One of the best suggestions I and others have made to pangoSE regarding this
proposal is a very strong use case or solid, easily-grasped
geographically-based examples of a problem (exclusively or largely unsolvable
in OSM today, with today's data and tools) that would make for a solvable
problem getting solved. There is a great deal of effort involved from
presenting "a solution" to the larger OSM community (first, so we understand
it, second so we might reach consensus about it, third so we might implement it
with a particular method) when no underlying problem is apparent. This is what
is meant by "a solution in search of a problem." What is it that pangoSE is so
anxious to fix that significant entanglement with a new naming system (linked
semantic wrappers) is required?
Perhaps there ARE problems that cannot be solved without such radical changes
to our naming machinery. I'm simply saying I have yet to read / hear one that
has been sufficiently articulated for me to consider this proposal further.
If problems are identified and articulated, that's a good and necessary next
step. But then so would be the greater buy-in of a well-presented proposal
that engendered sufficient discussion and perhaps eventual wide consensus to
proceed with the detailed and accepted proposal. We are a long, long way from
any of this. Let's start with what might be broken or difficult or impossible
to solve with what we have now and go from there.
I'm not saying OSM couldn't benefit by such a scheme (I keep calling it "Web
3.0-flavored" and maybe I'm right, maybe not; pangoSE chiming in about whether
his proposal and elements of Web 3.0 overlap or not is very much appreciated).
I am saying, let's have it presented to the community in a way that is usual,
potentially successful, "problem first, solution second," bite-sized in a way
that makes comprehension widely accessible and solves "something" (rather than
as it appears now: a hive of snarls that looks like deliberate obfuscation by
high priests of special knowledge). Clearly-stated concepts of what this might
solve must come first. Presenting a technical solution without articulated
problems it might solve is backwards.
OSM now has an existing "history of object edits." If you "do it right," it is
technically possible to leverage this into what you are proposing ("tracking
objects" to "follow" them?) with absolutely no change to OSM's present database
model. Maybe this is a good idea, maybe not. But pangoSE has not even
identified any costs that wold be associated with changing OSM's database
model, he simply sent us a link to it (which we can find ourselves, but thanks
for the effort).
pangoSE: please stop ignoring me in these threads. I'm extending effort to
listen, your lack of reply seems disingenuous.
SteveA
_______________________________________________
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk
