--- Rafael 'Dido' Sevilla <[EMAIL PROTECTED]>
wrote:
> The monolithic
> design is probably not
> going to be best for high scalability and for
> distributed computing
> environments.
> 
This is one serious issue the developers must contend
in the 2.5 development. Microkernels are fine - at
least mostly in theoretical computer science - but a
very good and free implementation of a
microkernel-based OS has yet to exist (I wouldn't bet
on the exokernel's commercial existence in the near
future).

Should the developers go around this limitation and
come up with a Linux microkernel, they'll be able to
scale up (all that needs to be done would be the
creation of various servers for the tasks needed,
similar to the Mach's implementation) or make Linux
extremely portable (maybe even devices much, much
smaller than the Palm would running be running Linux
reasonably someday?) But wait -- this involves an
overall rehaul of millions of lines of code! Aside
from it, lots of research (what happened to HP's and
IBM's commitments to fund Linux research?) and testing
would be necessary, making the transition probably the
longest compared to the developments of the stable
kernel releases from the development ones.

> They gave away source
> licenses to
> universities such as UC Berkeley, but BSD Unix back
> then could not be
> used by anyone else without buying an expensive
> source license from AT&T
> because they owned major parts of the code.

Before System 7, the AT&T license was more relaxed, as
they did gave away source code licenses for research
to major universities at that time. When the guys at
AT&T realized that Unix is fast becoming a bigtime
cash cow due to its popularity, they really closed the
source for good starting with System 7, with
maintaining the commercial viability of Unix as one of
their reasons.

This is why universities only taught operating systems
design in theory as they dropped teaching the source
code of Unix from the curriculum. However, it really
gave students a distorted view on how an operating
system is supposed to work (Andy Tanenbaum presented
these limitations quite well in his book on OS
design). UC Berkley was quite fortunate to have at
least made some progress in making a version of the
commercial Unix by writing their own code; with too
much AT&T code still inside BSD Unix they couldn't
distribute the OS, only the software produced on it.

> Around the time Linus began working on Linux (1991),
> Berkeley CSRG
> removed the remainder of the code owned by AT&T, and
> Bill and Lynne
> Jolitz rewrote the excised portions and began to
> complete the port to
> the 386.

Thus the evolution of OpenBSD, NetBSD (and its
offshoot FreeBSD) and other BSD type OS's.

During the days when the Linux kernel was really young
(just before 1.0), even Linus Torvalds said that he
expected a free BSD release would overtake Linux
should it be released. That simply isn't the case
today. I wouldn't say the same about the Mac OSX, as
it's gaining popularity and could really threaten
Windows on the desktop soon [A port of OSX on Intel
hardware may come(?) soon]  but doubtful however on
the server side, wherein fullblown Unix systems and
Linux dominate. 

As Torvalds would later admit, the Linux kernel itself
wasn't the phenomenon that shook the computing world -
rather it was the development process that Linus
started which made Linux very popular. The speed of
development, as well as the openness of development
attracted far too many people not only to work on the
kernel, but also on application programs and
filesystems. It proved that the bazaar could work even
in major projects such as creating an OS, and how
viable this method can be.

Paolo Alexis Falcone
UP Manila

__________________________________________________
Do You Yahoo!?
NEW from Yahoo! GeoCities - quick and easy web site hosting, just $8.95/month.
http://geocities.yahoo.com/ps/info1
_
Philippine Linux Users Group. Web site and archives at http://plug.linux.org.ph
To leave: send "unsubscribe" in the body to [EMAIL PROTECTED]

To subscribe to the Linux Newbies' List: send "subscribe" in the body to 
[EMAIL PROTECTED]

Reply via email to