Air conditioning has helped to destroy the beauty of summertime. As a boy
growing up in the pre-airconditioned South, you could walk
the streets at twilight amidst a restful summer quiet. It
was a quiet that is hard to describe. Not silence, but the
peacefulness of a community winding down at day's end. Kids
might be playing in the yards or loafing on a
porch. You might hear the quiet sound of a radio through an
open kitchen window. You'd see someone washing dishes, bathed
in the pale yellow glow of her kitchen light.
Outdoors
the sky would be turning darker as a shadowy purple
became the predominant tint to the surroundings. The
most prevalent sound was the synchronoized chatter of cicadas (locusts)
with their bizarre rhythm of cyclic rattling. Oh, and of
course their were the silent fireworks of the fireflies.
Now when you walk the street at
dusk, you see no one, not even someone washing dishes,
thanks to the ubiquitous dishwasher. Kids are nowhere to
be seen. The steady drone of each and every house's air conditioning compressor
fills the air, drowning out even the cicadas. You might
as well be walking through a 24 hour per day widget
factory. It is an industrial noise which blocks out all sounds of nature.
Sky watchers complain of light pollution; I would like to
add to their complaint, the noise pollution of air conditioners which have
helped to destroy the summer night.
One area where Linux could use the help of IBM
is in making better engineering tools
available for Linux. There are very few industrial
strength software engineering tools available
for Linux. The various CASE tools available
are usually little more that toys or "proof of
concept" projects left over from some one's
graduate program.
At present, Linux relies too much on a Monte Carlo shotgun approach to software engineering (thousands of folks randomly hammering on software to
shake the bugs out). What
is lacking is the industrial strength support
for formal engineering methods of design, testing, and QA. IBM has plenty of expertise in this
area. My hope is that IBM will step in and plug
the the hole.
Of course IBM's help in educating
developers in the benefits of formal industrial strength development
methods wouldn't hurt either. SGI has done
some work in this area, offering a suite of regressions tests for the kernel. However, how
many people know about them, how many actually use them?
The best memory state analyzer available for Linux is
GNU Checker. Far, far more than a malloc library, GNU Checker is GCC based. Checker automatically instruments every
memory access in your code when you compile
your code with it.
It detects bad calls to malloc/free, memory leaks, uninitialized
data structures, and all sorts of other memory problems.
I've used almost all the memory debugging tools and libraries
on Linux at one time or another. Most of them are pretty good
and each has its place. However, the Swiss Army Knife of memory
debugging tools for Linux has to be GNU Checker. It is the
most sophisticated of the lot, and it's the nearest to Purify
in power. Unfortunately it is also one of the
best kept secrets when it comes to Linux development tools.
Easiest solution would be to for FreeBSD to
change partition numbers. Problem solved.
FreeBSD is small beer in the OS marketplace,
so if the mountain won't come to Mohammed, maybe
Mohammed should go to the mountain.
As far a Linux users are concerned, the new IBM fiat is ``Linux everywhere'' -- mainframes, minis, handhelds, clusters, ad extremum. IBM is methodically
synching its divisions to the Linux way. IBM
has invested one-half billion dollars this year in Europe
and Asia on Linux development. They'll get the
notebooks right eventually. No problem. Nonetheless, thanks for your concern.
Speaking as someone who works at IBM, all future IBM NCs will be shipping with Linux. Despite the the advocacy posts here, NetBSD has been very problematic for us at IBM. The main reason we are moving to Linux is increased flexibility, and standardization of development tools. What most advocates miss (whatever their stripe) is that we are in business to make money. Cutting internal support costs is ``job one''.
Name recognition itself is worth many millions of dollars. If you have name recognition, your marketing department is halfway home.
When two very similar entities compete for the same economic niche, one will eventually dominate. Coyotes drive out foxes.
Linux is the dominant Unix operating system. A very similar Unix system has no chance in that niche (see second point above). Why give up millions of dollars in mindshare. You'd have to pay for a quarter-billion dollars in SuperBowl advertisements to even hope to approach Linux mindshare. That's the true free gift that IBM is getting. Other than Windows, the most likely challenge to Linux is from systems that differentiate themselves radically--BeOS for instance.
