- From: Jens de Smit <jens@layar.com>
- Date: Mon, 31 Jan 2011 14:26:47 +0100
- To: "public-poiwg@w3.org" <public-poiwg@w3.org>
On Thu, Jan 27, 2011 at 8:15 PM, Christine Perey <cperey@perey.com> wrote: > Hello, > > During our meeting January 26 I agreed to do some research on Pachube as a > basis for understanding the needs/possible synergies between this platform > and the POI WG work. > > The specific area we would like to better understand is how to manage the > frequent requests for information, and frequent updates to information in a > database. I thought that the requirements we have for objects which are > changing locations over time might have something in common with what > Pachube does. > > I want to introduce those of you who are not familiar with it to the system. > > From the home page: > <snip Pachube intro> > I suggest that you study: <snip links> > What are your thoughts? Hi Christine, What is impressive about Pachube is that they handle a lot of sensor values internally, values coming from a huge amount of data-producing sensors to a (potentially) large amount of data-consuming entities. However, the Pachube data format is "nothing special" as far as I can see. The data delivery method is standard HTTP, with the option of getting "push" data when a predefined condition is met, also over HTTP. If we really worry about frequent, _real-time_ updates it is possibly better to look at the data transport method than at the data format. Cue Thomas Wrobel :P In all seriousness, XMPP or something derived from that could be very important for these real-time tracking applications. This means that it is important that the data format we come up with is capable of adding and subtracting pieces of data into an existing set. However, looking at the direction we're headed, I do not anticipate problems with that. Still, keeping it in the back of our heads sounds like a bright thing to do. Regards, Jens
Received on Monday, 31 January 2011 13:54:34 UTC