The proposal is in two parts: (a) that the search engines currently offering the ability to restrict a search by domain agree to support a common query syntax for this command and (b) that search engines currently lacking such a domain restriction command add this feature in the near future.
Some search engines currently offer the ability to restrict a search by domain, which is an extremely useful feature for search engine users. For instance, this search on AltaVista:
mars landings -host:nasa.gov
would allow one to find pages that deal with the mars landings but which do not come from NASA web sites. That could be helpful if pages from NASA web sites dominate the top search results. Without this domain restriction, the searcher would have no way to eliminate these pages.
Likewise, this search on Infoseek:
tony blair +site:gov.uk
would find web pages about UK Prime Minister Tony Blair that only come from official UK government web sites. Without such a domain restriction, there would be no way to narrow in on the government sites in this way.
A domain restriction command also benefits webmasters. They can immediately discover exactly which of their pages have been indexed by search engines, which could eliminate needless resubmission attempts and feedback messages to search engines about the status of their sites.
Search engine reviewers can also use a domain restriction command to compare the coverage of the web that crawler-based services provide, assuming that the command is also coupled with accurate hit count reporting.
AltaVista, Infoseek and the Inktomi-powered services offer a domain restriction command, but each uses different syntax, as follows:
Lycos also offers such an ability, but only using a form and only when accessing its Lycos Pro advanced service. Northern Light offers a URL restriction command that can be manipulated in most cases to be a domain restriction command. AltaVista also supports the domain: command, which restricts searches to a top level domain (ie .com, .org, .edu, etc).
A) That services such as those named above unify on a common query syntax that could be entered into a search box to restrict a search by domain. This syntax would also then serve as a voluntary standard for other services to use, should they add a domain restriction command.
B) That services lacking a domain restriction command add one in the near future.
This proposal doesn't address consistency of results. For instance, assume that domain: was agreed to be the common syntax for domain restriction. Then assume you ran this identical search at AltaVista and Infoseek:
"mars landings" -domain:nasa.gov
At AltaVista, pages from any nasa.gov site would be excluded (ie, *.nasa.gov). But at Infoseek, only pages from the main nasa.gov site would be excluded. Pages from www.ksc.nasa.gov and www.jpl.nasa.gov would still appear.
Ideally, this proposal would culminate in the expected behavior of a domain restriction command, in addition to unifying the query syntax. In other words, should all subdomains automatically be restricted when a higher-level domain is specified? Or should there be two different commands, one that only restricts on the basis of the exact site specified versus one that restricts both the upper level domain and its subdomains inclusively.
For instance, site:nasa.gov could indicate restriction only for the nasa.gov site while domain:nasa.gov could indicate restriction for any site containing nasa.gov in its domain.
Behavior should be raised when this proposal is discussed, but the priority is to unify the syntax first. Behavior unification can be the focus of a follow-up proposal, if necessary.
This proposal is also specific to the domain restriction capability and its query syntax, not to user interface. Search engines might decide to have their own unique menu options to support domain restriction in addition to supporting this via a query command. But by supporting a common query syntax, users could learn at least one standard way of doing what they want when using different services.
This proposal is now put forward for discussion among the major search engines. By early June 1999, it is hoped that each search engine will have said if they will support the proposed standard or a modified version, or if not, that they will explain why.
The Original Search Marketing Event is Back!
SES Denver (Oct 16) offers an intense day of learning all the critical aspects of search engine optimization (SEO) and paid search advertising (PPC). The mission of SES remains the same as it did from the start - to help you master being found on search engines. Early Bird rates extended through Sept 19. Register today!