PDA

View Full Version : Cannot search Chinese Words


jlee
02-08-2005, 08:42 AM
The search function only searches for English words.

I used to enter both English and Chinese and I cannot do search in Chinese. The search always return no match results even there are one or many identical Chinese words.

Any solutions here?

kevina
02-08-2005, 03:19 PM
Matches wildcard searches will work for Chinese characters in attribute values, but Ultra Recall currently doesn't keyword set-based languages (like Chinese) adequately to comprehensively keyword search Chinese or similar.

Note: You can do a wildcard search in with a Quick Search by placing an asterisk (*) at the beginning and end of your search criteria.

The Chinese text keywording issue is that Ultra Recall is not properly identifying individual words within Chinese text. This is due to the current keywording capability relying word boundaries (which typically don't exist in Chinese text).

We are very interested in fully supporting keywording of set based languages (such as Chinese), but achieving this capability will require significant changes to Ultra Recall (which won't be completed for some time). I apologize for the inconvenience...

jlee
02-09-2005, 10:41 AM
Thanks.

It works for Chinese words in Item title and pure text in Item Detail area. However, if the Chinese word(s) is in html format in Item Detail (directly clipped from web pages), it doesn't work some times. Don't know why?

Question: Differences between keyword and wildcard search?

Another question: How long the search function mentioned in your reply will be completed?

Thanks,

J Lee

kevina
02-10-2005, 09:31 AM
If the word you are searching for is actually a single character in the html source (with a word boundary on each side such as a space or punctuation), then it will not be added as a keyword (currently Ultra Recall only recognizes two or more contiguous characters as a keyword).

Our research into the techology mentioned ealier in this thread is quite preliminary. It would require a fundamental change to Ultra Recall which will require much effort to implement (and a major update to any existing data). We have yet to prove the technology will be efficient enough (both in time to compute and space to store) for use in Ultra Recall. Therefore I can't give a prediction of if/when it will be available in Ultra Recall.