[OSM-talk] Returning to the question of collateral damage
hillsman at cutr.usf.edu
Thu Feb 10 18:21:33 GMT 2011
On Thu, 10 Feb 2011 14:28:41 +0000, Tom Hughes <tom at compton.nu> wrote:
Subject: Re: [OSM-talk] Fwd: Fwd: Re: collateral damage (was: What the
licensechangeis going to do to the map)
Message-ID: <4D53F619.5050609 at compton.nu>
Content-Type: text/plain; charset=windows-1252
>I know of no database with any sane way to go pack and cherry pick a
>series of ancient (up to several years old in this case) transactions
>and then roll them back.
>In this case I believe the edits go back before we moved to postgres
>anyway so (a) the current database wouldn't have logs for all the
>changes even if it never clean up old logs and (b) before that edits
>were not done in database transactions anyway.
But in the specific case I pointed out, the information is in the OSM system. I went back to the "my edits" section on my OSM user homepage, and I worked my way back until I found the session in which I had created the node. It showed the node ID, information from the two sessions in which I created and edited the shop node, and information from the two sessions in which Anthony edited the node to add the street address. That's how I got the node ID and dates I included in my message earlier this morning.
Now, it may not be as easy to work with the information from this part of the system as some would like, but the information IS there and IS in machine-readable form. I can't speak to any of the other changes/deletions/losses that have occurred, but in this case, it should have been possible to identify this node as having been edited by Anthony; then go back through earlier versions to see if it existed before ever being edited by him; and then, in this case, revert it to the state it was in before he first edited it. If the creation history were different, and he had originally created the node and I had edited it, the policy would probably be to delete it. But in that case, as a courtesy to the later mapper, it would be good to send a note to the mapper's OSM mailbox indicating what work had been deleted and why; apologizing for the inconvenience; and encouraging the later mapper to remap the area using only sources acceptable to the OSM community.
This is not about my losing a few contributions; the information I need to figure out what was lost in this part of the map, and hints on how I can reconstruct what I did, is there in the "my edits" history. And this weekend I'll go through it and see what I need to do. It's probably not a big deal. But I don't know where else Anthony's work intersected mine. This means I don't know what other of my work has been removed, or where to focus. The only reason I'm aware of this particular instance is because of Anthony's posting on this list serve, and the fact that I happen to know he has been one of the handful of active mappers in the Tampa area (although we've never met or collaborated). I suspect the other local mappers don't know that some of their work may have been affected.
We refer to the OSM community, and the need to respect the work of others. The way this particular situation was handled could have done a much better job of respecting the work of others. If software needs to be modified to make it easier to show such respect, then I hope those who have the skill and knowledge to make such modifications will do so. If the organization takes action against one mapper that can cause collateral damage to the work of other mappers, I think it has a responsibility to minimize the amount of damage and, where some damage cannot be avoided, to provide information that could help in repairing it.
More information about the talk