This is why IDOL Server also has the ability to remove duplicates based on the conceptual relevancy of the content to the other content in the index. The...
Mark, We are still debating with users what marks a duplicate content. whether they consider it duplicate if title and summary shown in results is the same...
Rup, I totally agree, let the users decide how they like it, what the requirements are. Others, There is a fancier way to try and determine duplicates, using...
Another nuance to this from the technical side is whether you're deduping the index, or deduping the results. The technique mentioned by Emory appears to...
Good point Garth, about the engine having only one page in the index, but recording multiple URLs for it. At least in simple cases, FAST ESP has both "url" and...
Does ESP then choose which of the "urls" is most appropriate, or does it return them all with the result and then custom app code has the task of choosing...
It picks one of them at index time. So "url" is always singular. Then "urls" will have at least one URL, possibly more. It uses spider rules at index time,...
I would say we are deduping the index as well as deduping the results. Index time deduping is setup using the stardard set of fields which when same a content...
Hi, when a user selects a result, and the document open's in its native application (pdf, doc), is it possible to take the user to the first instance of the...
We have a dilemma whether or not its best practice to have business logic built into search queries or search engine configuration itself. Here are some...
Doing it in the queries is much more efficient. Think about a filter that only matches 10% of the content in the index. If you filter the result list, you will...
I agree with Walter - put the limiting filter in at search tie and let the engine do the work. It is called 'SourceQueryText' in K2, which is what you are...
Hi Miles, it was VSearch.setSourceQuery() in K2. And similarly here now in IDOL we are using "fieldText" param with the action query. Result count issue is...
I think the general rule of thumb is when possible, push it to the search engine software for the efficiency reasons stated earlier. And for filtering, that...
IDOL provides a way to push some results to top based on parameters only known at query time. BIASVAL in IDOL can be used to boost the relevancy of a document ...
Also, since you're asking for a solution that works on IDOL, ESP, and MOSS, you might just look into the solution that's built into acrobat. Try out this...
The new ESP guides have a section on how to highlight the original document and the snippet then has a link back to this html rendered version with markup. ...
Please take a look at our new faceted search interface at http://www.statcan.ca/olcbrowse/english/clf/query.html? style=englisholc&la=en. We are now working...
Thanks for posting! I love hearing about the results of user tests, particularly related to guided navigation, as we frequently recommend that as the most...
The Canadian statistics site is OK http://www.statcan.ca/menu-en.htm it is a matter of finding which of the many trails to follow. Is it perhaps a private...
Super quick feedback. I don't find the UI very intuitive - it's not clear that those things on the left are categories (clickable), that they have sub...
I am impressed with the functionality you have created here. Doing multiple iterations of usability testing likely helped with the design as well. I am also...
Hi Sam, Thanks for your interest. I hope the link is working for you. We have addressed a few technical glitches that should be fine now. The main thing we...
Thanks, Otis, Do you think putting a plus sign would help? Is it the formatting of the side bar (within menu boxes rather than just on a plain background) that...
Kathy: One thing we've seen used to guide users to 'drill down' - essentially to get them in the mind-set of conversing with the search engine - is to give the...
Hi Kathy, I'm not sure I can give exact UI advice for the nav bar. I did forget to mention left vs. right part. I am *amazed* that people found it easier to...
Hi All, Does one of you know what the quality of IBM Omnifind is at this moment ? Can the functionality already be compared with products like Verity K2 v6 and...
Hi Mark, Would be either Enterprise or Discovery. My concern is about (in Verity language) gateways to Notes, scalability, language support, taxonomies. Isn't...