COMMENTS PLEASE: New JianTiZi Dictionary Basic Search...

[geek_frappa]

preview for basic search...

> http://chinese.primezero.com/strawberry/q.php?q=hsk

[sunyata]

nice work!

there is a bug I noticed in your search algorithm - it searches for any match within an english translation, which sometimes results in irrelevant results. For example, search for "read" and you will see what I mean...

[geek_frappa]

ya... i'm working to fix that right now... thanks, sunyata! if there is anything else, just let me know.

[geek_frappa]

ok try to "read" ... let's see what happens... better, but not very improved....

http://chinese.primezero.com/strawberry/q.php?q=read

[geek_frappa]

may i should allow them to semantically refine their search?

instead of "READ" ... ask user to "READ what?" ... read books? read newspapers? read palms?

[xiaozhu]

You should be careful with substrings, I input 'be' and the first 18 results were mostly types of beef dishes. I know for some stuff substrings would return useful results, but they should be the exception not the rule.

Also the the results need to be ordered better, at the moment I think it orders the results alphabeticaly (for the pinyin) for results where the substring is the first part of the English result, then another list ordered alphabeticaly for where the substring occurs anywhere in the English result. This is ok during development but don't keep it in the final product.

The user shouldn't have to provide semantics if they write 'read' the top answers should be to read books, letters, newspapers etc as these are more common uses of the word.

It's looks great so far but a user may get annoyed at having to sift through 100 or so results to find the sense they mean.

[geek_frappa]

thanks. sounds good. how are the searches for words with more than one or two letters? do you see improvement? or is it worse now?

i'm about to allow advanced options for the query (exact, all, any ,etc) ... do you care about the command-line format? or would you rather have a form?

[xiaozhu]

There seems to be no change in the results it returns.

The form is ok as long as it only optional to use it.

[geek_frappa]

no change?

hmmm...please try again...

i'm trying to knock out the bugs...

http://chinese.primezero.com/strawberry/

[chengdude]

My comment: Too bad your site is blocked, at least within my university on the mainland.

[geek_frappa]

where are you?

[xiaozhu]

Still no change! Ok before you do anymore work choose a few test examples say 10 or so. Save the results from those 10 then make your changes and see if their is an improvement. Otherwise it could take a long time to fix it!

[geek_frappa]

which 10 words or phrases do you recommend?

[xiaozhu]

It really doesn't matter. But anyway, for example:

I,

Think,

Online,

Chinese,

Dictionaries,

Are,

Really,

Complicated,

To,

Make.

Hao ba

[geek_frappa]

不行 ::I,

Think,

Online,

Chinese,

Dictionaries,

不行 ::Are,

不行 ::Really,

Complicated,

不行 ::To,

不行 :: Make.

i asked "what test words do you recommend?" not, "what are the first 10 words on your mind to make a sentence?"

This is not a real dictionary. It is a search conduit, that allows you to jump to better dictionaries or search engines. It is probably one of the worst dictionaries online. ^_^ hehe...

[xiaozhu]

Haha, my point is it doesn't matter what words you choose mate. Experiment with different words see which ones provide totally unacceptable results, like 'be'. Then modify the program and re-test to see if the results are better. I can give you more lists of words if you want

But,

I,

Won't,

Be,

Serious,

Do,

The,

Work,

Yourself,

geek_frappa!

[sunyata]

yup, there is still the same problem:

http://chinese.primezero.com/strawberry/q.php?q=go

[geek_frappa]

i will not allow two-letter searches.

at

be

go

... deng deng ...

so, problems with two-letter and one-letter searches are ignored for now because they will be eliminated... (i'm so lazy, huh)

any other problems?

But,

I,

Won't,

Be,

Serious,

Do,

The,

Work,

Yourself,

geek_frappa!

don't waste bandwidth. thank you.

[geek_frappa]

there is a bug I noticed in your search algorithm - it searches for any match within an english translation, which sometimes results in irrelevant results. For example, search for "read" and you will see what I mean...

ok, so... maybe i should make a search sorter ... than FIRST finds the EXACT match, then finds the combinations?

then create an advanced search that allows people to search EXACTLY what they want??