[josm-dev] Change to changeset comment handling, RfD

[Frederik Ramm]

Ulf,

> P.S: Yes, this isn't a theoretical situation. While riding my motorbike, 
> I frequently have a lot of such tiny changes spread over a huge area.

Perhaps we can try to be a bit more reasonable about this.

I have the impression that those who oppose changeset comments, 
including you, think that they are expected to write an essay.

Whereas I really think of it that way: If you've spent half an hour 
fixing elaborate footways and cycleways around a certain junction then 
of course you'll write "fixed soandso junction from aerial imagery" or 
so. Whereas, if you've been riding your motorbike all afternoon and 
added some speed limits or mountain passes or whatever in all of 
Southern Germany, then nobody in the world will object to your putting 
"added POIs from motorcycle survey" or something.

The way I like to think of this is: If you speak to another mapper later 
that day, and they ask you "what you've been up to?" and you answer them 
in one sentence - that's what you should put as a changeset comment.

I agree that if you were to add 100 POIs and upload each in their own 
changeset, it would be tedious to actually put an *individual* changeset 
comment to each. Ideally, you would keep a changeset open and have them 
all in one changeset, but if you must have lots of mini-changesets with 
only one or two changes each, then that's a case where I'd just re-use 
the same changeset comment. After all, if you *had* put them all in one 
changeset then the comment would have been "added POIs from motorcycle 
survey".

I completely agree that there is no black and white here. Forcing 
someone to enter 10 characters doesn't automatically make the comment 
good (see e.g. 
http://www.openstreetmap.org/user/anti_bot-nonsense/edits). Even a 
well-meaning user can probably write 80 characters without conveying the 
what, where, and why.

I think we should find the best way to make as many users as possible 
understand why changeset comments are useful, and supply them as good as 
they can when they use JOSM. If the best way to do that turns out to 
actually drop any user nannying in JOSM and just put a decent 
explanation in the user manual or so, I'm game.

Bye
Frederik

-- 
Frederik Ramm  ##  eMail frederik at remote.org  ##  N49°00'09" E008°23'33"