Chinese zoom search indexes entire sentences - performance?

[Chris Middleton]

Zoom Search for a Chinese language WebHelp creates an index (zoom_index.js) of entire sentences. Chinese does not use spaces between words. I do not know Chinese, but I can imagine that splitting the words might be an unsolvable computer problem. Yes?

So my question is:
Once I get my 25,000 word help translated to Chinese, will the index file be huge, and will it have an awful search response time?

[Tim Green]

Hi Chris,

Unfortunately, Chinese is not search index friendly. This is from the documentation of Wrensoft Zoom, which is the indexing engine bundled in Help+Manual:

• There are several things you can do to make Zoom produce faster search results on your website. First of all, the most important factor is the size of your index file. ... This is especially true if your website is translated to Unicode languages such as Japanese, Chinese, Korean, and others. These webpages can not be successfully indexed and will only result in a larger, slower, and less effective index file.

The point there being that because of the factors you mention yourself, a Chinese index will by definition be large and cumbersome. That comes with the territory. If you buy a Great Dane puppy you're going to have a big dog...

The only thing you can do is to maximize efficiency on those factors where it is possible:

• Define the PHP or ASP platform in your skin (NOT in the project, those settings are ignored if you're using a skin!). They are much more efficient and performant than the JavaScript option.
• If possible, increase the size of your minimum skip words. The shorter they are, the larger your index will get.

[Chris Middleton]

Thanks for the info. This subject closed.

And thanks, Tim, for your prompt, clear, and complete replies. You're some of the best support I experience.

Chris