Maneesh Yadav 97yadavm@scar.utoronto.ca
Sun, 30 May 1999 13:52:39 -0400 (EDT)

<Reminder let's keep this objective and not a "I'm a better coder than you
are", I have been through my computer programming, and yes I have I coded
my 3d asm engine and a bootable image of my incomplete ukernel, I have
worked with some electrical engineering profs at U of T with Tornado and I
am doing a lot of work with molecular modelling, genetic algoirhtms and
graphics here at the Scripps Institute, and I get paid to do it; no I
don't claim to be a smart person, but I think I know how to program> 

On Sun, 30 May 1999, Tom Novelli wrote:

> On Sat, May 29, 1999 at 05:30:03PM -0400, Maneesh Yadav wrote:
> > 
> > On Sat, 29 May 1999, Birkas Mate wrote:
> > > I had very similar feelings and thoughts as Maneesh had while i've read
> > > the letters in the list, but i really don't get the point, why does any
> > > others work hurt him (you) so much, even if it's a lower level stuff like
> > > Retro.
> > 
> > It really doesn't hurt me at all.  But when you see someone going about
> > something the wrong way with the wrong intentions, and when you know
> > they're saying stuff that they cannot achieve because they are not
> > familliar with the problem...it's hard not to stand up and say something.
> > Tcn is saying that there will be drivers, games and his OS will suppourt
> > all these wonderful things that new OS should (as cited in my previous
> > letters) it shows a real ignorance of the field.
> On the contrary, what you're saying suggests to me that you haven't gotten
> your hands dirty, that you haven't bridged the gap between theory and
> reality, that your head is up in the clouds somewhere.  I can achieve
> anything that can be achieved, all it takes is dedication.  I'm getting
> these things done, a piece at a time.  Tomorrrow I'll have it installable on
> the hard drive... orthogonal persistence isn't far off.  No it won't do what
> TUNES is supposed to do... but it'll be fast and simple, and that's what I
> care about.
> > He's already defining his memory mangment scheme, tasking etc; when he
> > hasn't seem to have even really consider the bigger issues, or the the
> > that no small team will be able to inpliment them efficiently in some real
> > time span.
> Where did you read that line of bs?!  Manpower is great for construction,
> farming, assembly lines, warfare... but software design?  That's like a
> collaboration between a hundred novelists -- the result will be huge and
> disjointed, if it's ever finished.  In all my experience, individuals and
> small groups get things done faster and better. 

Then you haven't worked on a project of sufficient complexity.  When you
really work on something new, the simple stuff has already been done, and
you need specialists in respective fields to co-dinate and work together
to reach some goal.  Overly large organizations of course don't work too
well (because we don't have the ability to co-ordinate in larger numbers).
Software design is one of those things where there is increasing
complexity, and in order to get anywhere, you need lots of really smart

 It also helps if you do
> more work and less reading and talking... not that an occasional argument
> isn't fun :-}
> Is TUNES is too big?  You split off your project, I'll split off mine...
> retro has enough momentum on its own.  Wait, what difference would it make? 
> These are two complementary projects (i.e., HLL and LLL).  TUNES wouldn't be
> complete without either of them.. hence the term "subprojects".
> > As I said, until you consider the things that are truly challenging
> > (distrbuted, parallel, portability, GC etc. etc.) this is nothing more
> > than an exercise in an undergraduate intro to OS's.
> I've considered all these things (there, now you know).  

No you haven't.  I asked questions specifically about those topics and I
didn't see anything concrete in there as to how they will be implimented.
I invite you to describe your basic mecahnism's agin and how the baove
will be factored into design, specifically... 

I never mentioned
> them because experimentation and research have answered all my questions,
> and I don't have any earth-shattering new ideas to share.  Undergraduate
> work, perhaps.. but it's progressing fast.  At this rate, you'll be using my
> OS to write your doctoral dissertation ;)
> Some folks believe you can automate everything.. simply create an
> ultra-high-level system and it'll take care of all the low-level stuff. 
> There's no substitute for hard work, human intelligence, creativity and
> common sense.  I'll do the critical stuff by hand.  What good is distributed
> computing if overhead takes up 90% of the cpu time?  And what good is AI if
> you're long dead by the time it comes about?

Certain parts of computing intensive stuff do have to be coded by hand,
but the ultra-high-level system doesn't exist yet, how can you comment on
it's performance? 

One more thing, you haven't addressed why you think Linux is bloated and
inefficent; by the time you impliemnt all the things it has, yours won't
be much smaller/bigger...even in terms of CPU time, Linux does a good job
of managing stuff....

> Oh well, enough of that.  It's like the pointless argument between pure math
> and applied math.  Now back to work.. I mean sleep..
> --
> Tom Novelli <tcn@tunes.org>
> "You just dropped a hundred grand on an education you coulda had for a
> dollar fifty in late charges at the public library?" -matt damon