I don't agree with RedHat's decision to ship
what was essentially a snapshot, in RedHat's
defense I have to say that they were faced
with a dilemma.
They could either:
Ship with what is probably the buggiest
release of GCC since early days,
Delay their release for some indeterminant
number of months while the GCC folks either
finish their ABI and decide to make an interim
release of GCC or finish GCC 3.0 althogether,
or
Clean up a snapshot (and almost any random
post-2.95.2 snapshot of GCC has been better
than 2.95.2) and release that.
They chose the last option, knowing that
there was no possibility of having 3.0
compatibility aside from option #2, and that
they'd at least get a stable and largely
standards-compliant C++ compiler.
After RedHat chose a snapshot, they
continued to follow subsequent snapshots
but because of the necessities of
release engineering created patches against
the original snapshot, which has lead to the
accusation (largely unwarranted) that they
have forked GCC.
Personally, I think they should have stuck
with 2.95.2, warts and all, but it was a
judgement call for them, and the path they
took is not without some justification.
(BTW, the FreeBSD folks are so disgusted
with 2.95.2 that they're considering making a
similar move.)
How're you
supposed to implement an atomic test-and-set if the kernel can pre-empt you in the middle of it?
This is a hardware, not an OS issue.
And sure enough, Alpha (like Pentia) has an
atomic test-and-set instruction!
Good thing, too -- you need this sort of
instruction to implement mutexes and other
locks both in kernel and user space.
There is no efficient software workaround
for the lack of such an instruction.
Cool; perhaps you did.
Nobody I knew did, but I'm hardly omniscient.
However, what it was called has little
to do with my
original point: the "Internet" of the day
(whatever it was called) isn't
what we call the "Internet" today.
Perhaps a bit of residue left over from the
changeover of the early '90's will convince
you.
Network 10 (addresses where the first octet
is "10") are not routed over the Internet.
Addresses in this range are used for internal
networks, for testing, and what-not.
Well, guess what the first octet of all
ARPANET addresses was?
Yup, it was 10.
Back when I first started working for them,
RAND's address on the ARPANET was 10.0.0.7
(and not coincidentally, they were about the
seventh node to join the ARPANET).
At the time of the MILNET split, RAND
received the new address 192.5.14.* (along
with a few other class C's) which it has today;
needless to say, the old address, like the
ARPANET itself, no longer exists.
My original point remains: several acts of
Congress opened up and funded the Internet
as we know it today, legislation that
came out of Gore's office.
You can love him or hate him for other reasons,
but he at least deserves credit for that.
The terms "internet" and "internetworking"
have existed from the beginning
as names for the technology.
But until the MILNET split, no one
called the network which used those
technologies "The Internet."
It was "The ARPANET."
Well, the most technically astute president
in the last several decades (and perhaps
ever) was Jimmy Carter, who was a nuclear
engineer.
He was also the last man who had a credible
claim to being a "Washington outsider,"
and one of the most honest men who has ever
occupied the White House.
And he was a mediocre president, at best.
Thus, I'm not necessarily faulting Gore -- he
at least knew who to listen to, whether he
himself was technically savvy or not.
So if I were voting just on the basis of who would
most likely be net-friendly (and not
just follow corporate interests and the wishes of
the national security apparatus), I'd give
him the nod.
But it's real hard to get much enthusiasm over
either of the major candidates.
In 1990, I had been using the
internet for about 10 years.
You're being just as misleading as you accuse
Gore of being.
The "Internet" didn't exist in 1980.
Its ancestor at that time
was called the "ARPANET,"
and it differed in many ways (aside from size
and speed) from today's Internet.
It was run for the Department of Defense,
with an (often ignored) "official use only"
policy and a (strictly enforced)
non-commercialization policy.
Nodes were either military bases, military
research labs and contractors,
or universities such as
MIT and Berkeley which were performing
military research.
Yes, there was an "underground" that used
it for socializing and hacking, made up of
people at these institutions (and a few of their
friends).
But such "unofficial" use was quite low-key,
and at least among ARPAnauts was defended as
being a matter of academic freedom (while
the DoD generally looked the other way).
But as traffic increased and budgets were
cut, the DoD's benign oversight changed.
In the late '80s the military insisted on cracking down on "unofficial use."
Many schools and contractors
wound up having to justify their access,
and came up short.
Efforts to start member-supported networks,
such as CSNET, got a lukewarm reception
since they (1) cost too much (no more
100% government subsidy) and (2) did too
little (email and limited file transfer only).
