Begin forwarded message: Date: Thu, 22 Nov 2001 13:24:21 +0200 From: "Mikko Honkala" <[EMAIL PROTECTED]> To: "Gordon Jones" <[EMAIL PROTECTED]>, <[EMAIL PROTECTED]>, <[EMAIL PROTECTED]> Subject: RE: [developers] XUpdate and publish/subscribe
Hello, sorry for the delay. We are quite interested on your experiments on doing live updates to the client's DOM. This is something we also have been thinking about. Think about an SVG map that would load very fast, and would update itself with more detailed data, when you zoom into it. In your case, you've also incorporated push - technology into the framework, which is interesting in itself. We have also experimented with JMS push messages in X-Smiles. First some answers to your questions: 1) The idea is interesting. XUpdate gives a syntax (I am not familiar with) and JMS gives the transfer framework but then there is the need for the framework for ordering updates and security settings. 2) I don't really care whether a technology is W3C or not, as long as it is complete and works. It is always possible to submit something as a W3C Note. Creating a real working group is much harder. 3) The only push framework we have been experimenting is JMS. On problem with that is that it needs an centralized JMS server, and is Java centric. Few questions: 1) Can you tell how much gain did you get from using push updates compared to periodically reading the graph by the client. Did most of the graph stay untouched in updates. 2) Would you be interested to share your code as open source? And apologies again for my late answer. -mikko > -----Original Message----- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] Behalf Of Gordon Jones > Sent: 8. marraskuuta 2001 13:54 > To: [EMAIL PROTECTED]; [EMAIL PROTECTED] > Cc: Finnian Reilly > Subject: [developers] XUpdate and publish/subscribe > > > > > Dear XUpdate and XSmiles people, > > We have been experimenting with XUpdate in what we believe is a novel way, and > one which we think would make XUpdate attractive to a much wider audience. > > We use it at the client end to modify a Browser's DOM with updates "pushed" from > a data source. In our experiments, the data we update is a real time graph in > SVG format. We have used the XSmiles and Internet Explorer browsers. (In both > cases using the Csiro SVG rendering engine, JMS to deliver the updates, and > Lexus to apply them). > > We work in an industry where publish subscribe of real time information is very > important (price information from exchanges). There are a number of commercial > products in the market place that supply publish subscribe infrastructure. We > are convinced there is a latent demand for this sort of capability in other > industries as well as our own. Our experiments combining Csiro, Lexus, XUpdate, > SVG, XSmiles and a publish subscribe protocol handler using JMS have > demonstrated that this latent demand could be satisfied more openly. > > We address this note to both the XUpdate and XSmiles teams first to thank you > for the excellent technology that you have given us, and second to highlight the > potential synergy of the combination. > > We have a few issues and questions: > > 1. We would like to see the idea of using XUpdate in conjunction with a > publish/subscribe browser environment promoted more widely. Do you think this is > feasible, and can you offer any suggestions about how best to achieve this. > > 2. As far as we can see, XUpdate is not yet part of the W3C family. This may be > because it has been perceived as a esoteric issue for XML databases only. Do you > think that exposing it as a client side technology would add enough interest to > the extent that it could be submitted to W3C. Certainly we have a concern that > it could become a defunct technology if it stays outside W3C, and this concern > was re-inforced by the recent request for developers sent out on the XUpdate > mailing list. > > 3. The publish subscribe protocol will need to be standardised. It clearly > differs from http: in that updates arrive after the initial page value. Neither > is it like the current streaming media protocols which handle real time data of > a quite different sort. There has been much talk of "push" technology, and yet > as far as I know, no such protocol has yet emerged. Apart from the commercial > products mentioned above, so called "push" technology has really been based on > hidden "pull". Perhaps the Xsmiles team could comment on the feasability of > introducing such a new protocol (pshttp:?). > > Your comments and suggestions about how to move forward would be welcome. > > Regards Gordon Jones and Finnian Reilly -- ______________________________________________________________________ Lars Martin mailto:[EMAIL PROTECTED] SMB GmbH http://www.smb-tec.com ---------------------------------------------------------------------- Post a message: mailto:[EMAIL PROTECTED] Unsubscribe: mailto:[EMAIL PROTECTED] Contact administrator: mailto:[EMAIL PROTECTED] Read archived messages: http://archive.xmldb.org/ ----------------------------------------------------------------------
