How not to replace email: lessons from Google Wave
jamey.thesharps.us
jamey.thesharps.us
1. Email lets you type an entire message before sending; IRC at least lets you type a whole paragraph, which is enough. Wave didn't let you type more than one character before sending.
2. The designers correctly realized chronological order isn't enough, you also want logical order, but unfortunately they provided the latter instead of the former, whereas you actually need both. (Gmail provides both: new messages and threads. Git provides both: view changes, and view current code. Google Docs provides both. Either view alone does not suffice.)
I would by all means encourage someone to take another shot at building something like Wave, with those problems fixed.
If it was indeed added later, this is an example of something Joel Spolsky cautioned against: don't do a huge marketing blitz for version 1.0 of your product. Aim for user count to increase in step with product maturity.
Though admittedly this might be hard to avoid with a product that has strong network effects.
Oh god, please don't, at least not for chat/irc/whatever-you-call-it. I've seen so many attempts to try to somehow add another layer of order on top of chronological order. All the threads and whatever features and it is always a "what the heck is that, where are my messages, why did they get rearranged" disaster.
Like threads in e-mail, Usenet, forums on WWW?
IMO those are perfect examples of a combined logic and chronology. I'll speak for myself, but I absolutely prefer threads in the mentioned examples. It makes the content better readable. For me, there's no question about that. Although that isn't any proof; it'd be best to have usability studies about it.
If you're referring to conversation view, then they're not as good as traditional threads since they don't handle keeping track of conversation between multiple people that may branch off into different topics.
It never tried to hide messages from you, the threads worked more like filters. Instead of finding the discussion you wanted you filtered out messages that weren't related to that discussion.
This is the sort of thing I love about the web, and composability in general I suppose.
The people who designed the HN UI didn't think or care about that feature at all, but just by virtue of the fact that the user has powerful general-purpose tools at his disposal, and the HN UI is just as valid as content as the actual content is, the user can piece it together for himself.
This would not be possible if HN were a traditionally-designed desktop application instead of a web page.
The web returns us lots of the composability that was lost when "progressing" from CLI applications to GUI applications.
Here's one way to show new comments at HN: https://www.talkyard.io/-32/how-hacker-news-can-be-improved-...
— part 1: "Finding new comments". Video, + live demo if you click "Recent comments" in the upper right corner.
That's not very intuitive. It's easier to see new replies to comments one posts by going to the comments view on one's own user profile page, but seeing new replies in general is definitely not (especially if you don't check for several hours).
I think that traditional mail and news clients with their threaded view go it right for that purpose decades ago.
IIRC, HN does roughly the same thing that reddit does for deeply nested comments (meaning that you have to click on a link to view more of the thread). The most deeply nested thread I was involved in [1] on HN didn't appear to be handled any differently than any other thread I've seen here.
A properly implemented mail and news reader will indicate new messages in a threaded view after retrieving them. For example, Thunderbird viewing a usenet group makes it very easy to see new messages in threads after clicking on the "Get Messages" button.
To me the Thunderbird approach actually seems a bit annoying. What if I return after lunch, click "Get messages", and then Thunderbird highlights new comments in orange. But I don't read all of them, instead I go for another errand. Later, I return ... and click "Get messages" again. But now the orange marks from after-lunch will be lost, right? And only the very most recently loaded messages will get marked?
As far as I know, the thread overview pane will still show the unread message in a bold font, but the indicator that you just retrieved the message would be lost for the messages that you retrieved right after lunch (though it would then show up for messages you retrieved after you returned from your errand).
In either case, it would still be obvious which messages you haven't read as of yet which is something that HN doesn't really show on subsequent viewings of a given comment thread.
Perhaps Thunderbird could be extended to show a different symbol associated with each message retrieval so that, in your example, you would have one symbol for messages retrieved after lunch and a second symbol for messages retrieved after you return from your errand.
Whereas here at HN, it's a lot harder for the software to know which messages one has read (because here messages are shown without the user indicating what s/he is reading).
Yes perhaps different symbols — actually I'd be happy with just Thunderbird's bold font :- )
But functionality wise they were pretty good.
Not sure what “back in the day” is for you, but I remember e.g. comp.lang.lisp being terrible in the 90s. And we all made jokes about flamewars, because flamewars were reality.
Even the dumbest of them would remember what articles you read and let you pick up where you left when you came back the next day. That's a straightforward feature that is sorely lacking on most web based discussion sites (Twitter does it, but is lacking in threading).
Here's a demo video of how HN and Reddit could be modified to solve that, i.e. how chronological order could be added to HN:
https://www.talkyard.io/-32/how-hacker-news-can-be-improved-...
This part: "1. Finding new comments". I've actually implemented this for real — click "Recent comments" in the upper right corner
Some apps also do it, the app I use (Sync) will tell me the number of new comment and highlight them when I revisit.
As far as "collaborative editing environment(s)" go, I would love to see more alternatives with such capabilities.
Sounds like Nirvana to me. No more eternal “reply to all” chains that span months? I’m in.
What I found is that it turned IM-like communication into a high-bandwidth medium comparable to what you get in a phone conversation. It managed that while also retaining nearly all of the advantages of an IM client.
What surprised me was Wave is obviously best for business, where email chains CC'd to a load of people who don't care are the norm, yet they killed it only months after enabling it for Google Apps
We spend a lot of time just figuring out what we wanted to communicate about inside wave and settled on planning a trip abroad, and eventually migrated to docs.
Ever since I've been paying a lot more attention to the somewhat indefinable 'feel' of a tool. Trello, for example, felt right despite its relative simplicity (or because of it?), so I felt like it had a bigger chance of working out.
If I had to define it, I guess it would be something in the direction of what someone in an early React presentation described as the 'pit of success'. Using the tool feels like it was optimized for how 'you' use it, and I suppose if everyone feels that way, it's got a chance of succeeding.
tl;dr: 'apart from the clunky UX' is probably a bad sign.
I mostly don't do software development except for the occasional learning project.
I was just listening to a talk by Eric Schmidt where he used Wave as an example of a mistake he had made -- for not closing it down soon enough despite it losing users right from the beginning.
I'm not really sure what Schmidt "does". Is he limited to running business processes and exploiting markets profitably or has he proven a talent fostering the invention of ground-breaking tech? If I want to successfully build the next Wave, I'm not sure if I want to be listening to him or not.
Transcript: "I made many, many mistakes as the CEO and I'm glad that the good things we did overcame the mistakes. A typical example is... Did anyone ever use a product called "Wave"? Some of you may have. We launched it and from day one its usage went straight down. And the question is: If you're the CEO, how many months of going down before you cancel it? ... Look at the numbers, three months, six months, nine months? Eighteen months. Not good. There are other examples of that category, and again, just bad decisions."
Schmidt's decision has been borne out by events. Did Wave take off when open sourced? No. Did anything Wave-like developed subsequently turn into a big market? Also no.
> Did Wave take off when open sourced?
Eh, how many "dump and run" open-source projects take off? Bringing an opensource community to life is its own brutally hard problem.
I think Google's real error was not killing it too soon, but allowing it to be developed in a bubble and then springing it on the world with a giant, "Ta da!" It let them ignore a lot of real-world problems and questions, investing heavily in untested design hypotheses. Once you get to the point of a huge global launch, you've baked in problems in a way that is very hard to fix.
Slack.
Trello.
Apple usually mail the core experience with v1, but all the ancillary roles have to wait till v3 or v4.
Schmidt was CEO from 2001 to 2011. During that time period, Google basically invented all the infrastructure we take for granted in distributed systems today, popularized the concepts of JavaScript-heavy web applications and Software as a Service, created from scratch a new browser (Chrome) and a new operating system (Android) which promptly dominated the market, and became one of the top five largest companies in the world.
I think he deserves some credit, yeah.
- Chrome - competitive play - iterative improvements over competitors
- Android - competitive play - purchased, playing catch-up to competitors
- Infrastructure - internal services
Basically, is he is more Xerox Copier division than Xerox Parc?
OTOH, when Larry took over, things became much more top-down, with Larry trying to be Steve Jobs and tell the company what they should be building next -- most notably when it came to Google+.