It took an act of congress to get these folks
hooked up again: the ARPANET was split into
the MILNET and NSFNET (NSF == National Science
Foundation), with the latter offering subsidized
access to any academic institution that wanted
it and paid its share.
And it was after this split that users starting
refering to it as just "The Internet" (though
from time to time I heard ARPANET referred to
as "The ARPA Internet" back in the '80s).
Guess who wrote the bill that funded NSFNET?
Or the National Supercomputing Initiative
which built its backbone?
Or the one somehwat later that lifted restrictions
on commercial use?
Yup, Smilin' Al (or, rather, one of his staffers
wrote the bill and he sponsored it).
When you look at it from a technical perspective,
saying that he "created" the Internet is a clear
absurdity.
But from a legislative and public policy
perspective, it's hardly an exaggeration.
I'm not convinced that Al "gets" the Internet
any more than W does.
But he did manage to listen to someone on his
staff who "got" it.
(And I've been racking my brain to remember the
guy's name--I met him briefly while I was working
at The RAND Corporation.)
It's too bad that so few technophiles go to work
as congressional staffers;
it leaves me with some doubts as to whether Gore
wasn't just "lucky" to be involved with the
Internet.
ts lame to use a small number of sites to base the conclusions on.
Wow; after 200 posts on the subject yesterday
and today, this is the first post I've seen mentioning this simple fact.
I'd hope that/.'ers would be a little more
statistics-savvy (they usually are).
His problem is that he
believes that the patent they are applying for
(in his name!) is overbroad, and possibly invalid
(prior art).
I understand that, but he probably signed away
the right to make that choice.
My point is that he didn't need to.
Perhaps he could have agreed not to assign
inventions to anyone else or to patent them
himself, thus reserving the right to make
the choice he now wants to make.
But that's not what he agreed to, and now he
wants to take back that agreement.
That's usually not possible...
They can't make him, but they probably
can sue him!
Although there are limits on how far employment
contracts can reach beyond employment,
they don't become null and void just because
you end employment.
I'm no lawyer, but I've consulted with a lawyer
every time I've signed an employment contract,
sometimes negotiating changes in clauses I
found offensive.
Given his beliefs, I suppect this fellow should
have done this, too.
Those pieces of paper people sign when they
become employees aren't just for show.
I'm wondering why this fellow signed an agreement
to assign inventions to the company if he
didn't plan to keep it.
He should have asked to have the offending
language removed or changed.
Granted, there is a good chance he'd be shown
the door.
Or he might have had to make concessions
on compensation or in other areas.
It depends upon the company, and how much his
services are worth to the company.
But taking a moral stand cannot be an
afterthought.
McCabe's Cyclomatic Complexity can be misleading, especially for assemblers, simulators,
and other software
which must represent a model of hardware.
The same is true for compilers, interpreters,
emulators, and other such software.
In the case of an assembler, dividing opcode
dispatch or address mode decoding across dozens
of metric <10 functions can
make the correctness and completeness of the
assembler harder, not easier, to determine.
(However, it is appropriate that a thorough
knowledge of the architecture of the machine
the assembler is for be required for such
understanding and determination.)
It's the nature of the beast, though it is
possible to build or obtain tools that allow
for a more direct representation of the target
hardware (e.g. GAS uses such a method).
Complexity metrics have the flaw that they
encourage (or in some shops, enforce) the
diffusion of complex but highly regular
logic which is better represented as a whole.
It may be impossible to rate the actual
comprehensibility
and maintainability of a module based on the
number of edges in its control flow, as
McCabe's metric attempts to do.
The predicates which generate those edges
may model
a regular structure such that drawing them
together into a single module makes it easier
to understand, even though its McCabe
metric is high.
Conversely, a control flow with a few edges
generated by predicates with obscure interactions
and dependencies, such as would be required
by the division of certain problems into
submodules,
might yield low McCabe metrics and be nearly
impossible to understand.
Inappropriately applied, metrics like McCabe's can
encourage the generation of artifacts
based not on the inherent structure of
the system being modeled, but on the satisfaction
of blind rules.
In some application areas (business logic, for
instance) metric like McCabe's work well.
They can be useful even with parts of systems
where they are otherwise inappropriate.
But there are, indeed, cases where they are
not only inappropriate, but dangerously
misleading.
I'm not arguing whether or not NASM is good
code.
But I am arguing that use of McCabe's metric
is probably inappropriate in this case.
No, they are actually doing the opposite from
that: instead of multiple bits/symbol, they
are using one bit/symbol.
