Wikipedia:Bots/Requests for approval/SmackBot 3b and 3c Source: en.wikipedia.org/wiki/Wikipedia:Bots/Requests_for_approval/SmackBot_3b_and_3c

The following discussion is an archived debate. Please do not modify it. Subsequent comments should be made in a new section.

Task: Improve exisiting ISBN cleanup routine

• Insert correct hyphenation according to the International ISBN Agency.

Method: Automatically using AWB.
Speed: 1-8 per minute.
Number: Several hundred per month (estimated), plus backlog.
Frequency: Approx every data dump
Duration: About five hours, plus occasional runs to clear backlog.
Testing: Several hundred test edits have been done.

• Both tasks approved for trials, post comments/difs/etc when ready. — xaosflux ^Talk 02:51, 22 August 2006 (UTC)[reply]

C. 75 diffs here http://en.wikipedia.orghttps://demo.azizisearch.com/lite/wikipedia/page/Special:Contributions/SmackBot.

Rich Farmbrough 09:12 23 August 2006 (GMT).

Go for it! approved. Martin 10:30, 27 August 2006 (UTC)[reply]

Thanks. Rich Farmbrough 14:05 27 August 2006 (GMT).

Task: Check ISBN valid, if not label article. Method:

• Scan database download using custom perl code and check for invlid ISBNs that SB can't fix (length, checksum, character set).
• Label automatically using AWB.
  
• Check manually for timing differences. (I.E. ISBN corrected or removed between dump and bot run.)
  

Speed: 1-8 per minute.
Number: A dozen per month (estimated), plus backlog.
Frequency: Approx every data dump.
Duration: Less than an hour.
Testing: Still under development.

• Both tasks approved for trials, post comments/difs/etc when ready. — xaosflux ^Talk 02:51, 22 August 2006 (UTC)[reply]

C. 100 diffs here http://en.wikipedia.org/w/index.php?title=Special:Contributions&limit=100&offset=20060823072327&target=SmackBot

Note the calculated check digit can be seen at User:Rich Farmbrough\Invalid ISBNs: there is no certain way of calculating which digits of the pseudo ISBN are incorrect.

Also worth noting there are some 1,100 invalid check digits, so I would expect that the monthly crop of new errors will be 30-60 rather than a dozen.

Rich Farmbrough 09:12 23 August 2006 (GMT).

• Looks good, APPROVED. — xaosflux ^Talk 03:38, 28 August 2006 (UTC)[reply]

The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made in a new section.