Hi,
I have read the whole discussion for some time, the whole thread
and now I think its time to sum things up a little bit.
You say that translators:
1) Cannot be in path
2) Are executed by user's action
3) Are not libraries
4) Are executed through hurd kernel mechanisms
And then conclude that this can justify a new top level directory
which is even VERY hurd specific.
In my opinion the argument that these translators are supposed
to go to /lib/modules has its weight because for all practical
purposes these things are the same.
1) They are both kernel extensions
2) They both can be executed upon user's action
(Just one example modules for zip drive insmod automatically
upon user mounting of zip drives listed in /etc/fstab. Worked
this way on my system.)
3) They are not libraries (linux modules are not for sure)
4) They are executed as modules of kernel, no matter what
is the mechanism that trigers it and no matter if its
super user or normal user to triger it.
5) They are both not executables and should not be in path.
If you have any other argument against translators being in
/lib/modules other than "I said so and you don't understand
how hurd works", please give it to us.
And yes, I installed hurd myself and played with translators,
I am even trying to make one right now.
-Jiri
PS: What I am trying to say that the difference is more rethorical
than fundamental.
--
Jiri Klouda <jk@zg.cz>
http://www.zg.cz/~jk
Attachment:
pgpiiqre3h0W3.pgp
Description: PGP signature