The Government Domain: Why Google Uncle Sam?

Peggy Garvin of Garvin Information Consulting is author of The United States Government Internet Manual (Bernan Press) and contributing author for The Congressional Deskbook (TheCapitol.Net).


Not so long ago, a variety of free niche search engines for U.S. government information marketed themselves to web researchers. Remember GovBot, SearchGov.com, and SearchMil.com? GovBot is gone. SearchGov and SearchMil live on, but only as facades: type a search in the box and you’ll be ushered into Google. Google’s Uncle Sam - released in 1999, the same year as Google itself - is one niche search engine that is still around today. Despite little promotion from Google, its Uncle Sam search has become popular with researchers, librarians, and others who train novices, create search guides, and recommend resources. Given some of its limitations, which I describe in this column, I have always wondered why.

Google Uncle Sam is not listed with any of the Google features, services, and tools in Google Help Central, and it is not lurking in the Google Labs. If you are using it, chances are you were referred to it by someone else and now have it bookmarked in your browser or in your brain.


[click to enlarge]

Other searchers find Google Uncle Sam by going to Google Advanced Search, scrolling to the bottom of the page to the topic-specific search options, and clicking on the U.S. Government link. The annotation next to that link says, “search all .gov and .mil sites.” And that is what Google Uncle Sam does.

The .gov and .mil domains covered by Google Uncle Sam certainly carry the bulk of the information that the federal government puts online, but they do not carry it all. USDA’s Forest Service uses the .fed.us domain for its home at http://www.fs.fed.us. The Postal Service is a dotcom at http://www.usps.com. The National Defense University is at http://www.ndu.edu. Those are just a few examples of what you are missing with Google Uncle Sam. Much more of a nuisance, depending on your research goals, is the fact that the .gov domain is not limited to the federal government.


[click to enlarge]

With Google Uncle Sam, your search results will include hits from the many state and local governments that use “.gov.”

Uncle Sammers should also be aware that there is no advanced search template. Click on the Advanced Search link, and you will be sent to Google’s generic Advanced Search screen, without specific government domains selected.

Fortunately, we do not have a search engine monoculture for federal government information. The other major search engine for the U.S. government webspace is the FirstGov search engine at the government-operated FirstGov site, http://www.firstgov.gov. The search feature maintains a low-key, small-font presence in the upper-right corner of the FirstGov home page. Although FirstGov has become more of a directory-based site, enabling discovery of information with much more guidance than a blank search box provides, it owes its start to the free loan of search tools from Inktomi (a search engine now owned by Yahoo!). When that gift loan expired, the FirstGov contracting team chose Fast Search and Transfer, implementing it quickly in 2002.


[click to enlarge]

FirstGov Search does not share Google Uncle Sam’s drawbacks. It allows users to search only federal web sites, eliminating results from state and local governments if desired.
 


As an option, it can search both state and federal, all states, or selected states. The federal sites include those from entities such as the Postal Service and National Defense University that are not in the .gov or .mil domains. FirstGov also has a powerful advanced search template. Options include limiting by format (HTML, PDF, MS PowerPoint, etc.), searching on word variations (e.g., tax, taxes, taxing), including or excluding results from a specific web site or domain, and sorting results by relevance, date, or size. It also has spell-check for those of us who type with our thumbs.


[click to enlarge]

So what’s not to like? Probably the most glaring fault of FirstGov Search is that it is not Google. It does not have Google’s super-strength relevance ranking, or its convenient page cache feature.

 

There are also some fundamental weaknesses that go beyond not-being-Google. I did some sample searches to test both Google Uncle Sam and FirstGov Search. With powerful software behind it, FirstGov has a lot of potential. To live up to this potential, it appears to need work in the areas of cleaning up its index and—if not quite getting to the Google level—improving its relevance ranking.

I searched on the phrase “cultural heritage” in both services. The first ten results from FirstGov were a little moldy and obscure. They included a 1998 media advisory from the Indian Health Service reporting on President Clinton’s proclamation of American Indian Heritage Month that year, a broken link to something from former Senator Breaux’s office in September 2000 about Louisiana’s cultural heritage, and a brief 2001 article from the National Oceanic and Atmospheric Administration about the cultural heritage of Indiana’s Lake Michigan Watershed. These results are not necessarily irrelevant to the topic; they just represent dead ends in my web research. They are not current, and they do not give me a good idea of which federal agencies are active in cultural heritage issues.

In contrast, the first ten results from Google Uncle Sam included some appealing links to the State Department’s international cultural heritage programs, the main page of the Bureau of Land Management’s program for cultural resources on federal lands, and a 2004 report on digital collections for cultural heritage that was sponsored by the Institute of Museum and Library Services. Admittedly, it is a little shallow to make this “first ten results” comparison based on an unsophisticated search, but that is how citizen-searchers evaluate the performance of a search engine. Before we stick our heads back in the Google sand, however, I must report that Google also was not perfect. Four of the first ten results were at the state level, and we do not have a way to filter these out. Also, one of the top ten was a broken link to a 1996 document from the International Trade Administration. Not helpful. I called the FirstGov results “moldy and obscure,” but sometimes what you are seeking could be just that. I encourage those who want to do thorough searching to give FirstGov and its advanced search capabilities a try.

Wanting to test FirstGov’s inclusion of federal sites outside the .gov and .mil domains, I did a search on “timber products” in both services. Indeed, Google did not return any results from the relevant web resources of the Forest Service site, with its .us extension (except for hits from a regional Forest Service site with a .gov extension, which both services found).

Two points for Team FirstGov? Not so fast. Other test searches showed that while FirstGov covers the federal domain better than Google in some ways, it does not cover the federal domain neatly. A search on “Director of Central Intelligence” turned up results from the decidedly non-governmental Federation of American Scientists (FAS). OK, they were links to government documents that FAS had obtained and posted. Maybe that’s why FirstGov indexed them. But the number one ranked result from the search I did on “cultural heritage” was from the private organization OCLC, and other searches verified that this was not an isolated example. Arguments can be made that these non-government resources were useful and relevant to the search request. But if I want non-government results, I’ll go to the big generic engines or to the many other web discovery resources that exist.

To wrap things up, I tried a search on a unique term, CONSIAT. It is the acronym for the National Institute of Standards and Technology (NIST) program called Construction Integration and Automation Technology. On this search, Google Uncle Sam and FirstGov Search proved to be equals. Google got 55 results, mostly relevant; FirstGov got 52, mostly relevant.

Lessons learned? When searching the federal government niche, follow the same
recommended practice as in general searching: use more than one search engine.
Don't neglect FirstGov Search. FirstGov has some improvements to make, but I
suspect they will be making them since this is their core business. Another
recommended practice for general research is to not rely completely on keyword
searching. Try directories and other tools. One of my test searches in
FirstGov was "expedited passport." (I needed one.) FirstGov threw back both
gold and dross in over a thousand results, but it would have been smarter to
have just clicked on the link near the top of the FirstGov home page that
says, "Get a Passport Application."

Additional Search Tools for Federal Government Research

Department of Defense Search  - DoD’s own search page provides a variety of search options for defense-related sites in the .mil domain and other domains. It also has a special search box for finding DoD news, speeches, and other press resources.

Vivisimo FirstGov Cluster Search - Vivisimo’s contribution to a diverse search ecology is its technique for clustering results into groups of similar items. This allows searchers to get an overview of all results at a glance, rather than page-by-page, and to quickly identify clusters of relevant results. Vivisimo has applied its software to the FirstGov index. Using the main Vivisimo search page, type a search and select “FirstGov” from the drop-down list to the immediate right of the search box.