I left before Sundar took over so I don't know how he runs things, but it seems like Google these days is just sort of chugging along, building solid products but not doing anything revolutionary.
The Larry approach sounds more vision-led and its interesting how it looks when it fails. Its difficult to see beyond the survivorship bias and only see it as egotistical and ignorant failure.
The real mistake with Wave was that a group of incredibly smart engineers were basically allowed to go into a bubble and build whatever they wanted for an extended period of time. Successful innovation needs to be anchored by a real problem to solve. E.g. the best programming frameworks always come from someone who is building a real app on top, not from someone who set out to build a programming framework. The feedback loop is necessary to keep developers focused on the actual needs of users, rather than on whatever problem they find most interesting to solve.
I think Eric is referencing the poor usage numbers as a proxy for the fact that the Wave team was not building something that users really wanted.
(FWIW my own startup, Sandstorm.io, had a pretty poor record on this point as well.)
I don't think it's fair to say: "They only bought it so it doesn't count." You could just as easily say: "They only hired Paul Buchheit, he was the one who came up with GMail, not Google." Real inventions aren't spontaneous point events, they take many years of development, and can die anywhere along the way if you don't continuously create the necessary environment to foster them.
Similarly, Google plus was not a total failure - Google Photos came out of it.
Making it even more ridiculous that they killed off Google Reader for it (or didn't bother to revive it afterwards)
EDIT: Also, they already had Picasa Web Albums since 2006. I don't know what exactly Google Photos has to offer over it in terms of functionality.
The machine learning assisted search. For example, if I want to find photos of whiteboards that I took to record a meeting, I type in "whiteboard".
That said, surveying today's landscape, it would appear that there isn't actually a huge amount of demand for either full resolution user-curated sites focused on photography or RSS readers. The vast bulk of users are fine with throwing their baby and cat pictures up on more casual sites like Facebook and Instagram or getting their news from Facebook/Reddit/etc.
Problem wasn’t full resolution, it was re-compressing your originals. Photographers that care about full res, care about original quality.
While there were (and still are) options like SmugMug for photogs, Flickr was the only prosumer one supporting unlimited original quality.
Also on the “stagnating” comment — at some point is it possible you’ve hit a reasonable photo site, and shouldn’t tinker much?
Turns out photographers who posted and curated 50,000 photos are not big fans of having the rug pulled out from under them by people in marketing trying to change the UI for their own engagement metrics.
I actually think that's somewhat fair. I could come up with some things that I wish were better with Flickr and there are probably lots of things I'm not thinking of. But I still use it and, in conjunction with Lightroom, I find it's a pretty good way to display a curated collection of edited photographs (as well as serving as a backup of last resort).
I did object to the big UI change a few years ago and still don't really love it. On the other hand, I don't actually use the web interface much. I upload from Lightroom and I mostly view from a tablet.
Source: I am one of the original authors of Writely, did not work on Kix but I was still at Google when it was being written, had a minor role in reviewing the design.
I don't know whether Trix originally used OT specifically, or the Kix team borrowed that idea from Wave. But the basic idea of clients performing small, tidy logical operations and then synchronizing them in realtime through a server was part of both Trix and Kix. Writely used a very different (and more primitive) approach.
I don't know anything about the origin of the comments feature, but my guess is that there was no connection to Wave, except (again) possibly a bit of inspiration. The Wave team was very separate, both geographically and organizationally, and the tech stack they built was very disconnected from anything that was going on with Trix, Kix, or Writely.
I was asking because I have a vague recollection that Kix borrowed OT ideas or even code from Wave. But then I was even more removed from both teams than you. Or maybe it was just an aspirational thing from management to find a silver lining in the Wave shutdown.
Kix is a drastically different architecture than Writely, so many lessons would not have been applicable. Writely was built on a contentEditable DIV, so the data model was literally whatever junky HTML the browser's contentEditable implementation would create, and synchronization was implemented by performing three-way diffs on the HTML code. It was very much a dancing bear, in the sense of "the amazing thing is not how well it dances, but that it dances at all". ...but I digress. My point is that the mechanisms and challenges of that approach are entirely different than the Kix approach. I think the important lessons would have been product-level rather than code-level.
