Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 158 Stack Exchange communities.

What is meta?
Here's how it works:
  1. Any Stack Exchange user can ask a question
  2. The community provides support, votes on ideas, and reports bugs
  3. Your voice helps shape the way Stack Exchange operates

On the Chinese SE and Japanese SE sites the Related section in the sidebar shows results which are not related to the topic of the question being asked.

For example a question like "What's the difference between X and Y" would be much better linking to other questions which contain X and Y. However they currently link to questions like "Difference between A and B" or "What's the difference between C and D".

Is it possible for languages using a non-Latin writing system to focus the search on the non-Latin characters?

See the Related sections in the examples below:


What is the difference in use between 以及, 和 and 与?


Based on Uphill's comments below, I am going to try and make this clearer. Hopefully this is more helpful. The SE site is not open source, so I don't know how the DB is indexed etc. I am just a language site mod, so forgive me if I don't know my stuff.

  • Ignore the English completely "latin characters". So you are just left with the Chinese characters. e.g. "以及", "和" and "与"
  • Context can be ignored.
  • Search should search for something like (not real code) WHERE title LIKE %以及% OR %和% OR %与%
  • Where you have a long string like 这个名字有点土 I am not sure how the topic is pulled in English? As a fallback I am not sure if it is possible or feasible to find best matching string based on individual characters, maybe something like this.
share|improve this question
It is rather impressive that the SE engine comes up with any similarities, finding a similar word when you only have a few East Asian glyphs that mean different things in different contexts, mixed with Latin glyphs no less, is quite difficult. Maybe you can contribute positively, instead of just filing a complaint about a hard problem and recommend awesome libraries that know how to do this? – Uphill Luge Jul 25 '13 at 0:47
If that makes you feel better and helps you to ignore that comment: yes. – Uphill Luge Jul 25 '13 at 0:59
I don't know, other "what is the difference" questions are quite similar to a "what is the difference" question. The system's not exactly broken like you make it sound. And "context can be ignored" scares me... – Cody Gray Jul 25 '13 at 8:04
@CodyGray - For your first point, "what's the difference between jogging and sprinting" and "what's the difference between milk and cheese". Why would someone asking about jogging need to see questions about cheese or anything else that is not related to jogging or sprinting? The only similarity is the grammar. – xiaohouzi79 Jul 25 '13 at 8:22
So the problem is not limited to non-Latin (or even non-English) language sites. It seems to be more of a question category problem. – Cody Gray Jul 25 '13 at 8:24
@CodyGray - It is limited to non-English because English sites that say "milk and cheese" show questions related to "milk and cheese" Chinese site shows questions related to "What is the difference between" – xiaohouzi79 Jul 25 '13 at 8:33
@CodyGray - For the context part, I don't understand why this needs to be something scary. Because it isn't used on any other site on the SE network, why bring it in for the CN or JP sites? It isn't there for the English site. Think about it, when a question contains "fishing" as the topic on the English site, the Related section doesn't check in the algorithm which context fishing is being used in, so why the big deal when I say it can be ignored in the Chinese language? Context is universally ignored across the SE network, let's make it easy and just ignore it here. – xiaohouzi79 Jul 25 '13 at 8:44
@xiaohouzi79, thanks for asking. This is really a good suggestion, don't understand why some people can feel so negative about it... – Question Overflow Jul 25 '13 at 10:28
To be fair, the problem also exists on English sites (e.g. What's the difference between the words of “Journey”, “Travel” and “Trip”?). Implementing algorithm for grouping problem is not as easy as how we human think. It also touches NLP (Natural Language Processing) which is quite an advance technology. Not to mention how to understand what's more important in deciding how related they are (e.g. the word "difference" vs. the compared words). Unless, this discussion is only for non-English sites, then you can ignore this comment. – Portal May 5 '14 at 6:38

You must log in to answer this question.

Browse other questions tagged .