[OSM-dev] created_by considered harmfull

[Dave Stubbs]

On 09/08/07, bvh <bvh-osm at irule.be> wrote:
> I'd like to seek a consensus on the created_by tag. In my opinion it
> is worthless. Not only are they handles differently by each editor
> (some add them, some don't; some change them; some don't) but it
> also seems some users are confused by it and substitute their own
> name as the value for that tag.
>
> For the non-advantage that created_by brings, it is rather costly.
> I did some tests on random areas and it takes up, pretty consistently,
> around 10% of all transfers. A typical example for an area of a few
> square km's with a good deal of ways and regular classifications
> is 624k with created_by tags against 550k withouth created_by tags.
> Underdeveloped areas or those edited extensively by Potlatch alpha
> instead of JOSM the situation is even worse. I believe at one point
> I saw a case where a whopping 25% of all 'information' was created_by
> tags.

Apart from the user confusion I can't actually see any "harm".

Bandwidth is a non-issue, and I'm not aware of any current database
space problems.

If it does want removing then a db based script could probably cull
existing created_by=JOSM's without much effort.

The purpose it serves was to highlight which editor was uploading
things, if for instance there was an editor bug. But as JOSM in
particular doesn't bother to include the version number, this is
probably less useful than it could be. Otherwise the tag lets us make
random statistics about what editor people are using.