sorry for broken link to ZDNet UK, just add a slash in the end, or use - http://wwwery.com/2011/05/06/p2p-on-browsers/
- I see - WiAB is pretty testable now with Jenkins We could use it for POW too. For clients i think - if protocol and API would be great - there would be example clients already written by WiAB/POW/JShare dev's by the time of public release. Basically i think we need a separate Protocol's mailing list for now, GWave is almost unusable by performance terms, Google group - maybe better then mailing list, far better - a separate forum with OpenID On Sun, May 29, 2011 at 6:14 PM, Thomas Wrobel <[email protected]> wrote: > On 29 May 2011 16:15, Perry Smith <[email protected]> wrote: >> Java and Python frustrate and scare me. Python has lots of issues between >> even minor versions. Java has issues between platforms. In both of these >> languages, I've never had a pleasant user or developer experience. >> >> I was going to suggest Ruby but didn't because I knew this was a Python/Java >> group. >> >> Would it be insane to have parallel implementations? That way, we would >> work out and clearly document any language specific details that might get >> hidden. > > Not insane - but I think we need one testable primary implementation > to deal with the "generic" bugs and issues that arise as the c/s is > developed before the implementation specific bugs. > > The more I think however I'm not sure we can avoid either java or > python - at least for the server side. We need to plug into an > existing server as I cant think of another way to develop a c/s for a > wave server. (and we dont want to have to make our own server!). Not > sure of any options really :? > > We could, however, have anything we like for client-side code examples. > > >> >> On a different topic, can you point me to the POW work? Is that using >> Python in place of Java for the entire implementation? >> >> On May 29, 2011, at 8:50 AM, Thomas Wrobel wrote: >> >>>> Can this be done as a very well documented and commented piece of code >>>> that actually runs? I can understand > code far quicker than I can >>>> understand TechSpeak. >>>> >>> >>> +1 >>> >>>> Pick a language like C (not Java or C++). Something that clearly shows >>>> precise intent. It can be a pseudo >>>> language but then we can't test it by running it. >>> >>> I essentially don't know any C, but I certainly approve of usable code >>> so I guess I could try to learn unless as nothing too language >>> specific is needed. >>> >>> In the end though someones going to have to convert it to Java needed >>> for wiab, python for POW and Javascript for webclients side no? >>> Downside of C for a c/s example lib might be no easy testing as theres >>> no server written in C? >>> >>>> >>>> After we're done, we'd not only have a spec but also something useful -- >>>> working code. >>>> >>>> On May 29, 2011, at 7:09 AM, Thomas Wrobel wrote: >>>> >>>>> Well, thats the problem, I haven't either ;) >>>>> I'm currently contributing a bit to the w3c POI standard, but its more >>>>> general advice on whats needed/useful for AR then solid contributions. >>>>> My experience is pretty low really, feeling my way. >>>>> I also don't know anything really about protocols beyond my own bespoke >>>>> stuff. >>>>> >>>>> regarding the name; I'm not sure thats such a good idea as its a bit >>>>> confusable with the "wave federation protocol" itself no? The c/s >>>>> standard might be similar in some ways but it wont be the same. >>>>> >>>>> >>>>> >>>>> On 29 May 2011 13:24, Adrian Cochrane <[email protected]> wrote: >>>>>> Oh, and Thomas Wrobel, I'd appreciate your help. I've never written a >>>>>> real standard before. >>>>>> >>>>>> How about dropping the "Con". "The Federation", less Firefly more Wave. >>>>>> -- >>>>>> Adrian Cochrane >>>>>> [email protected] >>>>>> >>>>>> >>>>>> On Sun, 29 May 2011 04:05 -0700, "Adrian Cochrane" <[email protected]> >>>>>> wrote: >>>>>>> Well, I just thought that if the name Wave came from Firefly, so should >>>>>>> it's concertium. >>>>>>> >>>>>>> To be clear, I'd take the task of reworking the