[talk-au] A way to go
mrpulley at lizzy.com.au
Thu Dec 1 21:48:05 GMT 2011
I've been thinking for a while about the best way to remove CC-BY-SA-only data when the time comes. As already noted, some people have started already by deleting large areas of data and re-adding from bing (losing all the tags) - which at best may be unnecessary, and at present (as no date has been given yet for the removal of this data) is just rude. As I mentioned last night, I had added some tags to ways when I travelled through Cobar which have now been lost (e.g. corrections to maxspeeds).
When the time comes, rather than just deleting all ways with even a hint of CC-BY-SA data, here's what I am thinking should happen - it will need a bot to do this.
The bot will need to go through the version list for every single node. It should start with version one of the node (or if this is by a decliner, then the first version by an ODBL-acceptor), then for each subsequent version it should compare the new data with the previous version and overwrite it (this includes the position of the node as well as tags) - unless the owner of the version has not agreed to the new licence, in which case the information should be unchanged. At the end of this process we will have a collection of nodes with ODBL-only data. Some of these nodes will have only been edited by ODBL-decliners - these will have no information (no tags, no positions) and will be deleted. Other nodes who have been edited by decliners and acceptors will have as much information as possible preserved. (e.g. decliner adds nodes on a way, I subsequently move some nodes when adding new ways - the nodes I have moved will have their positions preserved, although they would lose any tags added by decliners.)
In the same way, the bot goes through the version list for each way, for each version making a note of which nodes are included and in what order, and any changes in tags by ODBL-acceptors. At the end of this process, any ways by ODBL-decliners will have no tags (and no nodes as these will have been deleted already in step 1). Ways partly edited by acceptors will have at least some nodes left intact, and some information. (It might need someone to go back later to re-add nodes to the ways, and re-add some tags, but at least some tags will be left behind.)
In the same way, the bot goes through the version list for each relation, for each version making a note of which ways are included, and any changes in tags by ODBL-acceptors. At the end of this process, any relations by ODBL-decliners will have no ways (and no ways as these will have been deleted already in step 1). Relations partly edited by acceptors will have at least some ways left intact, and some information. (It might need someone to go back later to re-add ways to the relations, and re-add some tags, but at least some tags will be left behind.)
As an example, this is what would happen to my Cobar edits:
Decliner adds a way
Decliner-bot adds maxspeed tag + source 'default maxspeed'
I correct maxspeed (if needed) + 'source:maxspeed=voice (or sign or survey)' + possibly add tags
The bot I have just described removes CC-BY-SA-only data
If any nodes remain, at least my tags get left behind.
I add a way
The bot I have just described removes CC-BY-SA-only data (there won't be any for this way)
All my edits remain intact.
I hope this all makes sense! This seems better then what has been happening so far.
Has a date been set for the removal of CC-BY-SA-only data yet?
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Talk-au