The bit is initially encoded by a change in
duty cycle of a square(-ish) wave.
They then modulate a carrier with this
signal, remove the carrier and a sideband,
finally filtering the resulting sideband with
extremely sharp, patented-technology filters.
I'm real suspicious...
the initial square wave would have to be
several MHz, and it seems that the subsequent
processing would either wind up stripping
out the information or result in a multi-MHz
bandwidth in the resulting signal.
I could be all wrong, but until I saw the math
that shows exactly how the bits are encoded and
extracted, I'd be skeptical.
It was hardly evident at the time that X would be
the dominant non-MS windowing system.
Sun itself had SunTools/SunView followed by
its own
Display PostScript-like system, NeWS.
The latter predates NeXT's use of DPS by a
couple of years.
It flopped; Sun ultimately abandoned it,
after bleeding much effort and money.
And I have to disagree on the hardware angle.
NeXT should have dropped the proprietary
hardware angle long before it did; as it
stands Jobs' insistance on an all-proprietary
platform
bled the company to the brink of bankruptcy,
sacrificing efforts on the OS at a time when
Microsoft
had yet to obtain its total lock on the OS
market.
NeXTStep could have been a Windows-killer
if the company had put all its effort into
improving and marketing it (and providing
more assistance to developers).
We'll never know.
Jobs didn't see that the window of opportunity
for an all-proprietary platform that Apple
had entered was closed -- closed
by the accelerating
market in commodity PC's on the low end, and
a multitude of established Unix workstation
vendors on the high end.
The "turbo packet support" feature has just
appeared in test versions of pre-2.4 Linux
in the past week or two.
That's hardly "always had it," as our troll
claims (as if a system that isn't
even released yet could "have" anything).
So it's no further along than the FreeBSD
equivalent, and despite the claims of our
troll, is just as likely to be a copy of
the BSD feature as vice-versa (in other words,
unlikely).
But of course, the reality is that some other OSes
have had zero-copy for quite a long time.
So who is copying whom?
(You'll have to search the "linux-net"
list, not the "linux-kernel" list, for a
discussion of "turbo packet support" and
not "turbo sockets."
Not only is our troll such a Linux ignoramous
that he doesn't know where to look, he doesn't
even know what to call it.)
It's always a sign of an immature wanna-be
when such bare claims of BSD or Linux
superiority are made; it's pretty obvious
why our troll has decided to remain anonymous.
The fact is that Linux and BSD each
help make the other
better, with friendly competition among the
actual developers and a free flow of information
between them.
With rare exceptions, the attitude between
workers in the two camps is one of mutual
respect and even occasionally admiration.
(Far more dissing goes on within
the groups themselves than between them -- it's
the camp followers who make all the us-vs-them
noises.)
The Linux vs. BSD lamers simply don't
understand what free software development is
all about; unlike commercial software,
win/win situations are the rule, and not
the exception.
Well, I don't agree.
Remembering someone's life upon occasion of
their death is hardly "wallowing."
If it seems that way to you then I have to
wonder just how much meaning life has for you.
When someone famous dies, it is an opportunity
for us to reflect on our own accomplishments
and mortality.
From such reflection derives wisdom.
If you don't know or care enough about
Alec Guinness'
life to reflect on where it touches you own,
that's fine.
Just show a bit of respect for those of us who do.
Sir Alec earned his fame, and you might find
it worthwhile someday to investigate how and
why.
Until then, leave those who wish to celebrate
and memorialize his life and its connection
to our own lives free to do so without
your churlish comments.
Documenting the exploit isn't the same as posting ready-to-build-and-run source code. Describe the exploit on a packet-by-packet level, and forward it to Cisco, CERT, and other appropriate security fora.
Once you've given good, solid information that anyone with good knowledge of TCP/IP can understand and verify with a moderate amount of effort, you've done your part. It isn't your job to "prove" anything; if you get told "put-up-or-shut-up" then just move on.
Yup, though there likely was a reason why he did it: all the other system calls at that time were five letters or less. Rather than make "creat" break the rule (and perhaps mess up those nice neat columns of assembler code -- you did read enough of the article to know that Unix was originally written in assembler, hmmm?) he dropped off the "e".
Typing on those old ASR-33's got to be painful after a while; you had to press pretty hard on those keys and then klunk! it would press in and bottom out. There is a reason all those ancient Unix commands were so short!
