WORDS OF ADVICE ADVICE CONCERNING LIBRARIES AND SITE MANAGERS

When I got into the bibliography business, over a decade ago, text-posting was a new thing. Sites posting texts (both html transcripts and photographic reproductions) were first being established, it was a period of initial experimentation, so it was very understandable that each site went its own way according to its managers’ ideas of how such a site ought to be operated, and that every site manager felt free to behave as a law unto himself. The situation was a kind of free-wheeling, “Wild West” one, with no agreed-upon standards or conventions. As I write this, in 2009, the number of text-posting sites, many sponsored by well-established libraries and other institutions, has multiplied and the number of available texts has increased, both to astronomic levels, and the availability of a large number of texts in electronic form has become an important feature of contemporary literary culture. But, to my astonishment, the degree of chaotic anarchy has scarcely decreased. While I can name a number of sites which are superbly managed in the best tradition of librarianship, many others fall short of these standards, sometimes to a jaw-dropping degree. I am going to mention some gross offenses against good practice, all of which militate against users’ interests, and these will no doubt strike some readers as impossibly exaggerated, but I could easily document the reality of each and every one of them. And if you rely on posted texts for your work, gentle reader, I can also assure you that your interests are affected by the failure of posting sites to observe good standards. So this is a subject about which you should care. Although your primary reaction should, of course, be a feeling of great gratitude towards anybody who makes texts freely available to you, when you perceive that you are being victimized by shoddy practices, and that your work is being impeded by them, you should not hesitate to make your displeasure known.

What malfeasances do I have in mind? In the first place, when one begins to visit text-posting sites, it quickly becomes evident that there is nothing remotely like uniformity in their structure and design. Nearly all of them are, to some degree, different and some are downright idiosyncratic. The result is that when one visits a new site, one is confronted with the necessity of figuring out how to navigate it and find what one wants (and this sometimes involves an exasperating waste of time), since some are considerably more “user friendly” than others. And occasionally a site will be radically redesigned, not always for the better, so that one must learn how to navigate it all over again. I am not urging any rigorous standardization of design, but in my work I have visited hundreds of such sites, and the varying degrees to which site designers adhere to good ergonomic principles is very striking. Some sites are a joy to work with, and one immediately feels at home. In the case of others, one has the feeling of being constantly engaged in a duel of wits with the site designer (and sometimes coming out the loser). Clearly, it would be in readers’ interests if sites developed some kind of norms or guidelines regarding design and structure. It is my suspicion, by the way, that some sites are designed, and some important policy decisions made about their management, by low-level technicians with inadequate supervision by professional librarians. If I am right, this is a sure-fire formula for disastrous results.

The single most important design principle involves informing the reader of what holdings the site makes available. Although some site managers appear to think that a Search function is by itself sufficient, some means for browsing the site’s holdings is no less vital a necessity than is a catalogue for a traditional library. Ideally, there should be two browsable lists, one of authors and the other of titles. And the availability of this browsing feature needs to be prominently advertised on the welcome page rather than stashed away in some obscure corner of the site, so that it is immediately accessible to the viewer. It is extremely frustrating to imagine that the people who maintain text sites lacking this feature probably maintain some sort of running list of their holdings for internal management purposes, but that it has not entered their heads that they need to share this information with the rest of the world. The absence of any kind of browsing or catalogue feature goes particularly far towards diminishing the usefulness of sites which contain a huge number of offerings: the larger the number, the more important browsing becomes (imagine the Library of Alexandria without Callimachus’ catalogue, and you’ll have some idea of the condition of Google Books and The Internet Archive). 

It is also necessary for site managers to grasp this seemingly self-evident point: as soon as they begin to post texts, people are actually going to read them and use them, and to manage their material in such a way as to respect this fact, making sure that readers are helped rather than hindered. They also need to understand that, when they post texts, they are making certain tacit commitments to their readers, which they are henceforth obliged to honor, and that they can reasonably be accused of unethical conduct if they fail to do so. And this immediately brings me to the subject of URLs.

