udev reorders assignment [was Re: Solved Debian update - keyboard responsive, Lennart Sorrenson not so much]

[D. Hugh Redelmeier]

| From: Russell Reiter <rreiter91-Re5JQEeQqe8AvxtiuMwx3w at public.gmane.org>

| <snip previous>

I strongly advocate selective snipping, but I do recommend including
attribution so that the reader can tell whose voice they are reading.


| > In the case of ethernet interfaces (not in /dev, thanks to BSD brain
| > damage 30 years ago), you can discriminate by way of MAC address, but
| > I don't think that that is satisfactory since a card's MAC addresses
| > can be changed.  I've not yet bumped hard into this problem so I
| 
| Could you elaborate a little more on BSD brain damage. Also, I assumed
| MAC address's were stored on prom's not eproms and so were permanently
| assigned by the manufacturer.

I guess I should have said Berkeley CSRG brain damage -- those were
the people that designed the socket interface, for BSD (Berkeley
Software Distribution).

The sensible way of addressing devices, including networking
interfaces, would have been through inodes named in /dev.  See how
Plan 9 does it -- eg. /net/ether0.  Why invent a new name resolution
mechanism?  And a whole new API to deal with things using those names.

The socket stuff is a lot of warts on the side of UNIX.  Very useful
warts, I admit.

Old time UNIX purists don't like BSD and many of the extensions it
added.  They don't have the elegance of 7th Edition UNIX or Plan 9.
But they did add functionality that is very important.  BSD was the
first UNIX with support for paging, for example.  (I've been using
UNIX heavily since 1975, so when I say old time, I mean it in that
scale.)

All of this is only of historic interest now.  BSD won a long long
time ago.  It is good enough.  I read Plan 9's marketplace failure as
demonstrating that doing things right isn't enough of a win for folks
to change.