User talk:CSDCheckBot/log

Here are some feedback on the entries as of now:

Brethren - Mistery and Wizdm (3 entries) - The bot needs to do a better job at handling cases when an article goes through multiple tag addition / removal cycles. In particular:
1. It needs to ignore bots. The bots are not adding the tags and it should not leave messages for bots.
2. It is listing the creator as the one who removed the speedy tag. While this is correct for some of the removals, it should not be leaving messages based on the article creator removing the tag. The article creator is not supposed to remove the tag.
3. It missed the very first person to tag it (Intelligentsium).
4. The final tag removal was done by Cunard which was not captured at all.
Air Filter (band) - Another multiple tag addition / removal case. Again, it's the article creator that removed the tag. It also appears to have missed the fact that the tag was restored and the article was deleted per the criteria given. This appears to have been tagged for 11 hours before being deleted.
All American Security Systems, Inc. - This is an interesting one. It was tagged as G11. The G11 was declined on grounds that it had been previously prod'd (which is an odd rationale). It was then AFD'd. It was then tagged G11 again and then deleted as G11. The original tagger was correct and the remover was incorrect. It also missed the edit summary from the edit that removed the tag. The remover first removed the tag with an edit summary and then edited the article without a summary. It's reporting the second one when it should be reporting the first.
Delta Chi Lambda - Another multiple tag addition / removal case. However, in this case it's listing Ndchil as the tagger which is incorrect. Ndchil was the article creator and edited the article after tagging, but never added the tag.
Derrick Cobey - In this case, Someguy1221 moved the article from here to Wikipedia talk:Articles for creation/Derrick Cobey and then put it on hold per notability concerns (which is what it was tagged for). To avoid these mistakes, the bot should ignore cases where an article is moved out of article space.
Greg Rzab - Another case where it's getting the tagger and remover incorrect. Masterseru is the creator. Wuhwuzdat tagged it A1 & then later removed it as Masterseru had expanded the article. Cradleofrock tagged it G12 and Graeme Bartlett deleted it as G12. The bot needs to get the tagger correct.
Jamies Elsewhere - This was tagged A7, but deleted G12. Copyright trumps notability and it should be ignoring this case. Simply because it was deleted as a copyvio doesn't mean the db-band was incorrect.
Jerkin' - This was tagged as G3, but deleted as G4. I'm not sure the G3 was appropriate looking at the content, but the bot has no way of knowing that. I don't think the bot should leave messages in the case of a G4. The tag could have been correct and the fact that the tagger missed a previous XfD discussion is not a big deal.
Nsw mexicans - Another case where the article creator removed the tag. The tag was subsequently restored and the article deleted per the stated criteria.
Retrosbk - The bot needs to recognize when the speedy deletion tag was replaced by a hangon tag. It's listing the article creator as the one who removed the tag when they actually incorrectly replaced the tag with with the hangon. The hangon was removed by Graeme Bartlett and that is what the bot should have captured.
Vinrad - This was tagged per A7, but deleted as G3. I'm not sure the bot should leave messages in this case.
Weihenstephaner - It incorrectly reported the remover & edit summary. It was actually removed by Dank. Decltype was the next editor after the the speedy was declined.
Wizdm - Another case where the creator replaced the speedy tag with a hangon. The article was deleted per the specified criteria and no message should be left.
WSVAS - Notability and spam were both valid.

