hi re, sorry dropped thread, this is a longish one, more openguides
stuff below...
> share it among 3 people. i think we're going to get a lot of that. does that
> make any difference?
nnng; perhaps, in a worst case; if i start reasoning about things
based on those properties which i acquire from you; if i had other
information about more than one of those people, if more than one of
them became associated with that address, i might accidentally
collapse them into a multi-limbed, polypresent uber-person.
This, however, is largely my problem.
So if i start using a squishing algorithm to create identities from
FOAF data while you're publishing unique identifiers for non-unique
things; which are identifiers that i'm likely to hear repeated from
other sources... well, i can be more cautious about what i accept from
the world and start making new conclusions with. And you can be
cautious about what you emit... but do you want people to have to set
a 'this is a group account' flag and not emit email addresses? Perhaps
it's just okay to not worry about it.
> > right, i didn't realise the RDF/xcal representation was going to be
> > somewhere else as well. if the changes.rss could offer rdfs:seeAlso
> > links to fuller RDF representations that would be ideal.
> okay. i'm not too clear on syntax around this kind of thing. given that we want
> to retain some kind of pubDate, and have seeAlsos, etc, what would an <item>
> look like now?
well, i don't know what predicate/property to recommend using to
indicate pubDate as a concept in RSS1 that doesn't affect the
semantics of what you're talking about. perhaps just having a pubDate
/ ical:datetime / whatever on the item, and differentiating between
the item (the uri of the page about the object (the event))
and the object actually under discussion.
Have a look at:
http://london.openguides.org/api/rdf/category/Bookshops
That features an RSS feed with seeAlso links which are to the RDF
version of each URI. If you look at
http://london.openguides.org/index.cgi?id=Freedom_Bookshop;format=rdf
You see it differentiates between properties of that URI, the page
containing information about the thing, and the relative URI #obj
which is the URI which refers to the actual thing in itself.
Does this make sense? It is a nuanced issue the background of which
you can dig at with a search for "HTTP-Range-14". It may sound a
little obtuse but i found it later helpful when modelling metadata
properties of things (people, spaces), rather than descriptions of
or references to things (wiki pages, web service URIs)
Re. the broader hookup between OGL, EVNT and wirelesslondon:
- We'd get a feed from EVNT anyway, of stuff we'd been able to reason
was nearby. for WirelessLondon, we'd probably want to consume
everything you thought was in London, up front. We'd want to share
this with the Open Guide to London; and OGL will offer a kind of fuzzy
node-coder service, so you can send an address string, see if OGL has
anything that matches your place/venue, and attach your evnt to it;
this should be one-click for the user; then easy to POST new pages,
which perhaps can be wiki-gnomed and made into redirects later, if
auto-generated page node titles aren't to the Guide admins' tastes.
- Someone logs into Wireless London via the portal service. They could
do this via the web as well; we don't *force* them to authenticate
to the portal, they can pass straight through if they like. Each node
knows where it is, and shows a listing of local resources from the
Open Guide; ideally, EVNTs that we've managed to geocode within 1000m.
Once the user logs in, we have their personal information that they
gave us when they registered; it is just part of a big old triple
store into which we are also aggregating RSS, geo, FOAF data. If we
have mbox_sha1sum annotations from you, we say, "here's an event
you've annotated, therefore this is probably your evnt account, log in
here".
If we can really get people using this service - i've toyed with the
idea of *obliging* the Season of Media Arts for London organisers to
use the portal service, they don't get access to their tools via the
web - this means they have to go socialise in a cafe with open
wireless, or set up a Wireless London portal node at home ;) - well,
we get a lot of for-free association.
We know where the user is in space, and in time, and we know *who*
they are, what information resources they are connected to. And we are
starting to know a lot more about the things around them, in space and
time and more abstract network sociality; from EVNT and from the Guide
particularly.
Because it is being accessed with a knowledge of these properties, the
system starts to refract connections in the information near them. I'm
connecting to the network; I'm in a cafe in South Kensington; my
calendar says i'm supposed to be at a meeting at the Data Centre in an
hour's time. Saul will be there, and i have a postponed mail waiting
to go to him two days now, and...
I should stop there, but perhaps others have more of that picture.
Almost without architecture, just through the edges, we can almost do
amazing things.
-jo
I've been in discussions with some people about the possibility of linking
other services to OpenGuides. One thing that occurred to me would be the
utility of providing a fuzzy node name lookup. Here's a user story:
Users of website X want to be able to associate nodes on OpenGuides site
G with items on X. They go to a node lookup service offered by G and type
the name of something they are looking for. They are then presented with a
list of nodes on G that approximately match their search term.
Specifically, I'm thinking of search results gradiated into exact matches,
partial matches and approximate matches, just like the way IMDB does it:
http://uk.imdb.com/find?q=Hideous
I don't know anything about fuzzy matching techniques. Can anyone suggest
any pointers?
--
Earle Martin
http://downlode.org/http://purl.oclc.org/net/earlemartin/
I don't know who is intending to go. The dates are 31st Aug to 2nd Sep.
I intend to go myself, and I have submitted one talk proposal already.
Since the theme this year is "Perl Everywhere", it would be a nice
place for a talk on OpenGuides.
I'm offering to submit a talk proposal. The talk will include some of the material
from my Geographical Modules talk, but also some new sections:
- Wiki spam and how to deal with it
- UTF-8 and language problems
- Copyright and creative commons
Please give me feedback before I submit the proposal tomorrow (which
is the submission deadline)
Ivor.
Hi,
The CVS repository on un has been inaccessible for quite some time now.
I would like to suggest that we move it elsewhere: I would be willing to
host it on a colo machine I run providing pserver access. This would
additionally allow anonymous access which hasn't been the case so far.
Thoughts?
--
Dominic Hargreaves | http://www.larted.org.uk/~dom/
PGP key 5178E2A5 from the.earth.li (keyserver,web,email)
Hello,
We are mentioned on your site, but the person who has put us on there is
unsure of who we are and what we do:
We are a fashion company who put on designer sales of discounted stock, six
times a year in London. We stock some of the best known fashion labels and
have been going for 16 years. We would like to be featured on your site in
either a shopping section or an events listing, or both. Please let us know
how to go about implementing this. Our website is www.designersales.co.uk.
Thank you very much.
Best Regards
Sophie Shephard
Designer Sales UK