Google Wave Backstage

As Google Wave is about to be released to 100,000 beta testers tomorrow, reader snitch writes in with a link to an in-depth interview with Dhanji Prasanna, whose title is Core Engineer. It covers some of the technologies, tools, and best practices used in building Wave. "InfoQ: Would you like to give us a short technical outline of what happens to a message (blip) from the moment a user types it in the web client, until becomes available to every one else that is participating in that wave — humans or robots? ... Dhanji: Sure, a message written in the client is transformed into a series of operations that are sent to the server in real time. After authenticating and finding the appropriate user session, the ops are routed to the hosted conversation. Here these ops are transformed and applied against other incoming op streams from other users. The hosted conversation then broadcasts the valid set of changes back to other users, and to any listening robots. This includes special robots like the ones that handle spell checking, and one that handles livesearch (seen in the center search-panel), as well as explicit robotic participants that people have developed. Robotic participants write their changes in response to a user's and these are similarly converted into ops, applied and re-broadcast."

Comments

Cautiously Optimistic

[rehtonAesoohC]

I have to say that I am excited about the prospects of a chat/im/document/wiki/social network collaboration system all rolled into one, but I am very skeptical if they will be able to pull it off the way they have been touting it.

For starters, most people are very well ingrained into their way of using the particular applications that accomplish the things Wave does (all independent of each other), so I think a massive component to the success of Wave will be how good the integration tools will be. Will we be able to import contacts from Exchange straight into Wave? Will we be able to use waves in email services other than wave? IE: Could a wave user interact with a wave with someone who is using MS Exchange the same way as they interact with someone who is using Wave also?

That said, I think Wave could seriously revolutionize the standard of email communication, and I really hope for all our sake they are able to pull it off.

Re:Cautiously Optimistic

[edmicman]

If anything, I see this being the closest thing to actually *subvert* Exchange usage in a corporate setting. Granted, all I know is what I've read and seen in the video, but the concept strikes a chord with me. For example:

At work, we use Exchange, and I suffer from information overload. We aren't taking advantage of the calendaring features really, other than to schedule reminders of when we have meetings. The VAST majority of my work processes involve email exchanges between multiple people, emailing copy of spreadsheets and screenshots to all of them, who in turn respond to everyone else with their own docs, etc. I may be working on any number of tasks or projects at a time, and each of those has their own threads, sets of documents, IM exchanges, everything. I try to organize them via folders, categories, posting docs to a share and telling everyone to go there to view them, but it's a mess. Granted, a lot of the problem may be lack of organization all around, but this seems to be the case no matter where I've been. We could try and copy everything to a wiki, or try and force Sharepoint to work for us, but it just doesn't work, at least right now.

From what I understand of Wave, instead we could have a dedicated wave to each task or project. Everybody communicates via that (replacing IM and email), posts documents there (essentially replacing file shares, emailing multiple copies back and forth to everyone...and didn't I see there was some sort of version control built-in?), and everything from start to finish is contained there. It sounds like a wiki, kind of, but in real time and organizing everything communication related that you'd normally use other apps for and have that data stuck elsewhere.

Sure, Exchange interaction should be there. But why keep using Exchange if Wave can manage your data and workflow for you? Maybe I'm off, maybe that's not how it works, and maybe I'll be disappointed. But it sounds really cool at this point :-)

Re:Cautiously Optimistic

[IGnatius T Foobar]

If anything, I see this being the closest thing to actually *subvert* Exchange usage in a corporate setting.

Screw that, I see this being something that could subvert Facebook. There's really very little difference between groupware and social media anyway -- it's just how it's optimized and featureized. So let's move back to a world where everyone is working -- or playing -- on the site or server of their choice, yet everyone is still connected together, instead of forcing everyone to join one single site.

Re:Cautiously Optimistic

[GooberToo]

One of the most important difference b/w how Exchange and Wave work is that the later is hosted by Google and hence controlled by it.

The protocol is open and Google has publicly stated anyone is free to host their own server. So we can safely toss that concern right out.

With Wave we would all have to trust google with our data, which a lot of companies in the world may not be willing to do.

Nothing could be farther from the truth. The truth is, Wave is built from the ground up to concurrently integrate, in real time, both open (Google or 3rd Party) and closed (company x's own ) Wave servers in a transparent manner. With wave, you have to trust no one other than your self. Period. In fact, that is so much so the literal truth, you are less dependent on another company for your own technological collaborative destiny; which is absolutely not true for Exchange.

Wave understands locality and security so its possible to security integrate public Wave services with private Wave servers and services without ever sharing data outside your own company.

In short, nothing you've offered as fact is even remotely true.

social networking, business collaboration...

[circletimessquare]

nah, none of these things

google wave is going to be the backbone of a thousand homebrew MMORPGs, probably nethack interface style at first, but i don't see why eventually it couldn't look like WoW

heh, thanks google, for giving us our own battle.net to play with in the style of an easy programming interface

You've got it right.

[copponex]

It's a real time protocol with built in journaling, that is both free and open. Think of it as HTML written after the knowledge that connections will be mostly persistent and fast. Waves are going to replace damn near everything displayed live on web pages. It's basically an open and extensible combination of wikis, sharepoints, calendaring, and web forums.

Google OS + Waves + commodity hardware. If anything, at least the next version of windows will be much less expensive.

Wave need a killer app.

[Tei]

It seems the Killer App of Google IO and Google Gears is Wave, but Wave lacks a killer app. Withouth that, It will not be popular.

Wave may need a killer app that needs a 90% of the features that provide, or only a 10%. Also, a killer app will cement some ideas about what Wave is. Another problem with Wave, is that is nothing just now, is nothing and everything, but need to be something, and that nameless something is yet to be invented. I suppose Google want exactly that, some guy inventing a killer app for Wave, or even some usefull toys. But I don't think have it yet. Is everyone listening? Google has created Gears, and Gear can add "offline" features to any webpage. Google IO can add streaming features to any app and more. We need to listen to Google more, because is releasing some technologies and ideas that are worth our time. The XMLHttpRequest was behind our radar a few years, before people realized his raw power. I suspect theres some untapped power on some of the latest tools released by Google, and is not Wave, is what move Wave.

Of course, I can be wrong. Who I am? another random guy on the internet :-/

Re:Wave need a killer app.

[Joe Random]

Wave lacks a killer app.

Wave is the killer app (the reference implementation, I mean). It is, at its core, a replacement for email, IM, and wikis. In fact, that diversity may be its biggest stumbling block. As your comment shows, people will want Wave to be "something". People understand email. People understand IM. People understand collaborative editing. But what do you call something that rolls all of those together? How do you create a niche for something that encompasses functionality from what are currently considered separate niches? It's like trying to explain to someone 50 years ago about how wonderful smart phones are. "What do you mean, text messages? If I want to send a letter I'll go to the post office. Calendar? I already have one on my desk!"

I think that this massive level of generic utility is going to slow adoption somewhat, and adoption past some threshold is exactly what Wave needs to break into mainstream usage.

Still looks like portable "Word w/ Track Changes"

[xxxJonBoyxxx]

Every time I look at Wave and its threaded conversations I think of Word documents when you track changes. (shudder) I think the most popular option on Wave will be a "ignore everyone's inane comments and just let me look at the original content" option.

It looks like a cluttered mess

[rbanzai]

Every time I try to take a closer look at Wave it just looks like a horribly cluttered mess. It's like they said "Why use ten different programs when we can replace them with one? How? By stuffing the data from ten different programs onto one screen! GENIUS!"

Are there any videos of this product that don't look like digital throwup? There has to be more to it than what I've been seeing, because what I've been seeing looks absolutely unusable.

PyGoWave

[simon13]

If you're itching to try out Google Wave like I am, a bunch of developers have already launched their own wave server implementation. A combination of Python + Django Framework + Javascript. You can create an account and have a play around, or you can download and run your own. Note that its still in early alpha state.

http://pygowave.net/