standards by placing my >>>>>>> current plans online and taking all the criticism I can. >>>>>>> >>>>>>> As for using the original standards, it's just because then I wasn't >>>>>>> reworking the standards. As for Federation, I'd like that to be simalor >>>>>>> to the current standard (since it's the architecture of PyOfWave). >>>>>>> -- >>>>>>> Adrian Cochrane >>>>>>> [email protected] >>>>>>> >>>>>>> >>>>>>> On Sun, 29 May 2011 12:54 +0200, "Thomas Wrobel" <[email protected]> >>>>>>> wrote: >>>>>>>> Like call it Moya then, from Farscape ;) >>>>>>>> (hay, it did last longer....) >>>>>>>> >>>>>>>> On 29 May 2011 12:52, Paul Thomas <[email protected]> wrote: >>>>>>>>> face palm. more firefly references...ominous :/ >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> ________________________________ >>>>>>>>> From: Adrian Cochrane <[email protected]> >>>>>>>>> To: [email protected] >>>>>>>>> Sent: Sun, 29 May, 2011 9:58:12 >>>>>>>>> Subject: protocols >>>>>>>>> >>>>>>>>> avid Hearnden <[email protected]> Wed, May 25, 2011 at 8:36 AM >>>>>>>>>> Reply-To: [email protected], [email protected] >>>>>>>>>> To: wave-dev <[email protected]> >>>>>>>>>> There is a technical roadmap (i.e., rich design docs etc, published >>>>>>>>>> somewhere on the site - let me know if you can't find them) for a new >>>>>>>>>> protocol that overcomes many of the issues with the current one, and >>>>>>>>>> works >>>>>>>>>> much better with more advanced features (e.g. diff-on-open). I >>>>>>>>>> don't think >>>>>>>>>> it's a moving target - the doc has been ready for a few months, and >>>>>>>>>> I don't >>>>>>>>>> think anyone has significant changes to it in mind. However, AFAIK, >>>>>>>>>> nobody >>>>>>>>>> who's available has signed up to do the work, so there is no >>>>>>>>>> timeline for >>>>>>>>>> it. I was keen to get into it a few months back, and Alex North was >>>>>>>>>> too, >>>>>>>>>> but both our availabilities have significantly diminished. It's >>>>>>>>>> probably >>>>>>>>>> about 2-3 weeks of work for someone to see it through end to end >>>>>>>>>> though. It >>>>>>>>>> was previously blocked by a few things that have now been >>>>>>>>>> implemented. >>>>>>>>> >>>>>>>>>> I would strongly encourage not building too much on the current >>>>>>>>>> protocol, >>>>>>>>>> since it has a number of known limitations. The new protocol is >>>>>>>>>> simpler and >>>>>>>>>> achieves a better separation of functionality. If there are a few >>>>>>>>>> people >>>>>>>>>> (PyOfWave?) with the will and a bit of time, then it is very >>>>>>>>>> achievable to >>>>>>>>>> get it rolled out. >>>>>>>>> >>>>>>>>>> -Dave >>>>>>>>> >>>>>>>>> I will be proud to go over it, but (because I want to be liberal) I'd >>>>>>>>> first ask to start >>>>>>>>> with a forum or mailing list which I'd refer to as 'The Confederate' >>>>>>>>> after Firefly T.V. >>>>>>>>> series which gave Wave it's name. I've already exchanged some messages >>>>>>>>> with josephg on GitHub on >>>>>>>>> the shareJS Wave project on this. >>>>>>>>> >>>>>>>>> What I planned to work with, if I didn't get proper standardization, >>>>>>>>> is >>>>>>>>> the extended original >>>>>>>>> standards (to make up for some lacking features I say), a non-HTTP >>>>>>>>> alternative to Simple Data >>>>>>>>> Protocol, an fully designed Gadget API in a derivative of CoffeeScript >>>>>>>>> (to simplify offline clients), >>>>>>>>> and a URL scheme to serve for embedding, WaveThis, and a alias query >>>>>>>>> for >>>>>>>>> groups. >>>>>>>>> >>>>>>>>> I'll get started on it provided that I am provided with the necessary >>>>>>>>> information on how to do >>>>>>>>> it. However on my project, I've got some work on PyOfWave to finish. >>>>>>>>> -- >>>>>>>>> [email protected] >>>>>>>>> >>>>>>>>> -- >>>>>>>>> http://www.fastmail.fm - Email service worth paying for. Try it for >>>>>>>>> free >>>>>>>> >>>>>>> >>>>>>> -- >>>>>>> http://www.fastmail.fm - Does exactly what it says on the tin >>>>>>> >>>>>> >>>>>> -- >>>>>> http://www.fastmail.fm - One of many happy users: >>>>>> http://www.fastmail.fm/docs/quotes.html >>>>>> >>>>>> >>>> >>>> >> >> >
