[Tagging] Thoughts on how to replace or modify an exist/established tag (Was: Feature Proposal - RFC - sluice_gate)
simone.saviolo at gmail.com
Thu Jan 6 22:47:12 GMT 2011
2011/1/6 Peter Wendorff <wendorff at uni-paderborn.de>
> Hi Simone.
> I agree that some definition of "core features" would be useful.
> I think I wrote that in my last mail here, too.
> I simply don't see the benefit of a certificate on application side.
It's not an advantage "on the application side" in itself; it's more of an
advantage for the user. Say I'm a cyclist and I want a map: I'll prefer one
with CycleMap 4.3 over one with CycleMap 1.2 over one with no CycleMap.
Of course, this is a sort of "indication for the masses". While a
contributor or the developer of a consumer probably knows that a certain tag
has a whole set of problems or concerns, the "ignorant" user who found the
map on the internet will know that map X is good if he's a cyclist but has
no information about touristic attractions. In a sense, it's there to say
"OSM approves this map as a cycling map". While this is oversimplified, I
feel it would be the best way to favour usage of OSM-based products outside
the inner circle of the OSMers.
It's possible that OSM as a project is still in a stage that's too early for
such an idea. IMHO, such an approach would mark the transition from
"interesting and promising experiment" to "everyday-life useful tool for
> Let's consider two alternatives.
> Either the certification requires support for the most 5 important
> attributes A,B,C,D,E.
> An application decides to support A,B,C,D,E,F - and gets the certificate.
> Another application decides to support A,B,D,E,F,G and his call for a cert
> is rejected because of the missing C.
> But: Is that really an indicator for bad quality? Or - on the other way
> around: is it a sign for quality to follow the suggestions of the here
> defining part of the OSM-Community about a subset of tags?
You're correct, it's not the most absolute indicator of quality. Still it
makes sense. Think of the "cert" features as a set of things that "people
expect to see in the map". Think of the "Windows Vista Ready" (pfft!)
certification: it didn't mean that that software/device was of better
quality than another one without the certification, but if you were planning
to use it with Vista then it had better characteristics than the other one.
After all, no certification in the world is an absolute indicator of
quality, but it will at best be an absolute indicator of *that* specific
aspect. Is a car that gets five stars in the Euro NCAP tests better than one
that gets three? No, it simply passed *those specific tests* better. Still
you will know that at least in that aspect it will be better.
If application 2 doesn't support C, then it won't get the cert. If I study
engineering and pass all exams but one, I won't get the degree. I may be
better, much better than someone who does, but I won't be allowed to
advertise me as an engineer.
> What's the goal of these certificates? Advertising possibility for the
> application? "this printed map of Las Vegas supports core 1.1, cycle 1.34,
> foot 4.12 and boat4.4"? Add as much "features" to the feature list as not
> applicable to the used bbox to get more plus points?
The idea is to simplify communication between consumer developers and users,
so that users don't have to go through the wiki to have an idea of whether a
map is good for them. Of course, this mainly translates into advertising
possibilities. Then it's a matter of 1) rigorosity (the certifying organism
should check that the application actually supports features even if they're
not included in the default data) and 2) common sense on the part of the
users (I guess no-one will be interested in cycling features if they get a
map of the Atlantic Ocean, so the cert would be useless and ridiculous).
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Tagging