Slowing Down the Speed-Up: Library Standards
By Walt Crawford
American Libraries Columnist
Senior analyst, Research Libraries Group
Column for May 2002
Why do library standards take so long to adopt and implement? With the torrid pace of technological development, how can academic (and other) libraries stand for such ridiculous delays? Once something’s feasible, why don’t we have it now?
Volunteers, consensus, progress
True open standards—like Z39.50 and the ISO ILL protocols and unlike PDF (Adobe’s proprietary standard)—inherently take time to refine, adopt, and implement. In a field where most participant organizations are relatively small and many are nonprofit, it takes longer.
- Open standards require open development and consensus. That takes time—time for balloting; time to respond to comments and objections; time to ensure that the wording is clear, that the standard is implementable, and that it does not favor one company or party over others. Even so-called fast-track standards (such as OpenURL) take a year or more from proposal to adoption.
- Most groups involved in library standardization don’t have full-time standards representatives or unlimited travel budgets. Typically, committees for particular standards can only meet a few times a year to avoid overburdening the participants.
- Once adopted, the most important standards require continual refinement and publicity before widespread adoption. The more technology changes, the more some standards require maintenance.
- Complex multiparty standards require complex multiparty testing. Ask any participant in the ILL Protocol Imple-mentors Group about protocol testing; ask early Z39.50 adopters about the steps required to assure that implementations worked properly. Ideally, that testing should take place before you begin using the new standard—unless you appreciate being guinea pigs when critical library data is at stake.
- It’s easy to overinterpret what a standard actually does. Z39.50 allows computers to exchange searches and results; it does not and cannot assure that every database offers precisely the same searches carried out in precisely the same ways and yielding precisely the same kinds of data. Such overall uniformity would reduce all systems to an unsatisfactory lowest common denominator—but without it, broadcast Z39.50 searches (for example) may never be as effective as searches against large centralized union databases.
- It could be worse. MARC II (the first fully realized MARC format) emerged in 1968; the underlying standard Z39.2 was adopted in 1972; but it was not until 1984—a dozen years later—that nearly all library automation vendors showed reasonably full understanding of and compliance with the format.
Look in the mirror
It’s not just the slow pace of standards development. Take a look around, in your library and your institution. Do all your Internet-connected computers run Internet Explorer 5.5 or higher, or Netscape 6.2 or higher? Can you assure high-bandwidth, low-latency connections for all users? Do all the computers run contemporary, high-reliability, true multitasking operating systems—Windows XP or 2000, Mac OS X, or Linux?
If not, your systems are behind the times—enough so that some supposed standards may cause problems. For example, if you’re running Netscape prior to 6.2 or any version of Opera to date, full Unicode display is out of reach. How would you react if RLG (or OCLC) said, “From now on, Eureka (or First Search) requires Windows XP, Internet Explorer 6, and Pentium 4-equivalents running at 1.5GHz or above with at least 256MB RAM and displays showing at least 1024x768 resolution—and we assume everyone has Fast Ethernet or better. We have to be able to support contemporary standards, after all!”
It takes time for libraries to replace systems, just as it takes time for bodies to define and vendors to implement standards—for some of the same reasons.
When one agency has absolute control, things may move fast when that meets the company’s goals—but in ways that rule out competition and its advantages. When many agencies agree on common definitions, the process takes much longer, but the results serve most libraries and users better in the long run.
It takes so long because we’re human, because we have limited resources—but also because NISO members and others involved in standards bodies try to do things right, so that the results level the playing field and improve the future for everyone involved.
|