I hate to burst your bubble, but the BSDs had their chance. The world has changed since the PDP11 days of yore. If you want a scapegoat don't blame Linux. Blame Kurt McKusick of CSRG who refused to port BSD to the Intel architecture, despite requests dating back to 1986. By the time Jolitz had something to offer, six years had gone by and Linux was already on the rise. The BSD CSRG died shortly thereafter.
Let that be a lesson in elitism and snobbery. Isn't it ironic that McKusick's Moto 68000 is obsolete but the the Intel architecture which he spurned now owns 90% of the CPU market.
Slashdot Mirror
Outdoors the sky would be turning darker as a shadowy purple became the predominant tint to the surroundings. The most prevalent sound was the synchronoized chatter of cicadas (locusts) with their bizarre rhythm of cyclic rattling. Oh, and of course their were the silent fireworks of the fireflies.
Now when you walk the street at dusk, you see no one, not even someone washing dishes, thanks to the ubiquitous dishwasher. Kids are nowhere to be seen. The steady drone of each and every house's air conditioning compressor fills the air, drowning out even the cicadas. You might as well be walking through a 24 hour per day widget factory. It is an industrial noise which blocks out all sounds of nature.
Sky watchers complain of light pollution; I would like to add to their complaint, the noise pollution of air conditioners which have helped to destroy the summer night.
At present, Linux relies too much on a Monte Carlo shotgun approach to software engineering (thousands of folks randomly hammering on software to shake the bugs out). What is lacking is the industrial strength support for formal engineering methods of design, testing, and QA. IBM has plenty of expertise in this area. My hope is that IBM will step in and plug the the hole.
Of course IBM's help in educating developers in the benefits of formal industrial strength development methods wouldn't hurt either. SGI has done some work in this area, offering a suite of regressions tests for the kernel. However, how many people know about them, how many actually use them?
The best memory state analyzer available for Linux is GNU Checker. Far, far more than a malloc library, GNU Checker is GCC based. Checker automatically instruments every memory access in your code when you compile your code with it. It detects bad calls to malloc/free, memory leaks, uninitialized data structures, and all sorts of other memory problems. I've used almost all the memory debugging tools and libraries on Linux at one time or another. Most of them are pretty good and each has its place. However, the Swiss Army Knife of memory debugging tools for Linux has to be GNU Checker. It is the most sophisticated of the lot, and it's the nearest to Purify in power. Unfortunately it is also one of the best kept secrets when it comes to Linux development tools.
As far a Linux users are concerned, the new IBM fiat is ``Linux everywhere'' -- mainframes, minis, handhelds, clusters, ad extremum. IBM is methodically synching its divisions to the Linux way. IBM has invested one-half billion dollars this year in Europe and Asia on Linux development. They'll get the notebooks right eventually. No problem. Nonetheless, thanks for your concern.
Speaking as someone who works at IBM, all future IBM NCs will be shipping with Linux. Despite the the advocacy posts here, NetBSD has been very problematic for us at IBM. The main reason we are moving to Linux is increased flexibility, and standardization of development tools. What most advocates miss (whatever their stripe) is that we are in business to make money. Cutting internal support costs is ``job one''.
- Name recognition itself is worth many millions of dollars. If you have name recognition, your marketing department is halfway home.
- When two very similar entities compete for the same economic niche, one will eventually dominate. Coyotes drive out foxes.
Linux is the dominant Unix operating system. A very similar Unix system has no chance in that niche (see second point above). Why give up millions of dollars in mindshare. You'd have to pay for a quarter-billion dollars in SuperBowl advertisements to even hope to approach Linux mindshare. That's the true free gift that IBM is getting. Other than Windows, the most likely challenge to Linux is from systems that differentiate themselves radically--BeOS for instance.I hate to burst your bubble, but the BSDs had their chance. The world has changed since the PDP11 days of yore. If you want a scapegoat don't blame Linux. Blame Kurt McKusick of CSRG who refused to port BSD to the Intel architecture, despite requests dating back to 1986. By the time Jolitz had something to offer, six years had gone by and Linux was already on the rise. The BSD CSRG died shortly thereafter.
Let that be a lesson in elitism and snobbery. Isn't it ironic that McKusick's Moto 68000 is obsolete but the the Intel architecture which he spurned now owns 90% of the CPU market.