Keep in mind that Wave backbone was XMPP, the same XMPP that powered Google Talk, and could in theory allow any number of Wave instances around the world to talk to each other.
For me, though, it failed for two reasons:
* Wave was a solution in search of a problem, a need. It was fantastic as a tech demo --back when it was released, it was magical-- but it never seemed to fill a niche? It wasn't fully-fledged-enough to be a word processor, it wasn't really for communication, etc. I liked the description of other posters using it for gradually developing ideas --almost like a virtual pinboard-- but my guess is that very few teams work like this, and especially few outside of the bleeding edge tech space.
* Change is hard without a compelling reason. For example, I work in a (non-IT) organisation which offers people a choice of (legacy) MS Office + Sharepoint, vs. (the upstart) Google Docs. To me, Google Docs is far better for all a fraction of our work - it's faster, smoother, easier to share, and has (thanks to Wave!) collaborative editing... yet the proportion of users with the mindset to adopt it voluntarily I'd put at <5%, and probably even lower. And Wave offered a far less rounded, less compelling, less approachable product than Docs.
I have to suspect that nobody on the team had ever read Crossing the Chasm; if they did, they would have know that technophiles are a very small percentage of the market. If we want broad adoption, we a) have to solve an important problem for early adopters, and then b) broaden features and approachability enough that the majority of users will switch.
As the article points out, this is worse for network-effect products. Maybe interoperability would have helped. (And maybe not; Slack didn't bother with much of that.) But I don't think it was the primary mistake they made.
I'm not sure how to get that level of communication when the team is spread across multiple locations... webcams maybe
I love this phrase! We did something similar way back in the day, only using Emacs and "M-x make-frame-on-display" instead. Just like in your case, it worked fine as long as everyone was in the same room. The one tricky part of "multi-player Emacs" was that only one person at a time could use the minibuffer.
Google Docs, on the other hand, pretty much everyone I work with has adopted it even though it was more or less voluntary. The application itself works well but the ability to have everyone comment and make suggestions in a single doc is a huge win. Admittedly, we didn't and don't use Sharepoint so the alternative was pretty much to mail documents around.
> * Assuming you could stand Wave’s interface, it’d still be useless unless the people you wanted to communicate and collaborate with were also using it. Wave was intended to completely replace existing systems like email and chat, so it had no provisions for interoperating with those systems. To succeed, Wave required a revolution, a critical mass of people switching to the new way and dragging the rest of the world with them—and those haven’t worked out very often.
> * Making the revolution even more unlikely, initially Google offered Wave accounts by invitation only, so the people you wanted to talk with probably couldn’t even get an account.
I was excited about Google Wave at the time but first had to wait what felt like ages for an invite and then barely anyone I knew was on it and I quickly ran out of invites myself (because not everyone you send an invite will be as excited to use it as you are).
To replace e-mail it would have had to at least provide some kind of gateway between wave and e-mail to ease the transition -- or it would have had to integrate more closely with other tools and applications to "trick" people into adopting it (like how Google Talk integrated with GMail).
The "openness" was neat but didn't help them gain any traction because there was no incentive to actually build and run your own server because there were no users.
My social group has begun to use it it for a pretty wide variety of things, ranging from rough notes to blog posts to draft documents.
Also, can you structure documents into find hierarchy or tagging system?
Finally, no self hosting.
- It probably requires a left sidebar with a tree of the project’s documents,
- And perhaps a notion of document hierarchy, child pages or something to better browse. Aside from those, which aren’t technology challenging, it would make my startup switch to GApps.
Not sure what "feel connected" means. In both systems, you click a link and it takes you to the document in question. The only difference is that wiki links are shorter...
> Also, can you structure documents into find hierarchy or tagging system?
The hierarchy is just the file system, which only allows files to be in a single directory at a time.
> Finally, no self hosting.
OwnCloud with certain extensions probably does everything you want, e.g. https://owncloud.com/collabora/collaborative-editing/
https://marketplace.atlassian.com/plugins/confluence.chat/se...
With the caveat that this was of course several years ago, I seem to recall the reference implementation missing several major features. Private replies were notably absent.
It was all properly specced out, but the only available server software was incomplete - at least, it was when I tried running it.
