[josm-dev] Validator

[Dermot McNally]

2008/8/16 Dirk Stöcker <openstreetmap at dstoecker.de>:

>> Ah yes, so it is... It does force me to mouseover each violation,
>> though. For a large or detailed area, this is going to be very
>> impractical.
>
> Suggestions welcome.

I'd still favour my original suggestion - rather than group all of
these new errors under the single label they now occupy, group them
instead under their respective, detailed, labels. That keeps un-reffed
road errors away from, say, typoed landuse errors. But ultimately,
what I'm saying is "lose the tool tip", at least for information like
this. A good rule of thumb here, would be that tool tips are good for
followup information that might not be necessary in all cases, or for
inexperienced users. But you shouldn't put data behind a tool-tip that
is necessary in order to make use of the error message.

> You can ignore them once and for all the time :-)

I haven't yet worked out the full scope of the ignore option - but it
doesn't seem to ignore all similar cases, only the specific object I
have selected. In any case, some violations are things that you care
about (like missing refs) but can't always do anything about. (like
tertiaries, where it can be difficult to determine the correct ref).

> Don't know how make an ignore for specific tests in the test-set. But it's
> a new feature. Give it some time to mature.

For sure. Lest there should be any doubt, I really like what you've
done here. I'm just struggling with aspects of how the results are
output.

> Yes. The other validation types. Think of e.g. 30 different TagChecker
> types and one unclosed way. You wont find it in this batch of TagChecker
> stuff.

I'm not sure I see what you mean here. It doesn't matter to the user
if a crossing way or unclosed area message isn't generated by the tag
checker, even if another 30 messages are. Each violation message
represents something that the mapper needs to review and either repair
or accept in its current state. Do we really need to contain the
tagchecker messages behind an overall label? Maybe there's a technical
reason, but it feels unnecessary from a UI perspective.

Dermot

-- 
--------------------------------------
Iren sind menschlich