Might be an idea just to put some idea of progress here though, for those without access. Even if its just bullet-points occasionally.
Currently theres a list of pro's / con's summarising much of the conversation elsewhere; XMPP Advantages: XML Based (almost anyone can parse it) Already in use in other Wave–related protocols reduces the quantity of things a prospective wave developer has to learn reduces the number of dependencies (at least for wave servers, the clients don’t care about the part that already uses xmpp) great for mini–servers, such as one might tie into an existing web application to integrate waves as a data model for discussion (forums, blogs with comments, so on and so forth) or as a new data model for anything else real–time and with multiple contributors (code editors, other crap?) Disadvantages: Can directly extend FedOne's code, so we’ll have to touch Java… and Google–written code… *shudder* Just because we could extend an existing codebase doesn't mean we have to. There are lots of mature XMPP libraries in various languages (e.g. Ruby, JavaScript) which mean we don't have to go anywhere near Java unless we want to. JSON Advantages: Fairly efficient and compact, much more so than XMPP. Parsing tends to be faster as well. Already in use for existing wave client/server communication (but not in any usable form) Protocol Buffers Advantages: Most efficient, compact Disadvantages: Requires c++/Java/python libraries, meaning that other languages would need to re-implement the proto buffer Non-Readable(Binary) serialization/deserialization plus API. --- Following that theres some discussion on the JSON speeds/myths/zipping, and what seems to be a consensus on XMPP 2009/10/16 Devyn Cairns <[email protected]>: > Google Wave just moves way to fast for that to be useful... it allows us to > collaborate instantly. To put an archive somewhere public would be useless, > because things really do update that fast. > > On Wed, Oct 14, 2009 at 5:10 PM, Dave <[email protected]> wrote: >> >> I searched, but couldn't find it - is it a wave thread? Unfortunately >> not all of us have wave accounts yet... Could someone post an update >> of that discussion here? Ideally, I think it would make most sense if >> protocol discussions took place here - both for the wider audience and >> to ensure it's archived together with other protocol discussions, and >> findable from the wider web. >> >> My interested is in developing a thick client with good offline >> capabilities, and am currently looking at xmpp for c-to-s. My aim >> would be to maintain similarities to the s-to-s protocol where >> appropriate, and diverge where usage will be different. >> >> Later, I'm interested in maintaining server-side contacts, shared >> contacts, and other capabilities. >> >> I'm in favor of action over purely theoretical discussion - so believe >> we should prove any c-to-s proposals asap through iterative >> implementation, rather than delaying implementations until agreement >> on a 100% complete paper protocol. >> >> All the best, >> Dave >> >> On Oct 15, 12:34 am, Dale Francis <[email protected]> wrote: >> > There is a new thread for the discussion, created by myself and another >> > guy. >> > it has been split into multiple threads and is better organised so this >> > wont >> > happen and its more specific discussion >> > It's called The new client/server protocol discussion, everyone in this >> > group has access just search for it :) >> >> > > > > -- > ~devyn > > > > --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "Wave Protocol" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/wave-protocol?hl=en -~----------~----~----~----~------~----~------~--~---
