4 ms·

Google Chrome OS is 7 years too late

• blasdel 17y ago
  Yet another otherwise-insightful essay on ChromeOS that completely ignores the existence of Google Native Client -- rich native applications that are integral to the web. ChromeOS doesn't ship without it.
  • ZeroGravitas 17y ago
    I've not looked into NaCl much but maybe someone can answer me this: running "real" code in a browser but sandboxed for security? Isn't that Java? This seems particularly key when they're talking about running Chrome OS on ARM.
    • blasdel 17y ago
      Code compiled for NaCL is in a word-aligned subset of machine code that can be quickly verified before execution and can only generate interrupts to the NaCL host (no syscalls). It's currently i386, but x86_64 and several ARM variants are being worked on, hopefully with fat binaries. The only security hole is timing attacks, because there's no latency hits or blown caches from having a runtime or GC or even any preconceived notions about what your language looks like.
• ZeroGravitas 17y ago
  I'd suggest the author's (mis)conceptions about the limits of browser based apps is 7 years out of date. It's not even an explicit argument, it's just an unspoken assumption that web apps aren't the real thing. Actually, maybe it is explicitly stated: "But the browser, being a viewer of HTML5+CSS+JavaScript is not capable of ‘blowing you away', so we have to move beyond it." But, every negative point he raises about web apps (which aren't many or particularly fundamental) is being worked on feverishly by Google (and Apple, and others). But even if you take his argument seriously, he's only saying that rich-client will be thin, platform-specific wrappers round the APIs of what are fundamentally cross-platform web apps. Hardly a resounding victory.