|Think about Loose Coupling|
Deferring a consideration decisionby grinder (Bishop)
|on Nov 15, 2004 at 09:04 UTC||Need Help??|
There's a new CUFP, Cool Uses, re-used that, as of this instant, is unapproved. It's a web scraper that lacks a significant sleep time in its loop, to stop it from hammering the target site (pm.org itself as it happens) into oblivion. davido has already pointed this out, so ++ to him.
I've noticed this problem in the past and have forgotten to bring the issue up, so I'd like to put forward the following proposal before I forget about it again.
My proposal is that Consideration lacks a category, namely, Hold. A held node is not approved, and so is not visible to AMs. It can no longer be considered for approval, which prevents someone else from approving it. This latter point is important. Getting back to the CUFP node in question, I read the code, recognised it as a scraper and saw that it didn't play nice, so I won't approve it. No-one should approve it, but it only takes one person to not pay sufficient attention and it will be approved. Given the number of people on board, the probability approaches 1 after a time.
The current situation therefore relies on everyone doing the Right Thing, but I think that's too much to ask. The perlmonks hive mind works well in nearly all circumstances, but I think this is one area where it falls down. I freely admit to having approved nodes that should not have been in the past. Had someone more observant than I already seen the issue, they could have put a hold on the node, and therefore I wouldn't have made the mistake.
The "Hold" option would be a way of saying that yes, the node is in the right category and on topic, the basic premise is correct, but it needs a bit more work before getting the green light.
I toyed with the idea of putting the node up for consideration with "edit: the code needs a sleep() to play nice" but that seems a little too interventionist to my taste. I envisage a hold status as being something the author can fix herself, and when having done so, could petition the janitors to put the node up for consideration again. Admittedly, this is difficult to deal with in the case of an AMm but there are many janitors these days, and one is usually online at any given time of the day.
I also really like the term "punt" instead of "hold", but I suspect that the meaning would be lost on a lot of people (dictionary.com says "A design decision to defer solving a problem, typically because one cannot define what is desirable sufficiently well to frame an algorithmic solution").
- another intruder with the mooring of the heart of the Perl