Tuesday, December 9, 2008

Google's Native Client - Technology In Search of a Purpose

Google's Native Client seems like a step in the wrong direction for me.  It's an interesting technical challenge to securely run native code from unknown sources, and I could imagine that it's a challenge worth pursuing to some degree.


However, in the long run, it seems like we've made enough progress with virtual machines to stick with that approach.  Running x86 code in the browser just reduces the number of platforms on which you can run your code.  While X86 is certainly the dominant platform, I'm inclined to think we're better off staying at a level of abstraction from the operating system and the hardware when writing code to be distributed over the internet.

That said, i'm willing to let the market sort it out; this seems like interesting technology, and if it ends up being heavily used, then that implies that it had advantages worth pursuing.

2 comments:

Gavin said...

I think there are definitely some use cases where this functionality could very useful. I just started working on a side project where I will need access to an attached hardware device from within the browser. In my investigation sof far, I haven't found any way I could do it (aside from ActiveX and I'm not going there). Native Client (and Alchemy) look like they may provide a solution.

Anonymous said...

Its antithetical, been tried before and will fail