[Geowanking]RDF help sought

[Richard Soderberg]

On Monday, May 12, 2003, at 10:28  AM, Daniel Smith wrote:

>>
>> Would it not be appropriate to simply embed the full RDF into the 
>> photo
>> rather than having to invent a new URI format for it,
>
>
> It's not just about photos.  Some of it is about all the docs, binary
> or not, that you CAN'T go back and change.  If you
> want to do EXIF-hacking, cool.  It's just one little
> niche in the docuspace, and it's a nice solution for
> that niche.

This would be useful for annotating the notes from the MeetUp groups, 
too, so that we could show them on a map; it'd be useful for annotating 
blog posts, as long as you don't mind discovering the awful truth: I do 
most of my writing in a single place, the same place that my server is 
residing at.

I'm working on some sort of mental program involving keeping a list of 
geostamp and URLs, and then autorewriting geostamped URLs to point to 
the proper content -- and revision of content, if necessary.  It's 
still all in my head, though..

> It's funny, I'll explain ThereThen addresses to a lay-person,
> and they will get it right away.  They know it doesn't
> solve every problem, but they see that it does
> a simple intersection of interesting things (location
> and time) pretty well.

I suspect that's what happened here, minus the actual positive feedback 
aspect; there's this whole ongoing trend of the Internet to "cancel 
out" positive affirmations, like "hey, good idea, that makes sense".

> I'm getting frustrated at trying to explain it to a more
> ivory tower/secret handshake clique.  I've asked for
> help with RDF.  I've asked people like Jo and Tom
> Coates to give it a look.
>
> Nothing.  No feedback.

Drawing on my experience from much harsher mailing lists than this, I'd 
say that you've got tacit approval right now; the EXIF discussion was a 
good point, but it turns out that you were using photos as a nice 
example -- there's power in not restricting it to that kind of data 
type.  Keep onward -- I'm certainly watching :)

> I'll forge on.  I think I have it pretty well thought out,
> and though it would be really nice to have some peer
> review, perhaps I am asking way too early.  Fine,
> demo code it'll be.

Demo code is always good. Code for.. what?  Generating URLs?  Mapping 
them?

>> Wouldn't it be better to embed the long lat as a parameter passed to 
>> a cgi
>> rather than directly in the URL itself?
>> http://example.com/spatialsearch/?lat=37.75&lon=122.45...
>
> I don't assume a cgi.  I assume something like mod_rewrite,
> so that URLs aren't cluttered with obvious arguments (which is
> something that affects search engines)

mod_rewrite would be a nice route; using the PATHINFO route with a 
handler (mod_perl, anyone) would work well here, too.

> Also, it's not necessarily a query.  It's a direct alias
> for an existing doc, like going through a symlink in
> a filesystem.  The server side could certainly take
> it to mean "give me the closest thing you have
> to this location and time", but that would be
> a local implementation (or a fallback strategy, when
> there is no direct mapping TT -> original available)

It could be used, perhaps, to access older revisions of documents.  Tie 
this into a wiki's revision system, and you can access a specific 
version of a document by a stamped URL.

  - R.