掲示板 Forums - Crowdsourcing Projects megathread

Some suggestions to make participation in the crowd sourcing projects easier:

- Have something on the dashboard (possibly similar to a schedule), where you can select do X of Y each tasks each day e.g. tag 5 words each day, and it keeps track of when you have done them (otherwise, people like myself will completely forget about doing it). Something like this would also lend itself to being hooked up to Kao Points, which may encourage participation.

- For the mass tagging project, some definitions/descriptions of what each tag is intended for, plus some example words for each tag, would make it easier for the tagger, and probably yield more consistent results

- For the mass tagging project, a "None of the above" category, so it's easier to keep track of words that have been reviewed, but where no tag that fits was found

1. Both the tag lists for an individual word as well as the mass tagging page has a form for suggesting new terms.

2. I'd think that the explanations (when present) as well as the example words would be overbearing on the mass tagging page...maybe an expandable list at the top? I'd think that most people that use it more than a few times would not need the explanation/examples more than once/twice.

3. Definitely adding a widget and probably kao garden points, just need to get around to it.

I do plan on backporting many of the tags that are already present in the system (which came from jmdict), just haven't gotten to it yet. There's actually a huge number of them, but many are grammatical and not semantic tags, so I need to go and parse them out first.

Which display language are you using? English or something else?

I'm introducing a new set of widgets for the two widgets to help encourage participation, and they tie into kao garden coins for the two projects. Here's the word tag one - it's a bit of a different take from the current tagger.

I would like a "flip" version of the sentence marker as well, but I'm not sure how to best do it without data overload. Here is what I was thinking:

1. Instead of have a sentence with a way to mark up all the words, show a single word that has multiple defs that are commonly used (for example, 入る).

Then, I see two different options:

2A. Show a bunch of sentences with that word, asking for markup on just that single word. It would help keep the helper focused on the same word, but it could still be a lot of data visually (possibly 5,10,15 choices for each sentence.)

2B. Show a single meaning. For example, #2 for 入る (to join a club). Then say "check the sentences that have this meaning." There could be a quick English filter against the sentences, so you could (For example) get all the sentences with that word that also have the word "join" in the English translation.

I think 2B would be faster, and less info to deal with, but possibly a lot "no, this doesn't match", and there isn't much I can do with that info. So if the "enter" meaning of 入る takes up 90% of the sentences presented, the user is going to spend a lot of time searching sentences with less payout.

I'd love to get feedback on this before I implement something. In the meantime, the dashboard widget for the sentence marker will just grab a random, verified sentence for marking.

Hello, just started to use the mass word tagging tool today and I have a few questions.

First, after tagging a lot of word, I realized that it could be considered as spam since I added 242 tag today (I have two schedules with school disciplines and another for language/grammar words, so it was quick to add the "school" and "language" tag). I don't think those tag were incorrect, I hope at least, but if it's the case, I apologize.
I really want to help on this project and will probably end up do some mass tagging session from time to time, not regularly because I'm not really interested in the kao coin (even if it's nice), but now I'm worried that I'm going to be ban from tagging word. Would it be possible to have another system for that? Like, I will pay attention to the tag, especially if definition are going to be add later like suggested before, but I also know that with system like that, people always disagree. Would it be possible to have a threshold system instead? Like this word has tag 10 or 20 times with this tag, okay, it pass the check and can be send to the next step (like a manual verification if I understand it well?). Even if there's some spammer, the tag will not pass the check if it's too uncommon. Or maybe some more precise guidelines + definition for each tag? I don't know which one would be better but now I'm a bit worried about unintentionally over-doing it by being too zealous (which doesn't encourage participation).

And the second question is, where could I suggest/discuss subgroup of tags? I want to suggest a bunch of subgroup of tag related to advanced studies (like vocabulary for specialized sciences like algebra, molecular biology, etc) but I'm not sure where to do that. Should I suggest them one by one with the tool in mass tagging while adding it's a subgroup of the "science" tag or "mathematics" tag? Or is it more for later since some of them already exists with the JMdict and the backporting is a work in progress?
I think it would be useful for student who want to continue their studies in japan or just people interested in those subjects like me.

So there's a tag for greetings, but none for general communications?

We talk more about the point of the tags in that thread I linked, but the goal of the tags is two fold:

1. Make the tags a way to search for words

2. Use the tags as a way to select better (incorrect) answers in multiple choice questions.

Both of these require that the tags not be overly broad, as that would go against those 2 things. For example, someone recently suggested a tag for "abstract", since there are so many tags that categorize physical things, but less than do abstract ones. However, after some discussion, I decided it was not a good fit, because abstract is so broad that you could grab 3-4 terms out of it, show them to someone, and they'd have a pretty hard time telling me what category they came from.

It's an imperfect science, of course, and there will never be 100% consensus on tags, but that's ok.

Sentence market widget on the dashboard had been completely rewritten, will be uploaded tomorrow :)

How about now?