On Wed, 27 Feb 2008 09:36:44 +1300, Pete Dashwood wrote:
> "tim" <TimJ@[EMAIL PROTECTED]
> wrote in message
>>
>> You must be pretty good if you can write 15,000 lines of
>> multi-processing kernel code in a weekend.
>
> Well, leaving aside the fact that I AM pretty good :-), it was said with
> tongue-in-cheek...:-)
>
> As a matter of record, I once wrote a complete access method for a
> mainframe in a Bank, over a weekend...
>
> This is the size of the ext3
>> file system which is the predominant file system used in Linux.
That was not multi-processing kernel code. That's at another level
of difficulty entirely. From your description, your code assumed a file
system already in place and assumed an access method (RRDS) already in
place. Not that that's a bad thing - the "Google File System" does much
the same thing.
The equivalent to a Linux file system would be to start with a bare disk
and the startio macro and an assembler. You need to make sure your code is
proof against timing errors, it must perform really well for a huge
variety of system architectures (eg Numa with 1024 CPUs through an ASUS
EEE on a Celeron) and applications, and it must be hacker-proof.
> Like I said, my comment was not meant seriously...
I understood that, though it struck me that that the possibility existed
that you were just pretending to be tongue in cheek so you could show off
while maintaining plausible deniability. The content of the parent post to
this one may shed some light on this possibility.
Reiser's file system is a considerable achievement and probably qualifies
him as a geek. By the way Reiser's file system is actually a lot bigger
than ext3 - not sure whether that's a point in its favour or not. The fact
that other people on the kernel mailing list trenchantly criticised it is
neither here nor there - such a style of comment is pretty standard there.
>> His attorney's tactics smack of desperation I think. But I have no idea
>> whether he did it.
>
> Yes, I agree. I also have no idea whether he did it and would not make
> any kind of judgement until I had reviewed the evidence.
>
> Nevertheless, it is an "interesting" defence...
>
> Pete.
I suppose the attorney is trying to make the point that the Hans is a
different person. So the jury member should not assume that Hans would
behave like them. This is a reasonable and valid point IMHO.
Some have claimed here that programming ability is orthogonal to
personality. From my research and experience this is not the case. For
example, the incidence of autism in the children of Microsoft employees is
so high that MSFT has a special program for autistic children (geeks
breeding with geeks).
Tim
|
