FreeBSD's embedded agenda - filesystems

[Olivier Gautherot]

> >>>"The adequate has always been the worst enemy of excellence"

> >>'Good enough means exactly what it says, and doing more is foolish'

The one I knew is "The better is the worst ennemy of the best" :-)

Good to see these arguments: it means that people are alive and proud
of their positions - and that's what keeps a community moving!

I totally agree with Andrew that if a product does not make it through
the door, it has no value, only cost. That's the law of industry.
When Poul-Henning says we shoud strive for the best: well, we're on
a free project and no one will be fired if it takes 2 more months.
We're all proud to support this great system and I think we should
thank all those great contributors that strive for the best and drive
the industry.

Still, we have to settle on an approach (or 2). Embedded is a peculiar
market where *BSD is still lacking presence. The lack of filesystem
is probably central to this and needs to be measured but, before all,
we need to check what type of application we want to approach: for a
router, my best guess is "anything will do": you don't (or seldom)
write to the file system. A monitoring device will definitely be more
demanding. Going out with the wrong set of features or the wrong
timing may kill the project in the egg.

The question of the day, for me, is: what market(s) are we targeting?
Where do you see possible applications? I would love to see it on
audio/video over powerlines, monitoring systems, firewalls (obviously),
robotics and more. The OS in itself has a great potential. The question
is where do YOU feel the foot-in-the-doorstep is and what do you need
to make it happen?

Creating the filesystem is time and work. The ideal situation is to
have the product in hand and, on top of it, have the product be the
advocate: demonstrate that BSD is just as good, if not better, than
the others and then show that the license is the winner (I think we
all agree on this one). That's why I am starting the study from the
ground up - maybe an existing FS will do the job, the only way to
find out is to check our goals against their features.

Disclaimer: "Just as good" does not mean "another mediocre" but
"definitely another competitor in the marketplace" - IMHO, performance
will do the rest.

Just my opinion
--
"If there is no solution, there is no problem"
           Shadok proverb
----------------------------------------------
Olivier Gautherot
olivier at gautherot.net
Tel: +56 8 730 9361