[Talk-us] Special issues in LA remap

[Mike N]

On 6/6/2012 12:18 PM, Charlotte Wolter wrote:
> Hello Steve and everyone,
>
> Actually, the license redaction algorithm is not opaque and mysterious
> to me. It's unknown to me.
> And how were we supposed to find out about it when there is no
> communication? That is the ongoing problem with the license change and
> OSM in general: no communication. No newsletters, no general emails,
> nothing to let people know what's happening. Email lists don't cut it
> for us mere mortals who find them overly geeky. How about some
> communication for the rest of us?

   Like everything in OSM, you have to poke around to find information. 
   You're right that there isn't a succinct verbal description of the 
redaction algorithm.   I'm going by memory from several preliminary 
posts of the lists.   This is the only 'newsletter' about the status, 
but it's quite out of date:

http://blog.osmfoundation.org/2012/04/26/license-change-still-ongoing/

The exact algorithm could be inferred from the test suite:

https://github.com/zerebubuth/openstreetmap-license-change


    It's likely that those who provided the license check tools for the 
editors believe that this is better than trying to verbalize the 
algorithm, with all the accompanying questions when the tool plus 
history analysis tells you directly which objects have a problem.

   I've also noticed that simply reverting changesets from blars is not 
quite the best way because the redaction bot also removes the licensed 
information from the history file.

   So in summary, I think Toby's method of just moving forward with 
selectively applying new TIGER to replace bad data is probably the best 
way to proceed.