"Thank you" answers are generally not very useful, especially if they appear in masses. Some questions are attracting especially many of them and should probably be cleaned up by a moderator. A data-dump query shows:
8 Thanks
6 Thanks
- ASPSmartUpload v3.2
- sqlite3-ruby gem: Failed to build gem native extension
- FF3/Windows CSS z-index problem with YouTube player
5 Thanks
4 Thanks
- Tomcat 6.0.18 service will not start on a windows server
- Visual Studio Development Server using wrong port
- The imported project "C:\Microsoft.CSharp.targets" was not found
- How can I integrate the Picasa API into my iPhone application?
- Consuming a web reference in a class library (C#/.NET)
- Build Error - missing required architecture i386 in file
- Tomcat 6 server creation using eclipse IDE on ubuntu
(rest elided for brevity, use the SEDE query to see more)
Used query
The above lists were produced by this query:
select
count(id),
('https://stackoverflow.com/questions/' || to_char(parentid, 'FM9999999'))
from posts
where posttypeid = 2 and length(body) <= 200 and (body ~* 'thank|thx')
group by parentid
order by count(id) desc;
(PostgreSql, ~* is a regex search.)
More thank you posts
If after these lists above someone would still feel the need to clean up even more - or just doesn't know what he should do the next month - a general query for LIKE %hank% shows many answers that don't add any value. With length(body) <= 100 basically all returned answers are not useful:
select
('https://stackoverflow.com/questions/' || to_char(id, 'FM9999999')),
body
from posts
where posttypeid = 2 and length(body) <= 100 and (body like '%hank%')
order by length(body);
This query returns 2725 records. That's too much (and too boring) to go through by hand, and a simple delete in the database would be too extreme, so I guess they are here to stay.
Cleaning them up when they start "clustering" like in the above questions is probably the way to go; single thank you posts don't add that much noise that they couldn't be tolerated.
EDIT:
Here is an up to date list: http://odata.stackexchange.com/stackoverflow/q/886 - again tons to clean up.
Related query, showing just lots of short answers by 1 rep users: http://odata.stackexchange.com/stackoverflow/q/2587
That being said
... and kind of fitting the general topic of this post I want to add:
Thank you to all the moderators of the trilogy for spending their time for the good of us all, and for doing a very good job at it! You are always there when you are needed and make the sites a better place.
thank. Can you improve that query?