Search Returns Invalid Abstract (288445)


The information in this article applies to:

  • Microsoft Site Server 3.0
This article was previously published under Q288445

SYMPTOMS

When you view search results, HTML code or comments may be in the document abstract instead of the expected summary text.

CAUSE

Site Server Search usually builds the abstract (or characterization) for an HTML page from the first 300 characters of the document body or from a meta description tag. Improper placement of comments can cause the Gatherer service to return HTML code in the abstract.

RESOLUTION

If you are not using a meta description tag, make sure that comments are correctly placed in your HTML document, especially comments in tags. For example:
<IMG SRC="<!--#include virtual="/include/include.inc" -->/vdir/image.gif" width="582" height="20" border="0">This is the first text in the document.

This code above causes the abstract to show the following:

"/vdir/image.gif" width="582" height="20" border="0">

By removing the quotes placed around the include file path, Site Server Search displays the first 300 characters of the document, and not HTML code that is being incorrectly interpreted as words. The reformatted code would look similar to the following:
<IMG SRC="<!--#include virtual=/include/include.inc-->/vdir/image.gif" width="582" height="20" border="0">

The abstract displays:

This is the first text in the document.

MORE INFORMATION

Site Server uses the Microsoft Site Server/Data/Search/Config/Gathrprm.txt file to determine how many characters are used for the abstract when indexing documents. The default setting is 300 and can be modified by changing the value of the following line in the Gatherprm.txt file:
<AutoDescriptionLength value=300>
Modification Type:MajorLast Reviewed:3/28/2001
Keywords:kbprb KB288445
©2002 Microsoft Corporation. All rights reserved.