LINMO

Guidelines: Indexes

Indexes play a key role in the usability of LINMO-compatible web sites, just as they play an important role in the usability of hard copy (i.e., printed on paper) books.

Usability refers to the convenience, efficiency and satisfaction with which a product can be used. Indexes add to usability by serving as a search mechanism and facilitating finding specific information in a printed book or on a web site. A well-constructed indexing system can serve as a superior alternative to the conventional search mechanisms that are used on many web sites.

The concept of indexes on LINMO-compatible web sites differs substantially from that of indexes in hard copy books, largely because of the different technologies available. However, it also represents a difference from the philosophy of conventional web sites.

A LINMO-compatible web site can be looked at as a series of interconnected indexes, that is, with each page serving as an index for other pages on the site. This is analogous to the LINMO concept that each page serves as a glossary entry for at least one other page. Each index item is a hyperlink, and the text that forms the link is the exact, or approximate, title of the target page of that link.

There are several types of indexes on a LINMO-compatible site. The most comprehensive is the main index, which consists mostly of an alphabetic listing of every page on the site both according to its title and according to any keywords in its title. The latter can make it much easier for users to locate specific topics, as it largely eliminates the need to guess the appropriate word to use. Because of its central role, this index can be accessed immediately from any page on a LINMO-compatible site by clicking on the invisible link in the upper left hand corner of a page (just to the right of the link to the home page); this has the result of making any page on the site a maximum of only two mouse clicks away from any other page.

The main index on a LINMO-compatible web site can differ from the site maps found on many web sites in several ways. One is that it is an alphabetic listing, rather than a hierarchical listing. This is because LINMO-compatible web sites have little in the way of file hierarchies and treat every page equally (with minor exceptions for the home page and main index). It is also because it is easier for users to locate the desired content when it is listed in alphabetic order rather than in some artificial hierarchical order. Another difference is that entries are listed multiple times, according to each keyword, thereby further increasing efficiency for users.

Usability can be further enhanced in the case of large sites by supplementing the general index with multiple categorical (i.e., specialized) indexes, each of which is located on a separate page. Each of these indexes is listed in the general index, both at its start and in alphabetic order. The main advantage of these supplemental indexes is that it makes it easy for a user to see what is available on the site about specialized sets of subtopics.

A third type of index that can add to the usability of a site is a chronological listing of new additions and major updates of pages to the site. This allows users to find the most recent additions and updates at a glance. It also serves as a good chronology of the development of a site.

The fourth type of index on a LINMO-compatible site is the ordinary content page. Such pages can be considered as a type of index, in contrast to a pages in a hard copy book, because they contains references (in the form of hyperlinks) to other pages. Unlike the other three types of indexes, the index items (i.e., hyperlinks) are not provided in the form of an alphabetic or chronological list; rather, they are provided in context (i.e., of the body of text), with only the first occurrence of a specialized term or concept on a page becoming an index item.

Although creating general and specialized indexes can seem like a tremendous amount of work, especially in the case of large sites with hundreds of pages, it can actually be very easy if entries to the appropriate indexes are made at the time of addition or updating of each page.

Conventional search mechanisms can be technological marvels. Some of them work exceedingly well and can be highly efficient, particularly when used with extremely large amounts of highly heterogeneous data, such as the web as a whole. (This is certainly the case with Google, but Google might be the exception rather than the rule.) However, there can also be problems with such search mechanisms. One is that they often return large amounts of largely or totally irrelevant results. A second is that they do not necessarily provide a good idea of what information is available on a site about a particular set of subtopics. A third is that they can be slow. A fourth is that they can be difficult to implement for a person who is interested primarily in content development and who is not a programmer.