[OSM-talk] Fixme: A proposal

[John Sturdy]

On Tue, Oct 4, 2011 at 7:31 AM, Jochen Topf <jochen at remote.org> wrote:
> On Tue, Oct 04, 2011 at 06:56:51AM +0100, Ed Loach wrote:
>> Nathan wrote:
>>
>> > Would it make more sense to categorize by the tag? For example:
>> > *FIXME:name=yes
>> > *FIXME:maxspeed=verify that the entire road is 55 mph
>>
>> We don't need a proposal for this. It is such common sense that
>> people do this already where it is appropriate (i.e. where more than
>> one FIXME is needed on a single OSM element) :
>> http://taginfo.openstreetmap.org/search?q=FIXME#keys
>
> Its not common sense, its stupid. That way you make sure that your special
> fixme tag is not seen by tools that look for the normal fixme tag.

I think this reflects a problem with such tools, rather than with the idea.

> The value for the fixme tag is a free-text note.

Although there are many problems which don't fit neatly into classes
and are most simply addressed by a free-text note, that doesn't mean
that all problems are best addressed that way.

> You can just write:
> "fixme=Not sure whether the name is right, verify maxspeed". Even better:
> If you are not sure about the name and maxspeed, delete those tags. Then
> its obvious that something is missing there and somebody will add it in due
> course. No special tags needed.

This particular example is one where I'd agree that the free-text note
is suitable; but there are some specific characteristics of problems
that I think are worth marking in a systematic (tool-readable) way; in
particular:
  things that need to be verified on the ground
  stubs
  approximated routes that need GPS surveying e.g. joining two stubs
that you know connect (perhaps rare)

__John