"Valentin Ossman" <[EMAIL PROTECTED]> writes:

> I guess you're in here just for the sake of debate. 

Actually, no. I asked a question, and I also hoped it would be obvious
that I had read the description of your proposed product. In fact, I
had read most of the material on your site, and got interested. I
don't know where you got this "debate" idea from, but let me
rephrase the question.

I got the impression (which may be wrong, and you are welcome to
correct me) that the revolutionary character of the product lies in
the protocol stack you propose to implement in hardware that will
yield much better performance then the current software stack. To
utilize this capability you will need to "revolutionize" the kernel's
networking code, which strikes me, on the face of it, as a bit much
for a student given a reasonable amount of time. 

The web page mentions that the card will also be able to use the
unmodified software stack in an "OS-independent mode". In this case
you won't have to do much kernel work apart from an ethernet card
driver, but you will not get much "network acceleration" out of it,
will you? It does sound like a compatibility mode...

So what's the goal and the scope of the proposed activity?

You should be able to answer without any NDA since this question is
likely to be asked by any prospective job candidate, IMHO.

> I would recommend, in this case, the room next to the Academy of
> silly walks. (rent a copy of "The Flying Circus" from Anak Ha-Video,
> if you don't get my point)

It's not the case, and I didn't get your point, and I don't own a
VCR, but an explanation would be off-topic, I suspect.

-- 
Oleg Goldshmidt | [EMAIL PROTECTED]

=================================================================
To unsubscribe, send mail to [EMAIL PROTECTED] with
the word "unsubscribe" in the message body, e.g., run the command
echo unsubscribe | mail [EMAIL PROTECTED]

Reply via email to