← Back to context

Comment by ismaildonmez

4 years ago

Microsoft Research has a paper about the very same issue (2019): https://www.microsoft.com/en-us/research/publication/a-fork-...

6 comments

ismaildonmez

Reply
georgia_peach  4 years ago

That paper smacks of a Chesterton Fence. They haven't come up with a tested replacement for many of the use cases, i.e.:

  These designs are not yet general enough to cover all the use-cases outlined above, but perhaps can serve as a starting point...

yet bullet #1 in the next paragraph is

  Deprecate Fork

I think this is a case of security guys being upset about fork gumming-up their experiments. I don't really care about their experiments. The security regime for the past 20 years may have bought us a little more security against eastern bloc hackers, but it hasn't done squat to protect us from Apple, Google, & Microsoft! I have never had a virus de-rail my computing life as much as the automatic Windows 10 upgrade. Robert Morris got 400 hours community service for a relatively benign worm. If that's the penalty scale, Redmond should get actual time in the slammer for Cortana, forced Windows Update, and adding telemetry to Calculator.

  • cryptonector  4 years ago

    You fail to address any of the substance of their paper, or of my gist (TFA), then go on a rant about unrelated things. The authors of that paper deserve better treatment even if you hate Microsoft.

    • georgia_peach  4 years ago

      I did. Chesterton Fence. fork() has been in Unix from the beginning. Taking it out at this point will cause more problems than it solves. Until you have a working Unix distro (kernel AND common userland services) that elegantly covers all of the forkless cases, your paper and their paper are just opinions. Theirs is a formally written one. Yours is a clickbaity one. And casting vfork() as any kind of improvement here is just bonkers.

      And the rant is totally related: i.e. devs breaking things that worked just fine to begin with for the sake of doctrinal purity. It is usually a false doctrine.

      1 reply →