Page 1 of 1

result item text

Posted: Mon Feb 25, 2002 10:59 am
by penny
When a search is done on our site, the text that shows up for each item should be the text at the top of the page, but for some results, the text displayed comes from the middle of the page. There seems to be no reason why this is happening - the search string does not appear in the text, and the location on the page is different for different items.

I did not change the settings in the search script.

For two examples of this, please go to www.temple.edu and search for 'chancellor'. The text that appears for the first result is the 6th bulleted item. If you search for 'Temple' and look at the first result, the text that appears comes from about 2/3 of the way down on the page.

Thanks in advance for any light you can shed on this.

result item text

Posted: Mon Feb 25, 2002 12:55 pm
by mark
The abstract function controls what is displayed. By default it will attempt to display the first significant chunk of text (skipping nav etc). You can make it always display the beginning of the text or the text around the first hit. See the manual for <abstract> http://www.thunderstone.com/site/vortexman/node153.html and "search with hit context" in the webinator 2.5 examples http://www.thunderstone.com/texis/site/ ... ample.html

result item text

Posted: Mon Feb 25, 2002 3:50 pm
by penny
Thank you. The Chancellor is much happier now that the style is set to 'dumb'.

What I don't understand is what is considered 'significant' or 'meaningful.' Are there specific html tags that are keyed on? When I look at the source of these web pages, for example www.temple.edu/chancellor/, I do not see the difference between the 6th bulleted item, which was displayed in the default 'smart' mode, and the first bulleted item. Can you provide more insight on what makes the text meaningful or significant?

Thanks

result item text

Posted: Mon Feb 25, 2002 4:23 pm
by mark
Roughly speaking, it looks for the first complete sentence of non-trivial length.