While I looking at these, I noticed several cases where an article was tagged for speedy without the use of an edit summary. Since the bot is already monitoring for speedy tagging, it might be worth adding to its scope leaving a reminder messages to such taggers that using edit summaries for deletion tagging is recommended practice. -- JLaTondre (talk) 01:04, 29 May 2009 (UTC)[reply]

1. First off, it will be ignoring editors who replaced the tag with a revert-like edit summary altogether, which should catch bots. I suppose I can have it ignore bots too specifically, but I haven't done that yet. Second, sometimes the article creator will remove the tag and it will stay removed. Regardless of whether or not it was proper for the article creator to have removed the tag, if no one else sees fit to replace the tag, then the tag can effectively be termed incorrect, so I am not going to have the bot discriminate in that fashion. As for why that removal was listed as the final one, I have no means of determining whether or not a tag that has been removed permanently if a non-admin removes it except to have the bot wait to see if the tag gets put back. As it stands the bot waits an hour; I'm not sure how much longer it makes sense to have it wait. I suppose I'll push it up to two hours, but the longer it waits the more bandwidth it wastes checking edits and the more time the messages regarding the removal are delayed. I could have it wait longer in the case of an article creator removing it specifically, but I'm not going to have it wait forever. As for the final tag addition and removal, the bot had stopped running by then. As for the initial tagging, it thought AnomieBot had added the tag when in fact Intelligentsium had, and that's due to an underlying problem I have not yet found the solution to.

2. See answer to 1. If the creator removed the tag and it was not replaced for an hour, then the bot acted as expected. It was deleted after the bot had stopped running, so no other taggings would have been reported here.

3. If the tag was gone for an hour the bot acted as expected. It would have worked fine if everything you had described had taken place with the tag never removed for more than an hour. As for the initial tagging not being reported correctly, there's that weird bug again.

4. Without being able to see the history of the article I can't really comment here, aside from saying yet again that I am aware of the bug about the mixing up of the tagging.

5. It already ignores it when the article is moved to userspace. I had not anticipated people moving articles to Wikipedia space, I'll just have it ignore any article that is moved out of article space at all.

6. It's reporting the initial tagging as A7, not A1. Are you sure about this? I know about the problem where it mixes the tagger up, it should not be possible for it to mix up the criteria.

7. I had realized this, and had been planning on having G12 deletion ignore tagging by most other criteria. Just hadn't gotten around to it yet.

8. I'll add G4 to the list of deletion reasons that don't cause messages

9. See prior answers

10. This might cause some issues, such as an admin removing the tag and missing the hangon, and having the declining edit not be the one the bot captures, but I'll try this during the next log run. Having the bot catch the edit summary that the declining admin intends as the declining summary is important, but I'm not sure there's any way to make sure this happens 100% of the time.

11. This is not one of the pairs the bot is currently set to conflate. I don't think I'm going to make it one either. Whether or not a bio qualifies as vandalism is less of a judgment call than the other pairs I have conflated, and a bio which is vandalism should be tagged as such because it is more important to delete than your average A7.

12. The bug I've already mentioned.

13. See prior answers

14. I can't see the article in question, but I find this highly difficult to believe. Spam typically means the article is saying how wonderful the subject is. A7 requires there to be no claim of importance/significance (not notability; notability is irrelevant to A7). What kind of spammy article fails to assert the importance of the subject it is lauding?

The idea about messaging people who tag without an edit summary is a good idea. I'll see about adding this to the bot's messaging job after the present job is approved.--Dycedarg ж 03:17, 29 May 2009 (UTC)[reply]

I gave the history above to give you details to work from. It was not intended to be a nitpick, but rather to give you data to help you solve the problems.

Article Creator: I think a lot depends on the wording of the messages that are left. It is not uncommon for an article creator to remove a valid speedy tag and for it not to be caught for quite awhile. So stating that if it's been removed for only an hour "then the tag can effectively be termed incorrect" is false.

6. I double checked. It was initially tagged as A1 then converted to A7. I missed the conversion in the first look.

11. In this case (and probably most cases like this), A7 is typically going to be more valid than G3. A bogus bio still can be non-notable and vandalism is a higher standard. I'm not sure we want to encourage tagging of vandalism over non-notable which is what the bot would be doing in effect.

14. I don't know why you'd think that. Wikipedia gets tons of attempts to create articles to advertise small companies. There are no claims of notability in them and just a "come buy here" message. Spam has nothing to do with how wonderful something is, but whether it is trying to sell something.

