2011-07-22

Google+ – observations and future

A previous post gave an in-depth look at what Google+ is and how it was created. This post lists my observations after having used Google+ for two weeks. It relates those observations to what a technical lead on the Google+ team wrote about its future.

Observations

Observations about Google+:
  • Messages versus documents: Google+’s model is an interesting mix of two metaphors.
    • Sending messages: You “send” posts like email, to a list of receivers.
    • Accessing documents: If someone moves you to a different circle, existing content can disappear and new content can become accessible.
  • Competing with Google Reader: Google is planning to export streams as RSS or Atom feeds (see 2.4). Importing external feeds would be handy, too. But then Google+ is becoming a feed reading tool and will compete with Google Reader. In contrast to Google Reader, Google+ allows one to both produce content and to consume it.
  • Extended circles: That’s a weird feature that screams “spam” and invades privacy. Why would Google+ allow one to post to friends of friends? The idea of making use of one’s extended circles is intriguing, but more thought has to go into this (opt-in? groups?).
  • Friendship relations – not really asymmetric: While you don’t have to circle someone who circles you, the common practice is to put everyone who circle’s you into some circle; be it the one called “people I can’t stand”. Google+ gently nudges you into this direction by notifying you of people who share something with you and are not in your circles.
  • Reputation? While circles are not as asymmetric as one might assume, they do add more information to a connection between two people that Facebook’s “friend” relationship. I wonder if that information could be used to compute some kind of reputation [2] for a user – a numeric value indicating how much their output is valued by others. On Twitter, followers divided by following is a rough indication of one’s reputation. On Google+, Google would have to compute a number for you to make this work and to protect privacy.
  • What’s in it for Google? How do they expect Google+ to make money, long-term? One possibility is that advertisers become the true customers of Google+ and Google+ users become the product that is sold.
Features I would like to see:
  • Combining circles to construct new circles would be very useful (see 2.5, below).
  • What is the nature of “Public” for sharing? Shouldn’t it be a circle? Such a circle would nicely complement the ability to build new circles out of existing ones – e.g., to create a circle "everyone except my co-workers".
  • Tags: Currently, stream content can only be filtered by the sender (via circles). Tagged posts would make it easier for receivers to filter (see 2.2).
  • Shared blogging and forums: would be great for all kinds of social activities (companies, clubs, etc.). See 3.3.
  • Planning events: also missing. See 3.4.

Future

Google’s Joseph Smarr participated in the Ask Me Anything (AMA) “I'm a technical lead on the Google+ team. Ask me anything” [via GigaOM]. The following is a summary of his answers. Quotations are put in quotes. Smarr cautions that he can only hint at upcoming features: “Since I work for a big/public company (albeit a pretty cool one), I can't provide specific stats, dates for future features, or details of confidential code/algorithms. But I will do my best to be ‘refreshingly frank’ about everything else. :)”

1. Management and engineering.

  1. Ultimate vision for Google+: “We think we can make sharing better and make Google more social (and useful) in the process. ”
  2. Technology used for Google+:
    • Server:
      • Languages and libraries: Mainly Java Servlets. “Lots of Java, no JSP, no GWT, lots of Guice, no Go.”
      • Pre-constructing pages on the server: The Java version of Closure templates allows them to use the same templates as on the client and to pre-render a page. After JavaScript has loaded, it “finds the right DOM nodes and hooks up event handlers, etc. to make it responsive”. Naturally, later GUI updates will be performed on the client.
    • Backend: built on BigTable and Colossus/GFS.
    • Client: Closure library. The HTML5 history API leads to neat-looking URLs (as in with “slashes”). Older browsers fall back to hash fragments.
  3. Technical challenge: what and how to update. “For instance, adding/removing someone from a circle impacts (among other things): which posts they can see, the counts of people on your (and their) profile, suggestions (for potentially many people), and so on.” Some of it has to happen fast, some of it can wait. Getting the trade-offs right is tricky.
  4. How was engineering different from other Google projects? Everything was very agile – frequent release cycles, etc. Note that management was even more different from other projects, e.g. it was much less democratic [1].
2. Improving streams and circles
  1. Searching streams: actively discussed internally.
  2. Tags/hashtags: Smarr would personally like to see them, but can’t say anything official. Twitter uses hashtags to great effect. The usage overhead is small, because they are simply typed into the content by prefixing a word with a has symbol.
  3. Better blocking: public posts should be hidden from blocked stalkers. “We're working to make block even more comprehensive; this is definitely a priority for us.”
  4. RSS/Atom feeds of streams? “To be honest, I'm surprised we don't have them (we did for buzz)...lemme look into that! :)”.
  5. Circles: Several questions mentioned the wish for more powerful management of circles.
    • Sharing while excluding one circle is difficult. Is being considered by Google, but shouldn’t make normal sharing more complicated. Which is a great criterion for any addition in this area.
    • Combining circles for sharing: Smarr doesn’t think normal users need it, suggests implementing it for power users via the upcoming Google+ API.
    • Combining circles for filtering streams: is a frequent wish.
    It seems to me that all of the above could be solved by providing a few operations from creating new circles from existing ones: set-theoretical minus, union, intersection, etc. I don’t see how that would be terribly complicated. But it would be extremely useful as a power feature.
3. Other new features.
  1. Social bookmarking: Right now a user’s +1 pages are listed relatively obscurely in the profile. Google doesn’t want it to be part of one’s stream, because that might lead to users +1-ing less, for fear of flooding their streams. But Google is aware of the need for improvement. Good social bookmarking would be a valuable addition to Google+ and with +1, the foundations are already in place. Google would have to figure out how such a service relates to URLs that are shared with others. If they really plan to use Google+ to improve search then bookmarks should be more useful than almost all of the other social content that people produce.
  2. More features for Hangouts: For example, whiteboarding and playing games would be nice while hanging out. Smarr writes that new features are coming, but he can’t say what they are. Google wave had the technology for this kind of thing, so maybe they’ll reuse some of it (probably protocols only).
  3. Google+ accounts for companies, clubs, etc.: They are working on it. Obviously, profiles would have to be different and there would have to be shared ownership.
  4. Planning events: they are thinking about it.

Related reading

  1. Google+: An interesting product, badly marketed
  2. Next in social networking: reputation?

No comments: