[Tagging] service=drive-through or drive_through?

Mike N niceman at att.net
Wed Jun 29 14:53:26 BST 2011

On 6/29/2011 9:31 AM, Jonathan Bennett wrote:
> On 29/06/2011 14:19, Mike N wrote:
>> I don't see these edits as out of line or unusual. It's not so
>> different from the dozens of other projects to create more unified
>> tags so that data consumers have a chance of using the right tag.
> I suspect the "tags" you're talking about in other projects don't have
> quite the same significance as they do in OSM. Can you give us an
> example of what you mean?

    xybot for example.

> I see bulk-changing one tag to another in this way as being equivalent
> to changing a method name in an open source library without changing its
> functionality, just to make the name nicer. Anyone using that method in
> their code will get a compilation error all of a sudden, but nothing has
> actually improved in the library. You break some people's use of the
> data without having a net benefit.

   Which data consumers actually used this tag, and did they use the 
Wiki form, the last tagging list discussion, Editor Presets, or just 
invent their own idea of how to consume it?

> To put it another way, if the edits could be done using a simple
> algorithm, they haven't added anything to the OSM data itself, since
> that algorithm could be applied as post-processing. It's just
> rearranging deck chairs.

   This is an argument that tags should never be unified, but somehow a 
colossal document that lists every possible alternate tag be created. 
Data consumers would also need to implement such a document in code.

   On the other hand, an established tag that is clearly widely used by 
data consumers must not be changed without the agreement of data consumers.

More information about the Tagging mailing list