The first step is to choose the size of index you wish to build. The larger the index, the faster the search. However, for most applications the tiny index is more than sufficiently fast. If you're indexing an extraordinarily large amount of text and searches are not as fast as you'd like, you may want to use a small or medium index to speed up the search. The only drawback is that the larger indexes, of course, occupy more disk space.

One feature of the "medium" index is that it stores for each word not only which files contain that word, but also the location in the file of that word. Therefore, if you configure your search form to display matching lines in each file, and you have some very large files, your search may be considerably faster with a medium sized index (since glimpse can go directly to the matched word to output the matching line, rather than scan the whole file). This is the situation if you create a keyword index or other specialized index. Since a keyword index usually only includes one text file, the difference in size between a tiny and a medium index should be relatively minimal.

The following form section is displayed only if the system detects that a glimpse index already exists in the directory which you've chosen to index:

If you select "Rebuild Index," the index and all related files (except .glimpse_include and .glimpse_exclude, which you maintain (see below)), are deleted, and a new index is built from scratch.

If you select "Refresh Index," Glimpse will refresh the index to include the text of any files created or modified since the index was built. This can save a lot of time if your index is very large. However, this option does not remove from the index entries for files which you've deleted, so Glimpse will sometimes think it found a match but when it actually attempts to locate that file, it's gone. This should be transparent to you, but could slow down the search. After you've deleted a lot of files, it's a good idea to rebuild the index from scratch. Also, be aware glimpse will sometimes rebuild the index from scratch even if you request a refresh, if it determines that the amount of work required to refresh the index is greater than or equal to the amount of work required to just rebuild it from scratch.

If you select "Delete all index files," the index and all related files (except .glimpse_include and .glimpse_exclude, which you maintain (see below), are deleted.

Next on the form you'll see two TEXTAREA input boxes, one labeled "Include" and one labeled "Exclude."

Include

There are two uses for the Include box.

1. The first use is to force Glimpse to include in the index, files which it would otherwise ignore. Glimpse only indexes text files. Any files which it believes to be non-textual (including some ASCII files like uuencoded files and binhex files) are ignored.

   In most cases, glimpse determines the file type accurately, but it is possible for glimpse to make an error and exclude a mostly textual file from the index. If glimpse ever excludes a file from an index which you wanted to have included, simply list the file name relative to the directory being indexed in this box.

   For instance, if you created an index of your www tree, and if Glimpse excluded a file named "my_file" in your www/products subdirectory, you could force Glimpse to include it by placing the line "products/my_file" in the Include box (You do not have to specify www, because that is implicit, since that's the directory for which you're creating an index. If you were creating an index just of the www/products tree, you'd simply list "myfile" in the Exclude directory.)

   You can always determine exactly which files were actually included in the index by looking at the list of files in the ".glimpse_filenames" file which Glimpse creates in the indexed directory.

2. If you're creating a specialized index in which you only want to include one file or a small group of files, rather than list all the files you don't want in the Exclude box, you can put a '*' in the Exclude box telling Glimpse to exclude all files. You can then check the checkbox at the bottom of the form labeled ".glimpse_include overrides .glimpse_exclude," and explicitly list each file you want in the index in the Include box.

   Remember to specify the files relative to the directory being indexed (See above for examples.)

The contents of the Include box are stored in the file ".glimpse_include" in the directory being indexed. You can also edit this file directly, such as with the WebCom File Manager.

This file is only actually used during the index build process.

Exclude

List any files you want excluded from the index in this box. Like the Include box, the path(s) listed here should be relative to the directory being indexed, so to exclude a file named "bookmarks.html" from the index being created in your www subdirectory, you'd add this line to the exclude box:

bookmarks.shtml

You can confirm that Glimpse excluded the files you didn't want in the index by viewing the contents of the ".glimpse_filenames" file which is created when the index is created, and which contains a list of the files in the index.

The contents of this file are stored in the file ".glimpse_exclude" in the directory being indexed. You can also edit this file directly, such as with the WebCom File Manager.

This file is only actually used during the index build process.

This checkbox is used when you only want to include a small number of files in an index. To do this, put an asterisk (*) in the Exclude box, click this checkbox, and list the files you want to include in the Include box.

By default, Glimpse indexes words only and ignores numbers. Check this checkbox if you wish to have numbers included in your index (that is, if you wish to allow your site to be searched for particular numbers in your files as well as text).