As most of you know, you can’t comment on a changeset that hasn’t been closed. (If anyone doesn’t know what an open changeset is: this wiki page is a good place to start, specifically the section “Opening and closing changesets”)
But I’m wondering… What’s the reasoning behind this? Is it a good idea, and is there a better approach?
I will explain my thoughts in a bit, but first I’d like to hear everyone’s thoughts on this!
The reasoning is that anything you might want to comment on, could be gone by the time the changeset is closed. “In this changeset you have deleted the X object” - and then the user undeletes the object in the still-open changeset, invalidating your comment.
I guess that it can result in a moot comments, but the same may happen if user makes bad edit, someone comments but they were already preparing fix edit.
And it is a bit irritating when user makes bogus edit but you cannot yet comment on edit and you need to schedule a reminder. It happened few times to me.
I am thinking more of someone reviewing the whole situation a few days later. With a DWG hat on, I imagine a situation where there’s a dispute between two mappers, and I see that mapper A has commented on mapper B’s changeset, accusing them of vandalism or something, and I look at B’s changeset and I think “what is A’s problem, this looks ok” - and then only diving deeper into the changeset itself will reveal that the changeset indeed deleted something, and restored it later, and that the comment must have occurred in between the deleting and the restoring.
This situation will not occur if there are two subsequent changeset - one deletes something, has a comment that says “don’t delete this”, and the next changeset restores it. That is easy to follow for everyone.
Depends on whether you count “human error” has a “valid reason”
Most changesets are closed immediately anyway, so this rarely comes into effect. Of the popular editors, only StreetComplete tends to keep changesets open and add to it. It is an option in some other editors, but I wouldn’t expect it to be used often.
If someone encounters an open changeset they want to comment on, their options are:
Set a reminder for themselves - can be more effort than it sounds, since they need to keep a link to the changeset and remember what they wanted to say
Give up on commenting at all - “it’s not worth the effort” or “it’s too frustrating to wait” or “I’ll get to it later” (and actually forget all about it instead). I have no data, but my gut feeling is that this is the more common occurrence.
How likely is it that the changeset author noticed the same issue that the would-be commenter did? Who knows, but it feels wrong to prevent the commenter from commenting.
This is trivial to solve with something like this (lazy mockup):
You, as a DWG reviewer, will instantly know that the comment was posted before the changeset was closed and may not be relevant anymore - making the “diving deeper” easier and more obvious. Any commenter would be warned about the changeset potentially not being finished yet, which they might take into consideration.
(Edit: whoops, typing mistake in the mockup… You get the idea!)