The non-native scrollbars were smart and nifty, but, at the end of the day, they were jarring to use. I don't remember many single UI features that ever made me alternate so frequently between "neat!" and "WTH?!?". And I've seen lots of interfaces on many platforms.
http://ignorethecode.net/blog/2009/11/15/google_waves_scroll...
If it's true they were custom to make them work well "on mobile devices or netbooks with a limited mousing area", that's even worse. They know I'm on a desktop computer with a big screen, and they're delivering a netbook-optimized user experience.
Edit: Could any of the people downvoting me please state their position? I didn't think my comment was disrespectful or off topic, so I'm wondering what their problem with it is?
So apart from these obvious problems I think their biggest mistake was the try to replace email instead of evolving email to wave.
Mail AMP is a Google controlled superset of email. It is a given by many that chat is the ultimate social glue, and G already has their own chat protocol. In the interest of keeping people engaged, and keeping them walled in, I kind figured the long play of Google mail is to slowly morph all email into (a partially email compatible) Wave.
I love the idea described in the article of real-time collaborated _things_, as opposed to a timeline of ideas from a team. Specially for things like Github Issues or other bug management. The goal would be a shared living document, not a list of discussions. Really cool!
A lot of Matrix.org is inspired by what Wave could have been - although we are running late with threading (but some work is happening currently there). From our perspective, one of the biggest gaps was the lack of bridging in Wave to other existing comms interfaces (eg email) - as well as the quirky UX.
After all it is a store-and-forward transmission system with arbitrary length text and inline attachment of binary data.
all this makes individual emails self contained.
Various proposed replacements are more focused on the "conversation" it seems.
Back in the day, when I realized there was no backward-interop with email, I was afraid it wasn't going to go very far... network effect is hard to tame.
I am one of the four people who really liked Google Wave. I wrote a plug-in for it, and when Wave was shut down I stood up an instance of Apache Wave for friends and family. Basically no one wanted to use it except to play with it for a few minutes.
I want to check out the author’s C to Rust translator that he mentioned at the end of the article.
Here's something I wrote about it back then. https://news.ycombinator.com/item?id=7533023
Sadly, since Google killed it, I've never been able to get other teams to pick up the open sourced versions and now everybody just wants to use <insert latest startup collab tool>.
It definitely had a tremendous number of problems in execution, but I never really got why Google killed it so fast. It seems that a reboot and rethink of some core metaphors would have tightened it up into the new paradigm it needed to become.
There seems to be some historical revisionism here. Perhaps you could have said that about Gmail but no, when Wave came out people were very much used to crappy web apps. Wave suffered because it tried to make browsers do too much but that's been a common pathology for the entire history of the web, it was nothing to do with users not being "used to it".
If they'd written Wave as a desktop app it would have had more of a chance. I remember using it and the performance was awful, but not for any fundamental reason. Browsers just suck as app platforms.
Google Wave? No, Google Edsel, Mk1.
(Shameless plug: that's why I built CTX - https://getctx.io - a SaaS cloud search tool that indexes your data in things like email, Slack, Trello, GitHub, JIRA and Google Drive)
Make me wonder... What happen to projects that apache incubator retire like this one?
How about banks? I live in Poland and used to exchange Euros into Zloty at 3.97 with a 5% commission through my bank. Now I use Transferwise, and they give me a rate of 4.17 with a neglegible fee (a few cents). Now, that's an innovation that I welcome!
I don't have a problem trying to reinvent email. I do have an issue if they use market power to damage an existing technology that works well and openly across platforms.
Email has plenty of issues too, you can't just take any new features the smartphones have and call it a default of the old phones.
Email is "perfect" as long as nothing better replaces it, then it will no longer be perfect?
A few issues of email I can think of : no guarantee your email will arrive, and no real acknowledgement it did. Also, Mailing lists are clunky for users (I still receive regularly emails with 500 recipients, where a mailing list would be far more appropriate). I also think of attachments, which are very limited in size with a lot of providers.
So email is far from perfect, it's just that nothing better has arrived yet.
People decided that they liked portable computers far more after a brave company introduced a radical presentation of one.
Because nothing works perfectly.