(reposting) I think the volume of data that is potentially in this would not be suitable for a social bookmarking system. Imagine geocoding every news article in every news rss feed out there (say 10, 000 feeds).
As for the redistribution of the rss content, the separation has one huge benefit. It means that a consumer could pick multiple annotation streams for a single rss feed. For example, I could listen t obbc rss, a geotagging stream and a mood stream and put them together (that is precisely what I want to do!). If the geotagger included geotagging and the rss but not the mood, and similarly the mood stream had the rss data, mood but not geo data, I would still have to do the alignment of differnet data. By making the separation part of the specificatoin, the whole thing is a lot cleaner and extensible - it all boils down to correct usage of guid. I'm still for something like the original file format I suggested earlier. Simplicity is very attractive in this type of endevour. Matt On 7/22/05, Joel Chippindale <[EMAIL PROTECTED]> wrote: > This sounds like it could be built on top of one of a social bookmarking > sites. > > Then you could post your (geo)tags of BBC News stories to your social > bookmarks account using the site's API...and others could use the same > API to interrogate the site for the tags that have been associated with > a particular BBC News story by asking for all the tags associated with > the story's URL. > > It doesn't look like this would be possible using del.ico.us because > it's API http://del.icio.us/doc/api does not enable you to request all > the tags/extended descriptions used for a particular URL. However > perhaps one of the open source implementations e.g. http://de.lirio.us/ > or http://sourceforge.net/projects/scuttle/ could be modified to enable > this? > > Joel > > > -----Original Message----- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] On Behalf Of Matthew Hurst > Sent: 22 July 2005 10:09 > To: backstage@lists.bbc.co.uk > Subject: Re: [backstage] Geotagging BBC news stories > > > I think the separation of the original data from the annotation is > attractive. Yes, the annotation would be almost meaningless without the > original data, but it would save the annotation owner from republishing > the original content and any considerations of legality, etc. In > addition, it would drive traffic to the source of the original content > (always a good thing). > > There are (at least) two possible approaches: a query approach (give me > your annotation for this guid) and a syndicated approach. The syndicated > approach is less of a resource strain as the provider just has to > deliver something very similar to rss, they don't have to have a query > service running, just a document server (web server). > > In terms of developing the idea, we could certainly start with getting > feedback in this forum and even have some implementations provide a > service (Davy?). But I'd also like to hear any advice about how we can > rapidly formalize this and get the idea out there along with a proposal > for an api/document format. If only we had that wiki we were all so > excited about up and running... > > As the format is simple (guid mapping to some bundle of information), I > guess we are thinking of something like: > > .. > <annotationStream feedURL=" bbc feed that is being annotated" > annotationURI="http://www.latedecember.com/newsMood.xml";>. > <annotation guid="..."> > <!-- put your annotation here--> > </annotation> > ...more of above... > </annotationStream> > > Note that the stream points to both the rss feed and the annotation > information (which ought to describe the syntax and semantics of the > element at <!--put your annotation here-->) > > Davy, if you put a file up like this, I could easily integrate it into > my data sphere thing and put up some screen shots - we could claim that > constitutes the first implementation! > > Matt > > http://datamining.typepad.com > > On 7/22/05, Davy Mitchell <[EMAIL PROTECTED]> wrote: > > On 7/21/05, Tony Hirst <[EMAIL PROTECTED]> wrote: > > > backstage@lists.bbc.co.uk on 21 July 2005 at 21:52 +0000 wrote: > > > > it seems you have the same motivation as I had when I mailed this > > > >list about RSS Annotation Streams! > > > > The whole idea sounds great. I have plans to expose some of the Mood > > News data but never thought of anything quite so detailed. How about > > an XML with the URL as GUID plus the data (rating, colour etc) ? Would > > > it need a basic schema structure? > > > > Could lead to some very interesting client applications! > > > > Thanks, > > Davy > > > > -- > > > > Davy Mitchell > > > > http://www.latedecember.com > > > > > > - > > Sent via the backstage.bbc.co.uk discussion group. To unsubscribe, > > please visit > > http://backstage.bbc.co.uk/archives/2005/01/mailing_list.html. > > > > > - > Sent via the backstage.bbc.co.uk discussion group. To unsubscribe, > please visit > http://backstage.bbc.co.uk/archives/2005/01/mailing_list.html. > > > - > Sent via the backstage.bbc.co.uk discussion group. To unsubscribe, please > visit http://backstage.bbc.co.uk/archives/2005/01/mailing_list.html. > - Sent via the backstage.bbc.co.uk discussion group. To unsubscribe, please visit http://backstage.bbc.co.uk/archives/2005/01/mailing_list.html.
