On Thu, Apr 24, 2008 at 2:16 PM, Steve Hill <[EMAIL PROTECTED]> wrote: > On Thu, 24 Apr 2008, Dave Stubbs wrote: > > > > And if the occupancy is on a fish pond then it likely does > > > > How do you know it's a fish pond?
It would probably have a tag like man_made=fishpond. I don't know there's a tagging schema for that. I'm fairly sure that a piste:lift tag would rule it out though. > There is no tag that unambiguously > identifies the type of object it is. Why does there need to be? What use case does this enable? > Instead there is a whole load of tags > to identify the object, and you have to have a lot of background knowledge > about the structure of the data to know which tags identify the object type > (and thus the context of the other tags) and which tags are just describing > attributes of the object. Which you know because you're using the data. You seem to think that a level of ignorance exists, when that ignorance /can't/ exist because it if did then you wouldn't be able to process the tags at all. > > > if [piste:lift] is not null and [occupancy] is not null then: > > print "piste:lift:occupancy = [occupancy]" > > > > wow. that was hard. And also demonstrates how completely pointless the > > namespace was. > > > > How did you know that the piste:lift tag declares the object as being a > lift? That's right, you didn't unless you already had an underlying > knowledge of which tags identify the context and which don't. Your wiki page told me. Please show me the application that can do something useful with piste lift data without first knowing that it's looking for the piste:lift tag. How would you know to look for the piste:lift:occupancy tag exactly? magic? No, you know to look for it because you have defined what you're looking for -- the namespace didn't really help at all. You know you're dealing with a piste:lift and you know you want to find out the occupancy. The only situation where the namespace might help is if you randomly searched for *:occupancy and wanted to know what kind of object the occupancy was defined for. Which frankly is useless information, and a query no-one would ever make. Why on earth would you want to know, and what could you possibly do with it once you found out if you didn't happen to already know what a piste lift was in the first place? Why am I looking at the occupancy tag at all if I don't already know what the object is? > > > > > This is completely stupid - yes, they might avoid coming up with new > > > unnamespaced tags and *shock* propose new namespaced tags instead. Why > is > > > this a bad thing? > > > > > > > you snipped the or: they'll attach meaningless drivel to the start of > > every tag as a substitute > > > > What sort of meaningless drivel? like piste:lift: infront of occupancy, or climbing: infront of rock. It's garbage information. > > which is effectively what you're doing anyway. > > > > Except it is neither meaningless nor drivel. It has a meaning maybe, but it's an utterly pointless one, and ultimately it's just a waste of space. > > What you're doing provides nothing extra: I can throw it away and be > > left with the same information. > > > > No, you can't - if you throw it away you lose the context of the tags. The > *only* way to recover the context information is to know which tag to > retrieve it from, which is not something you can do from the data alone. > Thus you have lost something which is not recoverable from the data you have > - you now need to go find an external data set as well. Why on earth do I need to know the context for the occupancy tag in this situation? seriously, why am I even bothering to look at an occupancy tag if I don't know what the object is? why? I don't know what problem you're trying to solve here, but it really doesn't need solving. Seriously, I've had enough of this. Namespaces are for when you have no other context to work with, such as xslt stylesheets. On most of the tag suggestions you're defending it's just plain irrelevant. Dave _______________________________________________ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk