Hacker News Reader

Comment by frankfrankfrank | Hacker News Reader

It's a bit of a shallow perspective to say that Google shut down Wave. Sure, they shut down the branding and distinct feature that was called wave, but it was essentially just the experimentation environment for synchronous editing that was then deeply integrated into the rest of the GSuite. It didn't die, it just went through metamorphosis.

View on HN

Do you have a source for this? I don't think it played out that way. Google Docs and Sheets both had synchronous editing before Wave launched. At some point Docs was rewritten to use the underlying mechanism from Sheets. I'm not aware of any cross-pollination from Wave. The Docs and Sheets teams were fairly closely connected; Wave was developed in an entirely separate part of the organization, I believe the team was mostly in Australia, and there was never much communication AFAIK.

(Source: I was co-author of the original Google Docs implementation aka Writely; had some visibility into later iterations. I left Google in 2010 but have some idea of what went on subsequently.)

Does no one remember Etherpad? It had collaborative editing and Google snatched them up.

A year later they released Google Wave and Google Docs also got collaborative editing at some point.

Even then, Etherpad was not the first collaborative editor. It was just the first (to my knowledge) online one but I believe desktop ones existed for years.

And of course, the principle used behind collaborative editing (operational transform) is from at least 1989.

CRDT is the newer game in town now and it may replace OT entirely.

As a third party observer, I take it they both Google Wave and the collaborative editing in GSuite derive from Etherpad, and Etherpad itself is built upon the academic papers that predate it. If that is the case, while I’m sure some code may have migrated over, Google Wave “died” but Etherpad lives on.

Wave launched internally at Google to much fanfare at TGIF, and the whole company was somewhat amazed at the time by its realtime collaborative editing. Wave was built as a secret project in Australia, so when it did launch, internally there was understandably some rife / confusion as to what the strategy was.

The tech for appjet was primarily Javascript server-side, and I believe it used GWT for Java->Javascript client-side. This was seen as very odd to many inside Google, as javascript on the server was a new concept back then, and a far cry from the "officially supported" languages for development at Google.

While the initial launch of Wave went ~okay, the product itself had massive scalability issues. The Appjet/Etherpad team was then aqui-hired and quickly relocated to Australia. While one would think they were acquired for the similarities between Wave and Etherpad, they were instead tasked with "fixing" the javascript server-side performance situation. This made sense to management as appjet had been pioneering the concept since before Node was a thing.

In the end this wasn't a great outcome for the acquired team, but is typical of the aqui-hiring Google does.

Doc and sheets later added comments and realtime collaborative editing, but it was all reimplemented and never directly lifted from Wave's implementation.

s/tech for appjet/tech for wave/

From what I've heard: Google docs had multi-user editing, but it wasn't real time. Periodically the document would upload, be synchronized with other copies, and then update on the clients. Wave's real time cursor, synchronization, and conflict resolution algorithms were integrated into docs.

You're correct regarding the original Docs implementation. I had thought that the rewrite (launched around 2010 or possibly 2011) was primarily based on the architecture of Sheets. But it's possible there was some Wave heritage I'm unaware of.

Additionally, I believe the commenting and suggestion features were added to Docs after Wave was turned down; my recollection is that that bit is one of the main things that changed about Docs after Wave.

Google Wave was sold to us as a better replacement for e-mail. That was its main feature! It definitely died.

Sure they took some of the pieces and reused them in other products, but Wave definitely died.

10 years ago, I collaborated on Word docs by making changes and emailing it back. We’d look at revision histories to keep track.

Now the only email is the one that gives the link to the online document.

And I was the person who had to merge two peoples work if they both made changes on the same day with the 'master' copy.

Sometimes I wonder how we got out of that era alive.

But collaborative editing predates Google Wave. The innovation that Google Wave presented was using a unstructured shared document for all communication. That idea is pretty dead right now, and given that that is Wave’s main innovation, I’d consider Wave dead.

GSuite might have some code from Wave, but collaborative editing came from existing implementations.

It is well integrated in their office products, but Robots and Gadgets went away. Those were core to Google Wave, where you could search for games which you'd drop into a conversation and could play them right there inline with your chat partner(s).

I would agree here. Much of Wave was put into Google Docs, Drive, etc., as well as various other Google office collaboration products. Yes, Wave itself died, but it didn't fail.

Google does this with a lot of their stuff, it seems, too. Google Glass, for example, spread into so much of Android and further: Ok Google for the Assistant came from Ok Glass; the timeline made its way into Google Now and eventually into Assistant again; heck, you can see the beginnings of Material Design in the Glass UI. Google makes a ton of experimental products only to axe them swiftly and put their remains into new products.

Reply on news.ycombinator.com