search.mit.edu is an I/S-supported search engine which indexes world-readable web pages served through webservers at MIT. It finds pages to index by "spidering", or following links, from pages it has already indexed, and it sets its own schedule to revisit pages based on how frequently they change.

For additional information about how the search engine finds, indexes and maintains information about world-readable web pages at MIT, please see http://web.mit.edu/search/help/inktomi_at_mit.html.

Quick Examples
Here are some examples of popular query forms. If you see something here that suits your needs, please use "view source" to see how it's done in the HTML. These are intended to demonstrate functionality, not beauty or design.

• This is a search against everything that is in the search engine's indexes. It generally returns many results.
  
  Search everything known to 'search.mit.edu'
  
  
• All world-readable pages served through 'web.mit.edu' that 'search.mit.edu' knows about
  This restricts the search to only those world-readable pages that are served from 'web.mit.edu'.
  Search world-readable pages served through 'web.mit.edu'
  
• All world-readable pages served through a particular webserver indexed by 'search.mit.edu'.
  This restricts the search to pages that are served through a webserver which has been indexed by 'search.mit.edu'. An example is 'libraries.mit.edu'.
  Search world-readable pages served through 'libraries.mit.edu'
  
  
• A single "locker" (pages served from web.mit.edu)
  This restricts the search to a single "locker", which is a set of files served though 'web.mit.edu' and which are distinguished by the locker name at the start of the URL. In general, a locker and all files and directories contained in it reside in the AFS filesystem. The indexer only indexes those pages it finds by spidering AND which allow

  system:anyuser

  access to read the pages. As an example, search the Information Systems 'is' locker.
  Search the 'is' locker
  

  Other examples of forms include:

• The default "simple" query form on search.mit.edu
• The default "advanced" query form on search.mit.edu

FAQ: Frequently Asked Questions

• Q. How do I write a form to find anything at MIT?
  A. You really can't. 'search.mit.edu' is a single search engine which indexes world-readable web pages served via webservers at MIT. 'search.mit.edu' does not index:
  • world-readable web pages at MIT which it doesn't know about
  • web pages at MIT which are not world-readable
  • web pages not at MIT
  • documents that are not accessible via the web
  There is a lot of stuff at MIT. For general searching, a few places to look are the MIT Homepage Search, the Libraries Search Page, and a search page in the search locker.

• Q. How do I write a form to search a specific "locker" at MIT?
  A. See the example above.

• Q. How do I make a form search only a specific webserver at MIT?
  A. See the example above.

• Q. The indexer found my site, and I don't want it in the index. How can I get these out of the index?
  A. TBD

• Q. The indexer hasn't found my new site. How can I tell the indexer about it?
  A. TBD

• Q. The indexer hasn't found my new pages, and it still has the old ones. How do I get the old pages out of the index and put the new ones in?
  A. TBD

• Q. How do I make a search form return my site as the first item on the results?
  A. The query form doesn't determine which result is displayed first. The query form merely restricts the search to a subset of all pages indexed by the search engine, and lets you modify the format of the results. Other than choosing to display results sorted by relevance or by date, you cannot use the query form to control the order in which results are displayed. Thus, the only way to guarantee that your page will come up first is to create a query form to search only your pages.

• Q. What does 'search.mit.edu' index?
  A. It indexes world-readable web pages that are served from webservers at MIT. It finds new websites and web pages by "spidering", or following links from web pages it already knows about. See http://web.mit.edu/search/help/inktomi_at_mit.html for additional information.

• Q. How do I make a form search a webserver not at MIT?
  A. You should contact the administrator of the webserver to find out. The MIT search engine does not index webservers outside of the MIT network.

• Q. What are the "internal names" of the collections of 'search.mit.edu'?
  A. 'search.mit.edu' indexes two main collections of documents. Each collection has an internal name which you must use if you want to restrict your query to a particular collection. The collections and their internal names are:
  • webmit: the collection of world-readable documents served via 'web.mit.edu'
  • websrvrs: the collection of world-readable documents served via other webservers at MIT (that is, not including 'web.mit.edu')

• Q. What are all the form variables that can be used in queries directed to 'search.mit.edu'?
  A. They are too numerous to list here, but they are described in the "Customization Guide" in the Verity Ultraseek documentation.

• Q. Where can I find more examples of MIT-specific query forms?
  A. One way is to use the default query form to find all web pages known to the search engine which contain a link to 'search.mit.edu'. Try it out. For each result, you can click on the page. If you like the query form, you can use 'view source' to see how it was done.

• Q. How do I make a form that searches for "MIT-only" stuff?
  A. Currently, there is no way to do this. However, IS is working on implementing a search engine which will index "MIT-only" websites, and will allow queries only from web browsers presenting MIT personal certificates. Watch this space for details.

• Q. How do I write a form to use a search engine other than search.mit.edu?
  A. If you want to use google or any other search engine, you'll find information (if any) at those sites.

Defining What Your Custom Query Form Should Do
The first step in writing any custom query form is defining what you want to search for. Most web publishers will want to search their specific web site, such as the Libraries web site, or the Biology web site. You may also want to consider offering options such as also searching the MIT web site. Another option to consider is how you want the results displayed, e.g. with summaries or without, 10 or 500 results at a time.

How a Custom Query Form Works

• Queries directed against 'search.mit.edu'
• Queries directed against other sources at MIT

Developing and Testing Your Custom Query Form
Once you have an understanding of what is possible and have defined how you want your custom query form to behave, and know what to expect in the results display, you are ready to write and test out your custom query form.

• Start with one that works.
  Start with one of the examples on this page, or look around at other web pages at MIT. When you see one you like, 'view page source' to look at the HTML. If it sends a query to search.mit.edu, it's something you can probably copy and modify to suit your needs.
• Modify it to suit your needs.
  
• Verify that it works as you intend.
  

Why Doesn't My Query Form Work?
Here are questions to ask yourself when your form "doesn't work" as you intend.

• Does the "FORM ACTION" line use the host search.mit.edu?
  
• Have you verified that the search engine has indexed the site your form is intended to search?
  
• Have you used appropriate form variables correctly?
  
• Can you describe exactly what is wrong with the way it is working?