institutes that would like to benefit from the aggregated and integrated collection, but at a lower entry level of, say, 50 titles. The complete title list will give you a feel for the variety of publishers and reference titles.

Wide Assortment and Customization

It is a good idea to let subscribers customize the package for several reasons. One is that there are many dictionaries and encyclopedias on the same subject in Xrefer Plus. While it is beneficial to have access to two to three health encyclopedias, only the largest medical libraries would need (or could afford) all 10 medical dictionaries offered by Xrefer Plus, no matter how mouthwatering that sub-collection is with titles ranging from Black's to Dorland's and Stedman's respected dictionaries.

Another reason for the need for customization is that some of the encyclopedias and dictionaries have been available free of charge on the Web and are likely to remain free, such as the Columbia Encyclopedia (hosted at least by six different web sites), the Hutchinson Encyclopedia, the Collins English Dictionary and the Collins Spanish-English/English-Spanish Dictionary, to mention a few that I have reviewed before.

My evergreen free favorite, Bartleby.com, alone has some of the same reference works as Xrefer Plus, such as the Columbia Encyclopedia, American Heritage Dictionary, Brewer's Dictionary of Modern Phrase and Fable, Roget Thesaurus II, Simpson's Contemporary Quotations and the Respectfully Quoted collection. Some of the titles have links to other titles in the collection, such as from the Columbia Encyclopedia to the CIA World Factbook, but these are not reciprocal, let alone universal, among the components of the Bartleby collection as they are in Xrefer Plus.

The Issue of Free Alternatives

It is important to elaborate on and illustrate the point of the free alternatives that appear in Xrefer Plus because this will be a cardinal issue among librarians who will make the purchase decision.

The free versions of some of the works included in Xrefer Plus often stand alone, do not search the full text, are sometimes less current than in Xrefer Plus and, in case of the Bloomsbury collection, are misleading and utterly frustrating due to the brain-damaged software used by the publisher on its site. In my review two years ago I pointed out many of serious deficiencies of the Bloomsbury software, coupled with lack of quality control of the digital version. The pathetic software of Bloomsbury still cannot find the entry about cinquecento in its own implementation of the Guide to Art when searching for 16th-century Italian art, nor can it find the bibliography of Tintoretto, a well-known artist of the cinquecento. Xrefer Plus retrieves both these entries in half a second and lists them as the top two hits.

Of course, by Bloomsbury's software standard it is progress that it no longer repeats each entry twice and finally shows four hits instead of eight as it used to.

The Bloomsbury software also misses many articles. If you look for an article about degenerate art, it finds nothing even though it has a cross-reference from one of its entries to the substantial entry. Xrefer Plus finds seven articles for this term in addition to the cross-referenced entry from the Guide to Art and 21 additional articles from its other resources.

The Bloomsbury software remains not only mentally challenged, but also visually impaired and misinforming. It still encourages you to "click the blue heading to see the full entry" but as I noted before, "You can look for a blue heading until you're blue in the face because the headings are displayed in red." Xrefer smartly highlights the matching terms with yellow. I wonder when someone at Bloomsbury will get red in the face seeing this incompetent and careless digitization project?

The Hutchinson Encyclopedia also has a free version hosted by Tiscali. Maybe the programmers of the Bloomsbury site were involved in its implementation as the symptoms of carelessness are very similar. There are many ways the Tiscali software cripples the excellent content of the Hutchinson Encyclopedia, the most important of which is dropping the accented characters or replacing them with some weird symbol, making articles about a large number of French, Spanish, German, Portuguese, Scandinavian and Hungarian people and places irretrievable because in these (and some other) languages diacritical marks are extensively used.

When such articles are retrieved by other search criteria that do not include characters with diacritical marks, they look as appealing as punk-rock fans with melting make-up after partying all night Xrefer Plus handles accented characters with aplomb and allows searching either with or without the accent mark and displays the accented characters correctly.

It's no wonder that Xrefer Plus finds 34 articles in the Hutchinson Encyclopedia about Paul Cézanne, while Tiscali retrieves none because the painter's name appears as Czanne everywhere. Thinking about the many personal and place names that include á, é, í, ó, ú, ñ, à, â, ã, ä, ê, ô, õ, ö, ø, ç, õ, œ and š, you get a good idea of the dimension of the problem in the mutilated free Hutchinson version. This may not sit well with Tiscali users in several European countries who will not be pleased when the names of their famous scientists and artists or of their favorite soccer teams and players are not retrieved from the entire Tiscali site. This blunder is not easy to detect until you accidentally find the good entry about Dürer that appears as, and needs to be searched for as, Drer. In Xrefer Plus it can be searched as durer or dürer (uppercase or lowercase) and is displayed perfectly.

Another free version of an Xrefer component should not be taken at face value, either. You may believe that an English dictionary of Penguin is freely available on the well-known publisher's site, until you click on its logo. It takes you nowhere and always times out with an error message. The original Penguin offering, the Penguin New English Dictionary, was the same and is now part of the Xrefer collection. It is a good resource, but it was replaced by the Penguin Concise Dictionary and the programmer forgot to turn on the links. You would think that Penguin could afford to hire a programmer who can get it to work, but apparently not, and my e-mails about this abysmal situation have gone unanswered. While I am a fervent proselytizer for open access reference works, Bloomsbury, Tiscali and Penguin give a bad name to open access. When comparing free sources with their counterpart in Xrefer Plus, such features should be considered.

