[OSM-talk] This needs to be voted upon. (was: Re: Separating all metadata from coordinates in OSM into a wikibase instance)

stevea steveaOSM at softworkers.com
Mon Aug 10 06:41:39 UTC 2020


Even better stated, if pangoSE is interested in a longer-term goal (and, it truly must be this if anything at all) of moving OSM towards becoming a "full member of the semantic web," fairly large things must happen.

1)  Linked data and "the semantic web" (which has been emerging since circa 2001) must become more fully-formed itself.  It has been called "Web 3.0," but think about how well-defined even "Web 2.0" is today.  (There will be lots of answers here, but you see my point:  it's relatively early days for both).

2)  Much more than a "toss a poorly-formed and poorly-stated 'what for' against the wall on the Talk board" must happen first.  Inter-academic discussion, white papers, tracks at conferences, people with ideas who publish, synthesis of already-good-ideas-and-implementations in the semantic web with what OSM's longer-term goals, and much more.  This is an intermediate- to longer-term proposition, I'd say five to fifteen years.  (Happy Sweet 16, OSM).

3)  Drop-dead-easy-for-novices integration must be nearly seamless as these begin to integrate with existing workflows in OSM.  The chicken-and-egg thing going on right now is most people have never heard of Web 3.0 / the semantic web, know nothing about it, don't see it's (admittedly longer-term) benefits and so often have hostility or confusion about "why?"  This might have seemed my initial position, and isn't really true, so I now clarify.

4)  Real-life examples of how AI, neural networks, machine learning benefits of how linked data in a semantic web must be offered for people to see the benefits.  I don't mean pedagogical aids like teaching the basics of how triple statements, literals and URIs work (though, those are important early concepts), I mean "hey, that's a really neat and powerful exploitation of why we'd bother to link data."  The first examples might be geographical in nature to appear to OSM, they might not, but they should be sooner, rather than later, so people can more immediately realize benefits.

There:  I think I've tilled the soil a bit, and if pangoSE or somebody wants to plant seeds (again), I'd read #2 above and think much longer-term.  OSM could do this, but it's going to take more than a thread on Talk and a wad tossed against a wall.  And maybe a decade or two.

SteveA


More information about the talk mailing list