Linux kernel R&D worth over 1bn euros

I don't know what method they used (to lazy to bother to find out), but I hope they took into consideration that Linux is not a pile of code, but a constant stream of changesets, some of which take several attempts before they get merged and their development sometimes takes moths or even years. A lot of code gets changes time and time again. A a lot of code gets rejected, too, or even developed but never posted - and this code too would generate "R&D" costs (actually it's probably *the* cost of "R" in "R&D" - finding out what is feasible and what is not, and you can do that only by trying).

So I'd hazard a guess that the cost to develop Linux kernel is much more than even that.

> So I'd hazard a guess that the cost to develop Linux kernel is much more than even that.

One heck of a lot more.

There is simply no way to replicate with just "money" the sheer QUANTITY of testing that every Linux kernel goes through with the simple expedient that every user has the capability of communicating directly to the developers when they find a problem.

I do not believe it would be possible to deliberately create what we call the Linux kernel.

Any OS deliberately created would be limited to the hardware that it was designed for. And I am not going to say that such a system would not be better than what we now call Linux _on_that_hardware_, but nothing that isn't "organically grown" could work as well on all the different hardware platforms, 16, 32, 64 bits, embedded, mainframe and virtualized, all that the same time.

(ok, I'm not certain that 16 bit Linux still exists)

Bob -

On 16 bit Linux -

Eazy Peazy.

Nothing that a LiveCD of a 32 bit distro and a good pair of tin snips won't handle.

jez,

Thanks, problem solved.

Quoting: [...] Linux is not a pile of code, but [it ...] sometimes takes moths [... to] gets changes [into it]

Now I know how all the bugs got in. Probably put a leg across an open circuit frying it on the spot. Poor insects, to suffer so dearly for out mistakes.

YBT

Quoting:ok, I'm not certain that 16 bit Linux still exists

I don't think it ever existed... AFAIK Linux was from the get-go designed specially for Linus's 80386 box. Granted that Linus never thought that any other architecture would be suported, but going back to 16-bit from there must have seemed an extremely unattractive :)

> I don't think it ever existed...

It was a work in progress at one time: http://elks.sourceforge.net/

But I'm not sure it ever reached a working state.

GeoWorks Ensemble supported virtual memory on my ancient Sanyo luggable XT with its 30M IDE HD.