British Slant

Most of the titles in the Xrefer collection are of British origin, but it does not preclude its success in the U.S. Actually, they provide a fresh angle. With that said, some of the titles that are quintessential in the United Kingdom may be a hard sell in the U.S. market, and therefore should not be part of the core collection. The Xrefer page about the publishers duly extols their virtues and for Debrett's it claims that it has "recorded some 100,000 prominent people over the last decade, the meritocracy."

If this refers to its Peerage and Baronetage volume, it is perhaps more aristocracy than meritocracy, and the typical American users may not get excited to learn about such noble things as "one hereditary peerage has gone into abeyance, but two have come out of abeyance." I am personally relieved by the positive balance and although it will unlikely be a question on the many American TV game shows or in public libraries. (In case you are interested, the baronies are: the Barony of Adley for the former, and the Baronies of Arlington and Berners for the latter. Now you know.)

Some of the reference works have informative and unique illustrations that are reproduced in Xrefer Plus, such as the one in the various Andromeda publications, like the Atlas of the Bible. You will also find 17,000 reproductions from the Bridgeman Art Library.

All in all, Xrefer Plus has a rich assortment of high-quality and mostly current reference titles. Some of them fill a serious gap in digital ready reference where there are no free alternatives and very few high-quality competitors, except for the Oxford Reference Online collection. Its sub-collections in science, business, health, humanities (including what in the U.S. is called social sciences) and religion are particularly useful for the quality dictionaries from a variety of respected publishers.

The only genre that I badly miss is the almanac, specifically the Chambers Book of Facts and the Whitaker Almanack, which are fine sources with British focus. Both publishers are partners of Xrefer in publishing some of their other titles so there is a chance that they will be included. I will review in details some of the Xrefer titles later, but in the rest of this review I would like to focus on the Xrefer software that adds much value to this good collection.

THE SOFTWARE

Xrefer has a capable search engine and an intuitive interface that allows for exact phrase searching and Boolean operations in the entire body and in the entry header. When browsing in each resource, the list of header terms is a valuable feature as some users like to wander around in the vicinity of a specific term.

The help file is informative, but needs some clarification. For example, when it discusses phrase searching, it claims that "searching for 'Winston Churchill' will return a great many results. Not all of these entries will be about Winston Churchill, but his name will be mentioned within the body-text of an entry. To filter these entries out, and ensuring that only entries about Winston Churchill are returned, change the setting from 'full text' to 'headings'. This will then only return entries where Winston Churchill is mentioned in the heading".

This is actually not good advice as the search will be implicitly limited to those 68 items where Winston is followed by Churchill in the header of the article. This arrangement happens only in the quotations dictionaries, in all the other sources, about 80 articles, his name appears as "Churchill, Winston."

This is an unfortunate example, anyhow, not that I don't want to see his biographies or sarcastic quotations, but because there is an American novelist by that name (although without the Spencer middle name and the Sir title), and his entries are interspersed with those about the statesman. It is unavoidable. Therefore, choosing a simpler example for the help file should be done to avoid confusing users when they turn for help.

A much better alternative for the help file example would be Laurence Olivier limited to the header field in an AND relationship rather than as exact phrase. It would eliminate the need to explain the use of reverse name format (the comma luckily does not matter) and retrieves five more articles (all of them relevant) than "Olivier Laurence."

The software also offers a "spelt & sounds like" option, which certainly will appear in "reviews" that are based on press releases. While it sounds good that "searching for '~psycologist' will return results for psychologist," this search option, also known as Soundex Search, is less perfect than it appears. The query will indeed retrieve psychologist, but only as the 26th entry following such entries as oenologist/enologist, mycologist, phycologist, cytologist, gynaecologist/gynecologist, serologist, mixologist, sinologist and a few others.

I am hopelessly unauthentic in pronunciation, but I don't think that any of these words sound or (with the exception of phycologist) are spelled more like the query term than psychologist. The relevance ranking does not help and an alphabetic listing would give a better chance for the users to efficiently locate the sought after term (as the first character is much less often misspelled than the third and fourth). The relevance ranking could be much improved by comparing matching number of letters in the suggested term versus the query term.

Relevance ranking is the default in presenting the results of any search, but you may switch to give preference to entries with images. This is a good idea and is well implemented. Displaying the number of words in the entry, or using it as a sort criterion in the result list would also help users in deciding which entry to examine.

The two best features of the software are the display of the list of related cross-references and the mapping of the terms. When searching for Honolulu in the entire collection you get a result list of 172 entries — don't ask me why Honolulu is a female noun in Italian but a male noun in Spanish. On the side of the result list you can see five of the terms to which Xrefer Plus generated a cross-reference.

You may display all the additional xreferences, in this case 18 entries from the variety of sources, on a separate page. These are the most closely related terms to the query, based on the "see" and "see also" references in the various dictionaries, encyclopedias and gazetteers aggregated and integrated in Xrefer Plus. This is an effective tool, as it reduces the spurious hits of full-text searching.

The mapping feature visualizes which other articles may be related to the one used in the query. It takes a minute or two to create the map, but you see it progressing until the chosen number of 100, 200, 300 or 400 results are analyzed. Hovering over the map with the cursor shows the title of the entries that are related. If you click on one of the entries it will also display the sources. This is a promising tool that you can try out by going to the demonstration site.

Xrefer Plus brings valuable and new resources to your digital ready-reference desk with innovative software. It is a very good resource for large public and academic libraries. The variety of dictionaries and encyclopedias complement each other well, and often provide more well-rounded results than even the best single resource could, as I illustrated in my earlier review of Xrefer. It also offers the possibility of comparing entries in order to corroborate facts, an essential element in quality reference service, that is much facilitated in digital reference services, as you can see from my PolySearch Engines.