Re: [Gnu-arch-users] hurd followup

[Alfred M\. Szmidt]

   It's too off topic for prolonged discussion on this list, imo, but
   you could look at the Wikipedia articles on Mach, L4, and the GNU
   hurd, find RMS' retrospective comments that his original thought
   was that the hurd could be completed quickly, note the lack of
   device drivers, and ask what exactly people would really want that
   functionality for.

I've been a Hurd hacker since far to many years, and know the real
reasons why things are going as slow as they are; Wikipedia doesn't
have the real story.  One of the major reasons why things are going as
slow as they are is simply because there are no people working on the
Hurd, and because most people (today, this wasn't true many years ago
though) are simply happy with being able to run Emacs on their
GNU/Linux machine without doing much more.

It all boils down to "Worse is better", and why nobody today uses Lisp
or a Lisp operating system; even though that combo is the most
powerful way to make a operting system.  And also the quickest.

   If you want to make a fast, modular OS with such radically
   different capabilities there are prbly better approaches.  Build a
   lisp OS or something.

Sometimes it is easier to start from scratch, and ditch everything
else that has been written.

   I'd handwavingly suggest that in the early days the Hurd did have a
   serious chance at big impact.  If it had beat Linux to the punch
   people may indeed have swallowed the performance issues in exchange
   for a flexible, free kernel.  But it didn't, not even close, so we
   can only guess what might have happened.

I don't think the performance issue was ever really a serious problem.
It is now, if you compare things to GNU/Linux.