Interpreting Search Results
Click to go backClick to go home

This page tells you how to interpret the results you get when you type a query to a Panoptic search service. The example used is a search over websites operated by CSIRO Mathematical and Information Sciences.

Each section of the results page, from the header to the footer, is described in turn.

 
1. Header
 
The header can be customised by the organisation operating the search to include whatever logos or written information it wishes.
 
2. Search Identification
 
Searching Intranet [ Updated:  Aug 26 2002 ] Form: simple advanced
 
This box identifies the search service you are using and gives the time when the index for that service was last updated. The advanced search link takes you to a more customisable version of the search GUI which allows you to control various search features.
 
3. Search Box
 
The search box is where you type your search query. You can cause the query to be processed by hitting RETURN or by clicking on the SEARCH button. Depending upon which browser you are using, you should be able to use the normal text editing operations, such as left and right arrow keys, ctrl-A (to move to the front of the query), ctrl-E (to move to the end of the query), ctrl-K (to delete from the cursor to the end).
 
4. Result Summary
 

[ Query: investment -- Documents: 1346 fully matching plus 0 partially matching ]

 
Immediately below the search box, the line of text in square brackets tells you the total number of results (either full or partial match) which were found. Panoptic scans the first 200 of these results and reports how many of these were unique and how many were duplicates of others.
Panoptic also shows you the query as it was actually processed. If the query shown here is not as you originally typed it, it may be because the query has been reordered for efficiency purposes, or ineffective stopwords have been removed, or because you typed an invalid query which the system tried to fix up.
 
5. Tier Marker
 

Documents matching 1 out of 2 constraints

 
For simple queries, the presence of each query word or phrase in a document constitutes a constraint which Panoptic attempts to satisfy. In the example query, there are thus three constraints.
Panoptic results are presented in tiers, corresponding to a decreasing number of constraints satisfied. Each tier is headed by a tier marker listing the number of constraints satisfied.
 
6. Results
 
1. 100 CSIRO - HAIL seminars - Abstract
... Search Research David Hawking CSIRO/MIS, Technologies for Electronic Documents Tuesday 15 August at 11am Abstract The Canberra chapter of the CMIS TED group and the ACSys WAR proj ... Search Engines Meeting: http://www.infonortics.com/searchengines/boston2000pro.html Short resume David Hawking's PhD was entitled Text Retrieval over Distributed Collections. He joined CMIS in 1998, ...
http://www.cmis.csiro.au/conferences-seminars/hail/Abstracts/2000-past/DavidHawking.htm - 5k - Cached - 17 aug 2000
 
Each result starts with its position in the ranking, in this case 1. Then the relevance score is printed in red, expressed as a percentage of the highest score achieved by any document. Sometimes the highest ranked document does not achieve 100% because a higher relevance score was achieved by a document which did not satisfy all the constraints.
Next the title of the document is shown as a clickable link, followed by a query-biased summary in which occurrences of the query terms are highlighted. Please note that, for efficiency, the summarizer only looks in the first few kilobytes of the page and may not find the best parts of a long document.
Next, the URL (or URLs in the case of duplicates) is printed with the size of the page in kilobytes and a cached-copy link to the copy of the document as it was downloaded by Panoptic. If the original document was in a format such as PDF or Word, the cached copy will contain the unformatted text extracted by Panoptic. Finally, the last modified date of the page is printed (if it was made available by the Web server.)
NB: The URLs referenced in the example above may no longer be available.
 
7. More Results
 
Subsequent results are presented in the same format, separated where appropriate, by tier markers.
 
8. Next/Previous Page Link
 
When a query is submitted, the top 200 results are computed and saved but result pages usually contain only ten results. Links at the bottom of each result page allow you to move to next or previous result pages.
 
9. Footer
 
The footer can be customised by the organisation operating the search to include whatever logos or written information it wishes.


Panoptic Search Engine

© Copyright CSIRO Australia, 1997-2004.