Re: Wave state folders
is there a growing consensus toward Cassandra then? I guess the durability makes it a front runner. Does eventual consistency model fit, I know facebook are now using HBase From: Yuri Z vega...@gmail.com To: wave-dev@incubator.apache.org Sent: Mon, 9 May, 2011 17:49:20 Subject: Re: Wave state folders AFAIK no one is working on this at the moment. My understanding of the situation is that the infrastructure marking waves as belonging to a certain folder or marking wave is a tag is supported on the framework level. However there are some issues with current search implementation which should be changed in order to enable search queries based on user specific data ( like is:read). Also some changes to the user interface would be needed to add the new functionality. But, it still would be only a temporary solution, because we need a scalable and index based solution, something like Cassandra. 2011/5/9 Paul Thomas dt01pqt...@yahoo.com I think this was discussed before. folders are a bit old school. tags and specifically personal tags were suggested. From: Arne Eggert arne.egg...@googlemail.com To: wave-dev@incubator.apache.org Sent: Mon, 9 May, 2011 16:39:44 Subject: Wave state folders Hi, is somebody working on the feature, to organize waves in folders or somebody have an idea how to implement this. Thanks Arne
Re: Wave state folders
On Tue, May 10, 2011 at 1:55 PM, Paul Thomas dt01pqt...@yahoo.com wrote: is there a growing consensus toward Cassandra then? I guess the durability makes it a front runner. Does eventual consistency model fit, I know facebook are now using HBase I don't know Cassandra and whether and how it'd help here, but how about simply adding Lucene/Solr for indexation/search? (whichever the actual persistence backend) -- Thomas Broyer /tɔ.ma.bʁwa.je/
Re: Wave state folders
Yes, this is definitely an option IMHO. I mentioned Cassandra as it was suggested earlier in similar discussions, however, I don't actually propose use of any specific persistence solution. Lucene with file based persistence in fact may be the easiest solution for now. 2011/5/10 Thomas Broyer t.bro...@gmail.com On Tue, May 10, 2011 at 1:55 PM, Paul Thomas dt01pqt...@yahoo.com wrote: is there a growing consensus toward Cassandra then? I guess the durability makes it a front runner. Does eventual consistency model fit, I know facebook are now using HBase I don't know Cassandra and whether and how it'd help here, but how about simply adding Lucene/Solr for indexation/search? (whichever the actual persistence backend) -- Thomas Broyer /tɔ.ma.bʁwa.je/ http://xn--nna.ma.xn--bwa-xxb.je/
ShareJS: Wave's OT stack as a lightweight library
Dear wave refugees! As many of you know, I really want wave's technology to be usable in other situations. So I made ShareJS - a NodeJS server javascript client for doing concurrent editing with arbitrary data. Here's a simple concurrent wiki built on top of sharejs: http://sharejs.org:8000/wiki/Main Open it in a couple browser windows you can do google wave style live editing. Its wave's OT technology, rewritten in coffeescript. The software stack is type-agnostic. At the moment I'm working on writing OT code for arbitary JSON objects, so you should be able to whip up complex concurrent apps like route planners, spreadsheets, ... or, whatever :) Its pretty early days for sharejs - there's currently no undo support, no authentication and no rich text. You should help out. At ~2000 lines of code (+ tests), sharejs has 100 times less code than wave in a box. And its written in coffeescript (need I say more?) If you can think of anything particularly cool that you want to do with sharejs, drop me an email. I'll tell you which features are missing and how you can help write them. Lets make wave's legacy a whole bunch of awesome software. Code: https://github.com/josephg/ShareJS MOAR DEMOS: http://sharejs.org:8000/ Cheers Joseph
Re: ShareJS: Wave's OT stack as a lightweight library
Yeah - you should be able to do realtime collaborative editing in pretty much any application, so long as you can describe the OT semantics for your data. Plain text, rich text and JSON support should cover 95% of applications out there. I'm really excited to see what people make with it. Brett - You're right about wavelets. It should be possible to compile wave's OT code into javascript and pull that in to sharejs as another data type. You could run the wave editor on top of sharejs with a bit of work. Torben - is this compatible at all with the stuff you've been working on? You gave me the idea to do OT for JSON - though I think my implementation is quite different from yours. [On that note, I'd really appreciate comments about my spec for JSON OT: https://github.com/josephg/ShareJS/wiki/JSON-Operations ] Do you think I should implement federation in sharejs? A classic client-server model will work fine for most sharejs applications I can think of, and I'm going to focus on that first. I really like the idea of federation though - its a great way to authenticate users deliver notifications. -J On Wed, May 11, 2011 at 1:34 AM, Thomas Wrobel darkfl...@gmail.com wrote: This is pretty darn cool. For our own project we pretty much need all of wave (at least, concurrent federated multiple editors each user having their own subscriptions to various waves, different users posting seperate wavelets with in it). Still, I can see *so* many uses for this. You could, for example, use it to concurrently edit 3d data stored in a xml like fornat (x3d for example) - multiplayer 3dsmax anyone? :P -Thomas ~~ Reviews of anything, by anyone; www.rateoholic.co.uk Please try out my new site and give feedback :) On 10 May 2011 17:18, Joseph Gentle jose...@gmail.com wrote: Dear wave refugees! As many of you know, I really want wave's technology to be usable in other situations. So I made ShareJS - a NodeJS server javascript client for doing concurrent editing with arbitrary data. Here's a simple concurrent wiki built on top of sharejs: http://sharejs.org:8000/wiki/Main Open it in a couple browser windows you can do google wave style live editing. Its wave's OT technology, rewritten in coffeescript. The software stack is type-agnostic. At the moment I'm working on writing OT code for arbitary JSON objects, so you should be able to whip up complex concurrent apps like route planners, spreadsheets, ... or, whatever :) Its pretty early days for sharejs - there's currently no undo support, no authentication and no rich text. You should help out. At ~2000 lines of code (+ tests), sharejs has 100 times less code than wave in a box. And its written in coffeescript (need I say more?) If you can think of anything particularly cool that you want to do with sharejs, drop me an email. I'll tell you which features are missing and how you can help write them. Lets make wave's legacy a whole bunch of awesome software. Code: https://github.com/josephg/ShareJS MOAR DEMOS: http://sharejs.org:8000/ Cheers Joseph
