Despite stumble, GILS deserves second chance -
How can people find federal government information?
This has always been a challenge for agencies and for users. In the 1940s, the Federal
Register was one answer. In the 1960s, the Freedom of Information Act offered another
approach. The Paperwork Reduction Act of 1980 proposed a Federal Information Locator
This decade's entry is the Government Information Locator Service. The goal of GILS is
to have each federal agency provide pointers to indexes and navigation aids for its
information resources. Formal and uniform data descriptions--so-called metadata--were
supposed to support both public access and records management functions.
GILS was not designed to provide the information itself, however. That is one of its
problems. The first independent evaluation of GILS came out at the end of June. The
report, by academics William Moen and Charles McClure, was sponsored by several agencies.
You can find it for yourself on the Web at http://www.unt.edu/slis/research/gilseval/gilsdocs.htm.
The report finds that GILS has not worked as planned. Coordination and direction have
been inadequate, and agencies have been confused about its mission and purpose.
The result? Indifferent and inconsistent implementation. Not surprisingly, GILS has
attracted few public users, and those users don't much like it.
The report writers are too polite to say directly that GILS has been a failure, but I
am not so reticent.
GILS' data descriptions are incomplete, the databases are insufficiently coordinated
and hard to search, and connections to the underlying resources are mostly absent. Though
lots of people turn to the Internet for government information, GILS is not a major factor
on the Net.
You can find GILS if you look. But if you are searching for government information,
someone is more likely to refer you to a federal Web page or even to the search engine at
the Yahoo Web site than to GILS.
This all spells failure, in my book.
The real fly in the GILS ointment was the unforeseen growth of the World Wide Web. The
study acknowledges this directly. The designers of GILS did not anticipate how Web
services would evolve and explode. No shame in that. No one else did either. Of course,
the Web itself has its own search-and-retrieval shortcomings, such as search engines that
give you back 300,000 hits. GILS offers better data descriptions. But better descriptions
are not enough to provide what users want today.
The Web has made people expect to find data, not just descriptions of data.
Can GILS be saved? The study says that it can. It proposes to drop the records
management functions so that GILS can concentrate on information discovery and public
access. This makes sense. Mixing too many functions is one reason it didn't work in the
The report also suggests focusing only on electronic resources. The new philosophy is
that if a resource isn't network accessible, then it shouldn't be in GILS. This makes
sense too. If GILS is to succeed, it will only work as an Internet locator. People want to
find descriptions of information resources and then to retrieve the resources with a click
or two. Paper resources will not be covered, but you can't do everything.
Before reading the report, I thought GILS ought to be scrapped. But the report offers a
coherent and specific vision of a revised GILS. It stresses that GILS policies must be
evaluated against emerging technologies with a concentration on the Net.
GILS can be viewed as an experiment that fell victim to changing Internet technology.
We must accept that not everything is going to succeed on the Net. Failure is a learning
experience. In that spirit, GILS deserves another chance to succeed.
But I would set a deadline. If a revised GILS is not meeting its new objectives and
attracting users by Jan. 1, 2000, then the project should be killed entirely. Son of GILS
may work. Grandson of GILS is definitely doomed.
GILS offered the worst of both worlds: centrally mandated requirements without adequate
support or resources. This served only to stifle agency creativity.
The new report acknowledges that the Web has changed the world and that we have to
offer government information in the way that people expect to find it.
That is a reasonable message. GILS is worth another try.
Robert Gellman, former chief counsel to the House Government Operations Subcommittee on
Information, Justice, Transportation and Agriculture, is a Washington privacy and
information policy consultant. His e-mail address is firstname.lastname@example.org.