[josm-dev] Unclosed ways

Matthias Julius lists at julius-net.net
Tue Jul 15 16:04:18 BST 2008


Dirk Stöcker <openstreetmap at dstoecker.de> writes:

> On Mon, 14 Jul 2008, Karl Newman wrote:
>
>> Well, I was thinking you could mark it as a false positive using the
>> validator_ignore key, which could be used by maplint or other validators to
>> suppress warnings there, too. That key would only apply to a particular OSM
>> entity, not for disabling a test globally.
>
> At least I will not implement such stuff:
>
> a) Private stuff has to be private and what you think ignorable others 
> probably want to know. Every user must decide for himself.
>
> b) The tags would be outdated really fast. The validator code changes and 
> the reported warnings will be reduced and increased everytime forcing 
> updates of the ignore stuff.
>
> c) Ignore stuff becomes outdated really fast, as no editor cares for it 
> and adapts the dependencies when changing elements. To have it right, it 
> must be implemented as relation in many cases, which is pure overkill.
>
> d) I'm against moving programm depending stuff into global databases. I 
> even do not like these "osmarender:xxx" tags. Software problems should be 
> fixed in software and not in a database.
>
> Thought the idea of a local ignore list is fine. Would like to have it too 
> :-)

A local ignore list certainly could be helpful.

But I also believe to improve the usefulness of the maplint layer it
would be good to have a mechanism to ignore certain warnings.

I agree it is somewhat ugly to store that information with the data
but this is the place where it is available to all validators.  And
each validator/user can decide whether he wants to actually use it.

The JOSM validator could assemble a seperate list of ignored
warnings.  That way one could easily see if there is something bogus.

Matthias




More information about the josm-dev mailing list