[josm-dev] Validator

Frederik Ramm frederik at remote.org
Wed Jul 11 19:00:17 BST 2012


Hi,

On 11.07.2012 18:13, Dirk Stöcker wrote:
> The dialog already says:
>
> ----
> The following are results of automatic validation. Try fixing these, but
> be careful (don't destroy valid data).
>
> When in doubt ignore them.
>
> ...
> ----

I think that, for a newbie, the "when in doubt, ignore them" line can be 
dwarfed if there's a large number of warnings and errors.

> The idea behind this is that users actually fix these issues. When we no
> longer display them, then they wont get fixed at all. The system has
> already been tuned a lot,

Exactly. This is what the newbie will think as well: "The system has 
been tuned a lot, this editor is used by thousands of mappers every day, 
so it MUST BE RIGHT and I am responsible for all these bugs! Oh dear!"

> so that the warnings and errors have few false
> positives and information level is disabled in default.

We're not talking about false positives. We're talking about things that 
are indeed problems that need to be fixed, but we're showing them to the 
wrong person - someone who lacks the necessary time or first-hand 
information to fix them.

> As programmer I say an error is an error and it makes no sense to divide
> it into "my" error and "your" error.

Oh yes it does. Because if you have made a change that *causes* a 
problem, and then a message discourages you from uploading at all, then 
that's not too bad. But if you make a change that just happens to lead 
to *unrelated* problems being highlighted, and you then decide to rather 
not upload your change, then this is a loss for OSM.

> And a beginner getting "tons" of errors also touched "tons" of objects
> which probably is no good idea at all when you don't know what you do...

Is it not possible to touch one large object which has e.g. 20 level 
crossings with other ways and then get 20 warnings just because you 
added a tag to that one big thing? At least that's what the guy on 
talk-us claimed.

Bye
Frederik

-- 
Frederik Ramm  ##  eMail frederik at remote.org  ##  N49°00'09" E008°23'33"





More information about the josm-dev mailing list