-- JLaTondre (talk) 12:40, 29 May 2009 (UTC)[reply]

As I've said, I appreciate any and all feedback. I don't view it as nitpicking, and I'd appreciate it if you'd continue to bring these things to my attention as the tests continue.

Article creator: I'll have it wait longer if the remover is the article creator, but I'm not having it wait forever.

11. Hmm. I will have it not send a message if someone tags for A7 and it is deleted as G3, but it will send a message if someone tags for G3 and it is deleted as A7. As you said, vandalism is a higher standard and if people tag things as vandalism when they are legitimate A7s that's a problem. No need to message the taggers if admins are doing it though.

14. As I said, I can't see the article, but maybe your definition of spam is more generous than mine. Or you have a higher standard as to what constitutes a credible claim of importance/significance (again not notability, notability has nothing to do with A7) is more strict than mine. Also I'm not sure what size has to do with it, a small company can claim to be important just as easily as a larger one can. In my opinion G11 is a bitey criteria anyway, and if an article can be tagged or deleted per A7 as opposed to G11 it should be. And I'll note that according to the debate on WT:CSD there are at least some people who agree with me. Regardless this discussion isn't going to go anywhere if I can't see the text of an article that you feel meets both A7 and G11.

The current logging run has some issues. It's not declaring anything deleted as G12, G4, or G7 as successful like it should be due to a typo introduced when I added G4 and G12 to it. Also, I forgot to add the hangon tag like I said I would above. After the next time it updates I'm going to stop it and work on it some more.--Dycedarg ж 19:28, 29 May 2009 (UTC)[reply]

Misc

Display

Width: the width of the date column should be increased from the automatic. Otherwise, it is the column that limits the number of entries displayed at a time, since it seems to default to 3 lines. This will take some playing, and of course what is desirable is affected by monitor size; I will do some testing with the sample.
Ideally, The first row should be fixed, so it stays at the top when you scroll; I think at present this takes making the header as a separate table (and this of course complicates the sorting) The html elements necessary to do this easily not seem to be supported here.
In the sort, G10-12 should ideally sort after G1 thru 9, but I'm not sure it's worth the code.

Multitags

Multitags--I will often tag both A7 and G11, if both are relevant, to give a fuller indication to the user--how does it handle this.
Changed tags: if the tagger places one tag, and I change it when I delete, do we see this?
What if I add another?

Also, and particularly important,

Does the table detect the cases where an admin deletes without a tag being previously placed? It would be nice to have a way of sorting out when the same admin places the rationale, and then deletes. I an test this if you let me know when to do it. Alternatively, we could do a separate bot. It might take admin privileges--I'm not the ideal person for this, as I've never done a bot. DGG (talk) 18:56, 5 June 2009 (UTC)[reply]
Can we get the date o fthe first article edit, to detect articles deleted as empty etc. 1 oe 2 minutes after creation?

Display

This was supposed to be more of a test of the bot's abilities than a finished product, so I did not put a lot of effort into figuring out the best way to format it. I agree that the presentation could use some work. I don't really know if either of the second suggestions are feasible; I don't know how much flexibility there is in the table code at all as I don't use it all that much. The first one shouldn't be too hard though.

Multitag

This isn't really handled at all at the moment, as I did not anticipate people tagging with more than one tag. I suspect I could make it handle it properly easily enough though.
If you change the tag via an edit before you deleted it, you would show up in the log as another tagger. The original placer of the tag would get a message if his tag was deemed incorrect by the bot; this is not affected by your changing the tag.
This is not handled as I mentioned in the first note, if I fixed the issue in the first note this issue would be fixed as well.

Last two points

No. This was designed with assessing the work of taggers as the focal point, and only follows articles that have been tagged. I could alter it to follow all cases of untagged speedy deletion as well, but that would have to be done once I've finished with the primary task.
This would be easy enough.--Dycedarg ж 20:05, 5 June 2009 (UTC)[reply]