[OSM-legal-talk] A gradual transition to ODbL

FK270673 at fantasymail.de FK270673 at fantasymail.de
Sat Dec 17 14:20:57 GMT 2011


Even if some people do consider me as grumbler, I would like to suggest a gradual transition procedure that might be accepted by all sides. It should be called three-step transition because it consists of three steps:

1. The first-step transition is a software bot that transfers all benign nodes, ways and relations into the new ODbl database just for the mappers' convienience. Data is considered fine for automatic first-step transition if all nodes, tags and members are considered clean. No way or relation will be abridged or mutilated as mutilation is considered as copyright infringement. Only clean data is automatically transferred by the bot. All other data is flagged with odbl=unsure in order to facilitate retrieving them in JOSM. The OSMF says sorry for not being able to transfer this data automatically and asks users to check it manually. The OSMF states that bot execution is not a legal statement about license compatibility.

2. The second-step transition is a manual process. Data is still stored in and downloaded from the old, complete CC-database until it is manually edited by a user as users are asked to improve this data manually. When uploading data, the editor checks whether any node or tag contains original data (e.g. the original node position) from a declining or non-responding user. If so, the user is asked to repair these nodes and ways. Nodes are considered clean if they are moved, they always get a new ID after moving as required by Open Data Definition (which considers a multiple attribution of moved nodes as "onerous" attribution (§ 5) not compatible with Open Data Definition). Tags originally entered by decliners or non-responders are considered clean if a mapper changes the tag (where necessary) or clicks a checkbox to confirm that this information is fact. A mass flagging of unchanged, but unflagged nodes would be discovered by the editor, e.g. when it exceeds a certain threshold. An abuse of this checkbox cannot be prevented, but I would definitely prefer that only local mappers with local knowledge should be encouraged to use this checkbox. Really.

Benign data will be transferred to the ODbL database (and temporarily to the CC database in order to keep it consistent), tainted data will be stored in the CC database only until somebody works on it. Monitoring occasional CC-only-tainted edits gives active mappers an indication where active remapping would make sense.

As the second-step transition is a gradual process, the LWG can gradually monitor and correct individual mistakes. During this gradual second-step transition, the LWG will have to deal with one or two complaints per week instead of dealing with 10,000 complaints on April 5th. 

Mappers should have enough time to resurvey an area flagged with odbl=unsure, so the second step should allow some months. Urging mappers to remap an area within three months just in order to get it clean for a vandalizing and mutilating script creates a discomfortable pressure on them instead of allowing them to remap in time. We are told not to tag for the renderer and we should not be forced to remap from doubtful sources just in order to bypass the script in time.

Some decliners might be willing to agree if they watch a continuous map improvement instead of an abrupt transition and a tremendous data loss. If the OSMF gives mappers the opportunity to remap their area in time, it will come out of fireline. The responsability for the second-step would be transferred to the mappers who decide how much they want to remap, whether 0.01% or 0.1% or 0.5% per month.

3. The third-step transition includes the final termination of CC database as download source if the ODbL database has reached a specific threshold, e.g. 99.5% of all data and at least 98% in every country on earth. Unchanged data flagged with the tag odbl=unsure for more than 12 months will not be transferred to the final ODbL database. That's enough time for local mappers to verify and for decliners to rethink. After 12 months, all declining and non-responding mappers get a final chance to agree when the transition is obviously appearing successfully. The third stage might be done earlier if the threshold is reached earlier.

Doing all three steps in one night creates discomfort and resistance among mappers. Personally, I would prefer a smooth and peaceful transition with a minimum data loss in April 2013 rather than an abrupt cut with a tremendous data loss in April 2012. Furthermore, I am convinced that LWG and OSMF board would enjoy such a smooth and peaceful transition with only one complaint per week (instead of 10,000 complaints at once) as well. 

Benefits:
- a continuous growth in both databases with ODbL database finally reaching equal content
- remapping and verifying activity instead of meta-discussions
- many happy mappers
- some additional acceptors over time
- about 98% of all problems might be solved until April 2013 without any LWG script involvement

Burdens: 
- The OSMF needs to operate two databases during the transition which might cause a slightly higher expense and some additional hours of database maintenance.
- The OSMF loses its omnipotential power concerning the license change
- might be refused because of the NIH syndrome


-- 
NEU: FreePhone - 0ct/min Handyspartarif mit Geld-zurück-Garantie!		
Jetzt informieren: http://www.gmx.net/de/go/freephone



More information about the legal-talk mailing list