Practical Internet Groupware

Generating the Results Display

Before we build a test data set, let’s make sure we understand where all the data will come from. Table 8.5 shows an inventory of the required pieces.

Table 8-5. Extracting Docbase Elements from Search Results

	ProductAnalysis Docbase	Conference Docbase
DATE	element of URL	NNTP Date: header
TYPE	element of URL	element of URL
SUBTYPE	element of URL	element of URL
TITLE	HTML document title	NNTP Subject: header
AUTHOR	HTML document title	NNTP From: header
SUMMARY	CSS-tagged region of source document	region of source document

Figure 8-3. Finished design for search-results display

An element shown in boldface will be part of the results returned by any search engine, either as a piece of the URL or the document title. Note that for the ProductAnalysis docbase, whose URLs we are automatically generating, we have reengineered the namespace to include an extra element—the date. That gave us the latitude to reformat the doctitle as COMPANY/PRODUCT/TITLE so that we can include the specific TITLE in the abstract TITLE slot in accord with the final design.

An italicized item may or may not be directly available, depending on the search engine. Excite and SWISH return none of these items. The Microsoft Index Server, however, can perform queries based on custom properties including HTML <meta> tags and NNTP headers. As we’ll ...