At 09:28 03/04/2003 +0200, Wojciech Puchar wrote:
> > >> I [...] asked, "Why a full blown OS instead of some smaller
> > >> real-time thing?"
> > >> He replied, "because a megabyte of RAM only costs $1."
>
>what i'm scared the most in computer market are people that think this
>way.

The problem comes from those who buy, not those who sell:-p

Anyway, worst than that is the fact that:
- a company starts a new product. but since they don't know whether
it'll be successful or not, they minimize engineering expenses (cos'
they have to invest on marketing and busdev:) and produce a low grade
version (cuts always go in the same place) The idea being "If people love it,
we'll invest on fixing/improving it".
- people start asking for.
- but then no time to fix, no time to improve.
- they patch. fixes and improvemenst get sceheduled for a better day.
- repeat ...
- the product is finally too complex, and the only serious fix is a complete
rewrite, but the cost-benefits analysis doesn't justify that.
- more pressure on engineers: here goes the "patch me up" dance...

In short, the best products are those which have no users, as they can be
designed, developed, maintained, with no market (or users in the case of
free products) pressure...

Now back to the subject, I think it is safer for a company to use a well-known
OS than to write its own. While that's not optimal, it saves a lot of 
headaches.
Incidentally, it alos provides the possibility to let 3d parties develop 
addons.
As for speed/power/..., some examples that come to mind:
- Effnet used (dunno if they still do) to sell "fast routers" running NetBSD.
- NetApps use a modified FreeBSD for their appliances.
- Cisco are supporting Linux in many places.
- ...