There are two ways of presenting a site. The first is to assign a fixed, predictable, and permanent URL to each posted text. The second is to use a Javascript-driven “juke box” technology, so that each time a text is accessed, it is assigned a different and temporary one. The vast superiority of the former method at least ought be obvious, although to the managers of a discouraging number of sites it is, unfortunately, not. Individual readers are going to want to bookmark links to texts of interest. Scholars may want to cite URLs in their publications. Furthermore, in view of the ever-rising costs associated with traditional print publication, scholarly publication is destined to shift increasingly to electronic form. And, as soon as academicians begin to publish their research electronically, they almost automatically start to explore the possibilities of hypertext, with the result that direct links supplement or even replace traditional bibliographical references. All of this is facilitated by the assignment of unique URL to individual texts, but is rendered impossible by “juke box” technology. The assignment of unique URL’s to individual texts is, in fact, is just as much a feature of good librarianship as is the assignment of unique shelfmarks to individual physical holdings in a traditional library.

The key word in the preceding paragraph is “permanent.” Whether they realize this nor not, as soon as they assign a URL to a text, the managers of a site enter into a solemn relation of trust with their visitors. It is a strange thing that librarians who would not dream of tampering with, say, the shelfmarks of their manuscript collections (which in some cases have remained undisturbed for centuries), are capable of making arbitrary and capricious changes in the URLs of their electronic postings (and, when they do this, they rarely advertise what they have done), although changes in the latter wreak no less damage than do the former. I could name one very prominent text-posting site which has employed no less than three different and incompatible URL structures during the ten to twelve years it has been in existence. The very best sites advertise the addresses of their postings as PURL’s (Permanent URL’s), thereby issuing an iron-clad guarantee to visitors that they are regarded as sacrosanct and will remain unchanged. Such sites display a good understanding of the needs of their users, and ought to set the standard for the profession as a whole. Next best, if it is absolutely necessary for a site to change its addressing scheme, is to implement a system that automatically redirects users from old URLs to new ones, coupled with a prominently displayed acknowledgment of what is being done. Worst of all is to change URLs silently without providing redirection, and to do so in a way which cannot be updated by a simple global search-and-replace operation and require a separate manual intervention to update the URL of each and every item posted in their collection (and, again, a couple of sites have done just that.) When good principles in the handling URLs are violated in this way, violence is done to an important relation of trust with readers, and the establishment and maintenance of this trust should always be paramount in the minds of site managers. For this I guarantee: as soon as a URL is posted, it will be used, and readers need be able to rely on its continuing validity.

The concept of permanence, of course, goes deeper. Posting a text likewise involves an implicit solemn promise to the reader that the text will stay posted. But on some sites texts mysteriously disappear without any acknowledgement of their removal. Even entire sites vanish without explanation. Some text sites are maintained by private individuals, as labors of love. One feels great gratitude and respect for the individuals who maintain such sites, but at the same time one cannot help cringing at how short-lived they are, in all likelihood, destined to be. To speak very much about the long-term archiving of electronic scholarly material would take me too far off-subject, so suffice it to say that no site is very likely to enjoy great longevity if it does not have institutional sponsorship. And once an institution sets up or sponsors a text-posting site, it is, in effect, assuming a responsibility to keep it available on a long-term basis.

I am highly conscious that, although I am a professional scholar I am a very amateur librarian who has no business dictating rules to the professionals. But I would be so bold as to insist to librarians that the electronic reproduction of texts has become such an important function performed by modern libraries that the present “Wild West” situation needs to come to an end. Detailed industry-wide uniformity of structure and design may not be necessary or even desirable, but general standards of good procedure and some kind of code of ethical behavior need to be developed and observed by site managers, so that the greatest good can be derived from them, with the least possible harm inflicted. And, clearly, this development needs to be a collective effort. Electronic postings, surely, deserve to be treated with the same systematic care and respect that is shown towards physical holdings as a matter of course. I have no idea whether the management of text sites is yet formally regarded as a branch of library science, and taught (or even thought about) in the schools that provide instruction in that discipline. If not, it should be, and I respectfully suggest that it is high time that librarians begin talking to each other to develop a set of industrywide standards and ethics, for the better maintenance of such sites and to guarantee the good progress of the scholarship that depends on them. This will entail the development of some kind of “shame culture” in which errant site managers can be reformed as the result of their peers' disapproval, and the vocal reaction of readers when their expectations of a reasonable level of quality are not met. For the development and observations of such standards is not the exclusive business of librarians. It is the right and responsibility of every scholar who relies on posted texts, and also of the general reading public, to insist that sound managerial practices be developed and followed.