New Feature: Tags!

[roddy]

I'm going to make a post or two over the next few days about new features we've got as part of yesterday's forums upgrade. The most useful one I think is likely to be tags, which provide users with a way to tag topics with relevant words or phrases to make it a bit easier to find them in the future.

If you look at the bottom of any thread page, you'll see a 'tags' box - click 'edit tags' and you'll be able to add words or phrases, even if someone else has already done so (although there's a maximum of 25 tags per topic.

For example, if you visit this topic, you'll see that at the bottom it's been tagged as 'blcu'. You can then click on that to see a list of all topics tagged with that word.

Alternatively visit the tags page to see a list of all current tags - the larger the word, the more tags. Currently BLCU and HSK are largest, mainly because yesterday I tagged a few hundred topics with one or the other of those keywords.

When you are making a new post you can add any tags you want (up to five I think) and when reading you can add up to two to other people's topics. For the feature to be valuable will rely on people actually adding tags, so if you've got the time please do so.

It's possible for admin staff to tag threads en masse, as long as we can search for them. For example yesterday I searched for all topics with BLCU or HSK in the title and tagged them. This is a crude but effective way of getting things started. If you can think of any other mass taggings we can do similarly, let us know and we'll take care of it.

Any questions, just ask . . .

[roddy]

Have just spent an hour or two kickstarting the tags function by adding them in bulk based on title keywords - we now have 126 of them:

偏旁 废名 文言文 电影 蔡琴 accommodation advertisement anki beer beijing bfsu blcu book review cantonese chengdu chinese name chinesepod cjkos classical computer-aided csc dalian degree diet diqiucun driviing in china fei ming first episode project fish flashcards fudan university gcse chinese graded chinese reader 1 grammar grammar books guangzhou gym hakka hangzhou harbin headache history history books hokkien homestay hong kong hsk hwr input internet slang jinan jinan university jw202 kingsoft powerword kunming learning to write lef listening skills london mac mail mandarin mandarin cartoons map mba medical mobile phones movies nanjing nokia npcr ocr olympics palm pimsleur pinyin pinyinput plecodict poetry pool post qigong qingdao qq radicals read readers reading residence permit romanization rosetta stone scottish shanghai shanghainese shenzhen singapore soas studying in taiwan subtitles summer supermemo tags taiwan taiwanese textbook textbooks texts tianjin tones trainspotting translation tsinghua university uk system cartoons upgrade vegetarian visa writing samples wubi wudaokou wuhan xinjiang xp yunnan zdt zhongshan university zhuyin

And when I have time I'll do a load more. Would really be appreciated if users can make use of this feature and add tags when posting a new topic, when reading or replying to topics or just when they think it would be useful. Perhaps you could have a look at your subscribed threads and add some tags there.

You can see the current tag cloud here. Please note that changes might take up to half an hour to appear, so you might not see changes instantly.

[TCcookie]

Question: Are there any sorts of conventions we should be following to make the tags more searchable or more neat? The first thing that comes to my mind is whether nouns should be singular or plural. Should Chinese tags, if any are assigned, be written in both simplified and traditional where appropriate?

[roddy]

To be honest at the moment I'm just hoping to get people using them. However when you start to type tags into the box, you'll notice you get autocomplete options coming up - you could use that to check if, say, 'domestic flight' or 'domestic flights' has been used previously.

On simplified / traditional issues - when we've run polls we've had a significant majority of users saying they use simplified, so that's the default preference for the forums - ie, if it doesn't make any difference to you, we prefer you use simplified. However, for those who are used to traditiona;, we'd rather have you participating than thinking 'damn, I'm not allowed to use traditional', and therefore we're not going to complain if you do so. So all are welcome.

It would be good if people can pay a little attention to what tags have been used previously, either by looking at the tag cloud (always available via 'tag search' under the search menu) or the autocomplete and choose tags already in use wherever possible. However it's not a major issue as if necessary admin staff can edit tags in bulk. Or at least I assume we can, it's all a bit new to me too

[muyongshi]

Of course the biggest thing here is that this is going to be a test for all of us in spellig (sh) (that was on purpose by the way)

Does it really distinguish between the traditional and simplified? I could see it happening on certain words but I would think that it doesn't separate something like 电影 and 電影. I'm just thinking about like google how it will search in both even though you only type in one...

