Re: Suggested dual human/binary interface for proc/devfs

From: Dale Amon (amon@vnl.com)
Date: Tue Apr 11 2000 - 14:20:45 EST


Olaf Titz <olaf@bigred.inka.de> writes:

>> processor 0 {
>> vendor_id:GenuineIntel
>> cpu family:6
>> model:6
>> ...
>> }

Also rather reminiscent of gated.conf although perhaps
a lot simpler :-)

There really is much to be said for the name:value
duple format. It's not just for someone new to a particular
conf file either. How often do you end up reffing the
man page just to be sure your wetware hasn't had a parity
error?

It also makes some sorts of errors stand out like sore thumbs.

And I am another of those who have suffered under vendor attempts
to make conf data binary files. I loved NeXT, but I simply
*hated* NetInfo... or at least I did after the first time the
binary db corrupted and I had to rebuild the entire CS
NeXTlab system user accounts, nfs imports and exports, and
you name it from scratch... starting from single user login
on the server.

Binary conf data is a very bad idea. It has been tried and
tried again by vendor after vendor and it is still a very,
very bad idea.

------------------------------------------------------
Use Linux: A computer Dale Amon, CEO/MD
is a terrible thing Village Networking Ltd
to waste. Belfast, Northern Ireland
------------------------------------------------------

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/



This archive was generated by hypermail 2b29 : Sat Apr 15 2000 - 21:00:16 EST