Google’s Cloud Source Repository … WTF :)

So Google has released a beta for  “Cloud Source Repositories” as part of their Cloud Platform.    At first blush it looks analogous to github, bitkeeper, sourceforge, etc. etc.   My reaction to this amuses me.

  • Does Google’s left hand know what it’s right hand is doing?
  • Is this more proof of how a modern high tech company balances trust v.s. velocity?
  • Is this a sign that all the cloud OS vendors will discover that source control is just a feature?


It’s painfully ironic that this new thing comes out only shortly after Google murdered Google Code.  It’s really hard to trust a vendor that kills product offerings (i.e. Google Reader, Google Code, etc.).  I suspect most developers are bewildered that they didn’t “just” migrate the Google code users to this new thing.   It really suggests that the team that shut down Google code was entirely independent from the one that build this thing.

There is a B-School narrative about how firms need to balance providing backward compatibility against the need to force their installed base to adopt new technology.  The first is necessary because customers need to trust that their investment in your offerings won’t suddenly depreciate as you create new versions.   The second is because old tech can quickly become a legacy tarpit that enables upstart competitors to leave you in the dust.

Microsoft was good at the first and very slow about the second.  And this dialectic is the primary cause of angst at Microsoft – and has been for about a decade plus.  When you are a monopoly you can spend a very long time resolving these tensions.    The unbelievably long road they took to get to an OS with a reasonable GUI is an example of that, though most people have forgotten that era.

So the B-School narrative around this suggests that a monopoly, i.e. a firm how’s customers are locked in (aka very loyal) can be extremely slow to offer the cool new thing.  That story is changing though.  If the velocity of high tech change increases firms are forced to be more aggressive about chasing it’s tail lights.

Google is willing  to abandon customers more casually than I’d expect.    Well until I recall the last paragraph.  Maybe.

In Google’s case I tend to think it’s more about the culture.  They cough up these products like Google Reader and Google Plus and Google Code and then the team and management loses interest or switch jobs.  The shiny new thing catches their eye.  Maybe they reward new to excess.

Anyhow what ever.

My third take on this is more interesting to me.  I remain fascinated by this new species of operating systems.

The set of things that make up an OS is much larger than they teach you in school.  It’s not just device drivers, and resource management.  These days it’s package management, app stores, developer tools and networks, evangelism, partner management, etc. etc.

Back in the 80s VC would sometimes utter the damning phrase “Oh that’s just a feature.”  For example somebody would author an amazingly complex package that would check grammar in documents.  “Just a feature.”   What that meant was that some other product Microsoft Word for example would swallow that feature.  And sure enough that is exactly what would happen.  Painfully – because each and every time the feature implementation inside the product that swallowed it would suck compared to the stand alone product’s.

So one thing I’m thinking is that cloud based source control (bug tracking, wiki, build tools, etc. etc.) is just a feature.  That it will be swallowed by the cloud platforms.  This morning that seems pretty damn inevitable.

The term platform has always been intended to suggest a stable surface you can build on.  But really these folks come from California.  Earthquakes!

1 thought on “Google’s Cloud Source Repository … WTF :)

  1. Kimbo Mundy

    I’d add deep learning AI to the list of features in this new OS. Those algorithms are qualitatively different when they have large datasets to learn from.

    Once we let deep learning loose on the meta-problem of what to spend money/resources on, then we’ll have reached the singularity. I wonder if those data-driven charities will be the ones to program that layer. I’d sure prefer that to Wall St. Consider:

Leave a Reply

Your email address will not be published. Required fields are marked *