[OSM-talk] Validator tags

Matthias Julius lists at julius-net.net
Mon Nov 24 19:02:12 GMT 2008


DavidD <thewinch at gmail.com> writes:

> 2008/11/22 Matthias Julius <lists at julius-net.net>:
>
>> Whatever the syntax is you won't know why someone chose not to include
>> a name tag.  The here discussed proposal is only about telling the
>> validator not to complain about the missing tag.  It is up to the
>> mapper to be reasonable and not just include ignore tags because all
>> those maplint warnings are too annoying.
>
> I don't really see why this is needed. Wouldn't it be a lot better to
> have a tag for when a street has no visible name sign? Then I and
> other mappers know what is going on and there is no risk of someone
> with local knowledge thing the street is already named.

People can put that information into the note tag.  The ignore rules
are not meant to be for that case. 

>
> I remember when osmarender used to render all street names in the same
> font size on all streets even if the streets where too short for the
> name to fit. As a result the osmarender:renderName=no tag was often
> added to small streets to make the render look nicer. Then osmarender
> stopped adding names to short streets and could use a smaller font
> size on smaller streets. As a result many streets that could have had
> a name rendered did not and all the osmarender:renderName tags needed
> to be deleted or where pointless. Someone then had to go around and
> delete all the useless tags.
>
> I suspect the same mistake is being made here. The validators will get
> smarter and the tags will end up pointless or harmful. Tagging for
> validators should IMO be viewed the same as tagging for renderers.

I agree with you here.  One has to be very careful when using those
tags.  And they should not be used to circumvent shortcommings of the
validators.  But there are cases where the validator simply can not
know that it has a false positive.  IMHO the usefulness of validators
increases significantly if those false positives can be suppressed.

Matthias




More information about the talk mailing list