Good grief, all you did was repeat half of what I wrote, but with five times the words. You write as if I were trying to defend CLI's. I'm not. They are an arcane and essentially primitive use of human language abilities. My point was that GUI's aren't any further advanced in the use of the human capacity for visual expression than CLI's are of the human capacity for verbal expression. Their advantage is that they allow for recognition to replace recall, and in a relatively compact way. That's a huge advance in "usability" over a CLI. It's a damn shame it's pretty much the only one.
Your example shows the power of narrative and recognition (e.g. landmarks like the Wal-Mart), not the power of visualization. An example of the latter would be a simple map. And that's what I want in a UI.
Why can't I draw a simple map for my computer and say "do this?" Instead, I'm forced to click on this, then this, then this. If I want to do the same thing to a different set of objects, I have to go through the same series of clicks. Do GUI's let me draw a series of steps and then apply them to the objects I choose? No, they force me to script -- in other words, back to the g*d d*mn CLI again!
The first fact is that the majority of people are spatially-oriented
No; most people are verbally-oriented. The ability to visualize and juggle spatial concepts is less common than the ability to describe, narrate, and request things. The special skill of the mathematician is one of expressing in symbols things that ordinarily can only be visualized. In other words, mathematicians (and good programmers) have the ability to formulate and utilize abstractions.
Nonetheless, there is a lot that could be done to make GUI's more useful. Take a look at how limited most GUI's are in terms of their use of visual metaphor: click, drag, drop, click. Things that wouldn't tax a 3-year-old. You would communicate with a real visual interface by drawing, by gesturing, by manipulating a visual representation in much more interesting ways than merely clicking and selecting.
No, today's GUI's are just more compact versions of the menu interfaces of yore, with nice pictures instead of words but with little else that utilizes visual intelligence. We use them because we don't know any better. We use them because, like earlier text menus, they take advantage of the fact that the human brain is much better at recognition than blind recall. We use them even though they make little use of the visual realm.
MS Windows and MacOS have done the same thing to the GUI world that Unix did to the CLI world: they are "good enough" that no one has managed to step back and start from scratch, questioning the assumptions that were made 30-odd years ago at Xerox (for GUI's) and Bell Labs (for CLI's). It would be a monumental effort at this point, given the enormous inertia and mental stenosis that affects the industry. We should salute anyone with enough guts to try and take UI's beyond the mental world of the 3-year-old.
The difference between reselling that Metallica CD and sharing it with a few thousand of your closest net-friends is that in the former case there still is only the one CD, while in the latter case thousands of copies have been made and distributed. These are two very different situations from the perspective of those who assert rights to the recording (and profit from those rights). That's not to say that there is anything "natural" about a one-medium, one-copy rule. But that is pretty much the condition under which the current system of music distribution has evolved, for good or for ill.
The problem with a purely free-for-the-downloading approach to music is that there is no money in it for those who make the recordings. Recording music would be pretty much a hobby activity, or at best one subsidized by musical performances. And making good recordings isn't cheap, with $1500 microphones, $40,000 mixing consoles, and so on. True, digital technology is going to reduce some of these costs a bit, through hard-disk recording and the like, and some forms of music (computer- based electronica, for instance) are especially cheap to produce this way. But recording live performances is always going to be labor- and gear-intensive, with at least a certain amount of expensive equipment and personnel required to produce a quality result. All-free, all-the-time isn't good for music or musicians, or ultimately, music-listeners. (And those luddites who feel that such a situation will benefit live music enough so as to be a positive trade-off -- save your arguments for later.)
So the question boils down to this: how can music recording make enough money to be self-sustaining when anyone can make copies at will, for free? The fact is, it can't. And unlike free software, where services to, with, and about a free product provide ample economic opportunities, a recording musician can't make much money with a free recording.
The RIAA members may grossly overcharge for their services, but at least sometimes musicians get recorded and paid. The 'net provides an alternative to what the RIAA does in manufacturing and distributing recorded media, and at much less cost; they want to be able to continue to make fat profits on a system that is rapidly becoming antiquated until they find a way to fill their money pipeline via electronic distribution. You and I don't expect, or want, that to happen. But the fact remains that free distribution won't pay for artists and it won't pay for recordings to be made. What will replace the RIAA members, and how will it make its money? Answer the latter question, and the former will take care of itself.
I personally don't think that having individuals rent music is a particularly viable solution to financing recordings -- I just don't see any enabling technology on the horizon that the average music listener would accept. And that's what this discussion boils down to. But I could be wrong: this could be a very fruitful discussion. It's not going to go anywhere if we keep focused on RIAA vs. Napster, however -- I don't see either of them as representing viable solutions.
