[Ilugc] Take on Minix & Linux by Giris
- From: girishvenkatachalam@xxxxxxxxx (Girish Venkatachalam)
- Date: Wed Aug 30 11:27:26 2006
On 8/29/06, amit saha <dandolls@xxxxxxxxxxx> wrote:
For Girish
With all due respects , girish I ask you one simple q what do u mean when u
say " dont do projects of academic interest " ? cud u plz enlighten me who
ur intended audience is wen u say somthn so "irresponsible" like that ? What
kind of projects should a student do ? What should be his goal? Money,fame or
to learn something new. To learn something which his text books didnt teach.
I dnt know abt u ,whether a stud or a prof...wat ever ! but at som point of
time u must hav been a student (i hope i am not making assumptions)...n then
what did u do projects for ? to learn....or for som thn else...neways..if u
thnk otherwise( wich i presume u do)..do let me know abt projects u did wen u
wer a student... I wud like to learn a few thns frm u then !
Amit,
You have asked too many questions here.
Pls understand that my words were interpreted out of context and
sentimentally analyzed. Nobody took the time to understand what I am
feeling deep inside when I spoke those words.
Of course my words were incendiary in nature and I had taken an
extremist stance which I now regret. Extremism does do good but it
leads to lot of bloodshed in the short term.
I personally have no anathema againt anything or anybody. In fact I
love Win XP for its stability , the applications supported by it and
ease of use. There was a time when I used to hate Microsoft and used
to love UNIX. But after several years I realized that any religious
allegiance or hatred only does harm and no good.
At the same time when MBA folks tell me that technology doesn't
matter, only result matters; I disagree with them vehemently bcoz
unless you use the right technology and the right tools you cannot
achieve your end.
And you cannot become good at something by being technology agnostic.
For now, to answer the most important question as I understand at the
top of your mind, here goes.
By "projects of academic interest" , I dont mean academia or industry.
There are several real life projects developed in universities. google
is an academic project, TCP/IP was developed in UCB and SUN is
Stanford University Network.
So ppl will laugh at me if I say that academia should only teach and
engineers should only do real life projects.
What I mean by "academic interest" are the sort of futuristic research
done by IBM T J Watson research center, AT & T Bell Labs and MIT media
lab projects. No, I dont mean to make a blanket statement here.
What I mean is many of these things dont matter at all to the common
man. They are done for pure research purposes.
This is what I don't like. No, I have nothing against research. But
from whatever I have observed about India and the Indian mind all
these years I feel that everyone goes into R & D mode at the drop of a
pin taking shelter in theory alone.
At the same time I have also met a lot of very good coders who start
coding at the drop of a pin.
What I think works is a proper balance of the two.
You need a solid conceptual and theoretical grounding first. On top of
that you build a product and then, once people start using it, you
will understand that what people want is really very different from
what you started with.
And you will have to make lot of changes for portability, for bug
compatibility, do ugly things like use the sleep statement for getting
it to work or put in hacks to make it not fail under certain
circumstances.
IOW, I mean to say that even if you have the brain of Albert Einstein,
unless your product is proven in the crucible of the real world, when
thousands of people use it in situations that Einstein himself couldnt
imagine(largely bcoz Einstien is not as stupid as an average Joe user
:-), until then, the product cannot be trusted.
i.e, I mean to say that a product should become truly idiot proof and
you should fix corner cases and other bugs that crept in due to your
assumptions which no longer hold good in the real world.
Now, coming to Minix, the statement sent by Roshan that Minix really
gave folks a chance to write device drivers... This might have been
true in the early 90s, when nobody knew about the open source
phenomemon. At that time, nobody had access to any kernel code. Only
employees of Sun , Novell , IBM, DEC etc...
But now, a lot of water has flown under the bridge. Today open source
has moved out of academia and hacker's hobbies and become more
mainstream than any other thing.
It came as a deluge and every single company in the world today should
owe a certain gratitude towards the great people who wrote open source
programs.
Believe me, I have seen some of the best companies in the world taking
open source code, only slightly modifying it, throwing in some masala
and then making products out of it. Very few of them actually follow
GPL laws. In fact many companies don't even acknowledge that they use
open source.
Whereas BSD licensing is very very business friendly. You can use the
code and do whatever you want with it. Moreover BSD code is also rock
solid unlike linux. Linux may have 1000s of features but when it comes
to doing something and doing well , nobody can beat BSD.
So it is only in this sense that I said that talking about Minix two
decades after it was born , especially when Minix has transmogrified
into linux and since Minix is hardly used is braindead to put it
euphemistically.
At the same time if you have nostalgia about something like the TCP/IP
stack or C or basic UNIX shell commands like cat or vi, then that
makes sense.
Bcoz though equally old, these programs have weathered real life use
and proven themselves. The fact that TCP/IP or cat has hardly changed
speaks very highly of the foresight and the brilliance of the
designer.
Hope I make myself clear now.
regards,
Girish
PS:- Very sorry for writing such a long mail and wasting your time!
Other related posts:
