Menu
[not needed anymore- removed by CK]
[not needed anymore- removed by CK]
I don't get it...
>I added an extra column to the "native speaker" stats to include the number of OK ratings that these members have added.
I've OK-tagged hundreds or even thousands of sentences, and the corresponding number is zero.
Do you have an explanation ?
These stats are actually misleading. They are said (see above) to represent the added OK ratings given by a member, when they actually are presented as if the OK ratings had been APPLIED (by others) to the sentences of the member.
Hardly the same thing...
So, in my view, they're just a new way to self-promote oneself, which you are a champion at...
[not needed anymore- removed by CK]
Have you spotted how many actual users use this new functionality ?
It's but a handful of admins and corpus maintainers...All this effort for NOTHING !
What a shame !
You KNOW your stats are MISLEADING ! YOU DO !
I thought that the tag OK is obsolete now, since we started using the green, yellow and red buttons. I stopped using it a long time ago.
So, if it's not obsolete, what is the difference between the green mark and the OK tag?
For me, both of them have only one purpose, that is to specify that the sentence has been proofread by someone who consider it okay. However, I do not understand the green, yellow, red (GYR) system and my use of the OK tag is only as an "anti-other flag system". My point of view is for the French corpus, and I see it like this:
For a sentence to be in the corpus, it has to be "good". "Good" having no clear meaning, for a French sentence, I see it as fulfilling "grammatically correct (conjugation, choice of words and stuff) + somebody has used this sentence before and will use it again in the future". If it is not grammatically correct, then this sentence has to disappear from the corpus. If nobody would never use this sentence like this, some people call this "to sound natural at the ears of some native speakers", then it is the same as the monkey experience when you put words together, hoping for hitting a good sentence, so it has to disappear as well.
Therefore, for me, the R and the Y of the GYR system are obsolete. If it is "not OK", this sentence has nothing to do here and should be deleted (or changed), we use flags to detect that. If the sentence is "unsure"... How a sentence could be unsure, I don't know, but anyway we should discuss it and arrive at a good solution, again we use flags or directly @name comments to do that. And finally, the "OK" is, for me, the only one that could be used, but if the sentence stays in the corpus, from what I explained above, it does not need and OK mark. It is by its own existence OK.
As long as a sentence has no flag on it, I consider that there is no problem with this sentence or that it was not proofread yet. So the "OK" thing might be an indicator of who has proofread what, but everybody making mistakes, I would proofread the proofreader anyway...
Finally, the only time I use the "OK" tag is when it was tagged need native check, because it is a way to explicitly say "user not native but sentence as good as", and potentially avoid the return of the tag, later.
I use the red rating when one or more users say that a sentence is wrong, and when the author - after discussion - refuses to correct it.
In terms of functionalities, there are some differences:
- It is possible to search sentences that are tagged OK, but it is not possible to search sentences rated OK.
- It is possible to browse all the sentences tagged OK (and filter them by language), but is is not possible to browse all the sentences rated OK.
- Tags can be applied by advanced contributors only. Ratings can be applied by anyone.
(and I'm probably missing a few other points)
My personal position on this matter is that users should favor using the ratings rather than the tags, to mark a sentence as OK. However, I don't want to push people to use ratings if they feel the tags are more appropriate for their use case, or if they are simply more comfortable using tags.
Generally speaking, the rating/collection is still not a stable and established feature that is fully part of the contribution workflow. It can be useful in its current state, but there are several things to consider, and several improvements to implement before it becomes fully part of Tatoeba's features.
[not needed anymore- removed by CK]
It would be nice if there was a page just to proofread sentences. ;)
I agree with this. An idea I could think of for this could be a page accessible from one's profile and similar to the "translate [user]'s sentences" one, but with the rating options instead of the possibility to translate/edit/link/delete. Maybe it could be called like "rate [user]'s sentences" in the menu
This is achievable by simply making an advanced research with one's username, isn't it? Or am I missing something in your explanation?
That could work, but only if the user has no more than 1000 sentences. With something like the "translate [user]'s sentences" page (what I have in mind about something like this), there isn't such limit, and you don't miss sentences because the user owns more than 1000. In addition to this, imho, this could avoid the typing or the copy-paste of long/complex usernames, so maybe it could be quicker (everything is just a personal suggestion on this general idea)
right now I'm seeing that the translation page for a given user works, but maybe implementing a way to exclude the ratings of a given user in the advanced search could speed things up