[roddy]

It won't be able to match up simplified and traditional versions of the same word, so we could have duplicate tags. We don't quite have Google's resources yet.

[roddy]

Of course the biggest thing here is that this is going to be a test for all of us in spellig

Yeah, I just edited 'driiving in China' to contain a more conventional number of vowels.

Made what I think is a useful change just now - expanded the range of fonts in the tag cloud, because before you could read 'Beijing' and 'BLCU' and everything else was very small.. Should be more usable now - take a look.

I've also increased the number of tags which you can add to existing threads. Originally you could add five to your own thread, and two to others. You can now add five to any thread.

Good to see that people are starting to tag threads. You can see a tag cloud based on usage at the bottom of the advanced search page - font sizes here are based on which tags have been clicked most, while the other cloud is based on how many threads have that tag.

[roddy]

And now we've got a tag for Dashan!

[renzhe]

but I would think that it doesn't separate something like 电影 and 電影. I'm just thinking about like google how it will search in both even though you only type in one...

Google does all sorts of black magic. It will take "Btrdny serpas" and give you Britney Spears.

They have a HUGE database of what people type in and use this to guess what other things people might have wanted, they also do all sorts of lexicographic analyses etc.

[roddy]

So Renzhe, what do you on-topically think of the new tags feature.

[renzhe]

I tagged all the First Episode threads as soon as the feature came online. I love it!

[roddy]

Well I'm glad you like them - looks like people are starting to use them too. Almost up to 200 different tags.

Just made a couple more changes:

You can now have single character tags - ie 了.

The Chinese comma ， is now also recognized as a tag separator so you can type 你，你好 and it'll be recognized as two tags rather than one.

[muyongshi]

That is a big help!!!! Just hopefully we don't get a lot of accidental tags like "s"

[roddy]

Upped the maximum length of tags to fifty characters - stuff like 'taiwan language institute' and 'beijing language institute' weren't fitting into the original 25 characters.

Well over 200 tags now - any helpful folks who want to do some tagging can do so either just as they browse the forums, or by looking through their Subscribed threads, or by searching for threads on topics they are knowledgeable about or interested in. And the world will thank you.

Also, raise any tagging issues here and I'll see what I can do.

[roddy]

Now at over 400 tags - thanks to everyone who's been tagging threads.

[roddy]

We're now up to 636 different tags. I'm not sure how many actual topics are tagged, which might be a more useful measure, but never mind.

Thanks to everyone who's been taking the time to tag topics - it really does make information easier to find, and I'm noticing the tag pages ranking highly in Google in some cases.

If you're not sure what I'm taking about, see the little tag box below - members can add new tags by clicking the 'edit tags' link, and you can get an overview of what's being tagged already here - also accessible via the search dropdown.

[roddy]

I've just installed and run an automated tagging add on, which means that all topics have been automatically tagged. The caveat, however, is that:

Because it's automatic, it does some odd stuff - eg if someone wrote help(urgent) then you might end up with help(urgent) as a tag. Similarly it can't parse Chinese (although I've told it about Chinese pronunciation and am running the whole thing again so that should help - edit, no, it didn't). To limit the number of nonsense or pointless tags you see I've reduced the number of tags shown in the tag cloud to just the 500 (might change that) most commonly used.

That means tags you've added might not appear in the tag cloud. Rest assured they will still appear on the topic, and still be accessible via the tag search.

Any manually added tags will not be affected. However you might see auto-generated ones appearing alongside them. You will see odd stuff (for example a thread with a Chinese title might have the entire title as a tag) but basically it's best ignored - having the English tags in place is still pretty handy.

You'll also notice you get automatically suggested tags when you're posting a new topic (at least you should). Ideally people will edit those and improve on the suggestions, but if not it'll be better than none at all.

Thanks to everyone who has been taking the time to tag threads - I know you're out there, even if I don't know who you are.

Edit: dammit, I swear this was working right up to the point I posted about it.

Edit edit: Ok, it's working now, although I'm not sure about the value of the automatically generated tags, and they're swamping the much smaller number of manually done ones. Have a look and see what you think.

Edited April 22, 2009 at 05:05 PM by roddy