And the leased lines it used were damned expensive, too; the cheaper leased lines used by some leaf nodes were actually as slow as 8Kbps. This was true as little as 18 years ago, back when TCP/IP was just being invented. (Betcha didn't know that the ARPANET didn't always use TCP/IP.) So it's silly to say that TCP/IP wasn't designed for such low datarates -- at the time there wasn't much that was faster.
For instance, they often times have to hire, promote and train often times based upon color or sex rather than skills.
"Affirmative action" only affects companies who contract with the government, and is hardly universal even then. It's usually been fairly toothless, with a few notable exceptions. You may be referring to antidiscrimination laws that affect almost all companies above a certain size, but these merely prohibit hiring/training/promoting based on specific things other than skills and experience -- a different thing entirely than what you imply.
They have to pay the lionshare of taxes in the US - [which they pass on to us no doubt]
Large companies actually pay proportionately less in taxes than small companies and individuals, leading to such absurd situations as one of the largest and most profitable companies in the US, General Electric, paying no income taxes at all.
They have to comply with often time, arcane rules for ergonomics and safety set for them by a department with nothing better to do. [Nothing funnier than seeing the local Bell rep in a hard hat in our equipment room. Look out for those falling bits!]
I'd agree with you if service reps only had to work in the semi-office environment of the typical machine room. But data and telephone infrastructure is hardly limited to environments such as yours; they can be every bit as dangerous as construction zones -- and frequently they are construction zones.
As for the larger issue, until you've worked on a line and seen the sorts of things employers try to get away with in terms of safety and working conditions, you've no ground on which to criticise the bulk of what OSHA does. Visit a data-entry facility if you want to see just why some rules might be required for keyboard workers -- they are veritable sweatshops with computers. It may seem a little silly when they get applied to software developers -- but then again, having known several with work-related repetitive-stress injury, perhaps the rules aren't quite so "arcane" for us either.
They have to, without question, deal with and support unions.
You're pretty young, aren't you? Union clout these days is miniscule compared to what it was twenty or thirty years ago, with about a third the members today as then and whole industries de-unionized. Unions currently are weaker in the US than in any other industrialized nation, without doubt.
It takes an act of Congress, literally, for them to do business outside of our borders.
There's this thing called the US Constitution that requires Congress to do this. But I've not noticed this little fact putting much of a crimp in the sails of multinational corporations lately, have you?
In brief, I don't know what broadsheet you copied the above "facts" from, but they show a pretty short-sighted view of the situation. Study the history of US industry from the late 1800's to the start of the depression and you'll see an eerie picture of what happens when corporations increasingly rule the US. Perhaps you'll gain a little appreciation for why the government is involved in the affairs of corporations today.
Actually, Microsoft doesn't need to have specs at all. It just needs to carry along a bunch of legacy code that gets glued into successive versions, perhaps with some API modifications or a compatability layer. "Conformance" to such a non-spec can be determined by regression testing.
A surprising number of software projects cook along for many years and through many revisions without ever having complete specs. And though the lack of specs may be bad, code re-use is usually a Good Thing, specs or no.
As for me? Yes I have started to buy MORE cd's because of napster. but then, I have stopped buying mainstream crap. I've been buying non-label artists' CD.
This, in my experience, is a genuine trend, and is likely one of the unspoken reasons the RIAA folks have their knickers in a twist. Even if piracy ultimately promotes sales, it screws up the big label's marketing machines since it lets people expose themselves to a wider variety of music. Think about it: the fewer titles a label needs to produce and market to sell a given volume of CDs, the lower its costs for studio, production setup, unsold units, and so on. But if people buy a greater variety of music (because they are exposed to that variety on-line, legally or illegally), their costs go up, and thus their profits go down.
This is, I think, a major reason why the recording industry appears so clueless in dealing with the on-line world. The natural progression would be in the direction of greater choice--and they don't want you to have any more choices than they want you to. Of course, the major labels' practices also mean that most artists wind up on the sidelines, except for their chosen few. They want their audiences to fall in large, well-defined (and well-controlled) groups.
Slashdot Mirror
I don't agree with RedHat's decision to ship what was essentially a snapshot, in RedHat's defense I have to say that they were faced with a dilemma. They could either:
They chose the last option, knowing that there was no possibility of having 3.0 compatibility aside from option #2, and that they'd at least get a stable and largely standards-compliant C++ compiler.
After RedHat chose a snapshot, they continued to follow subsequent snapshots but because of the necessities of release engineering created patches against the original snapshot, which has lead to the accusation (largely unwarranted) that they have forked GCC.
Personally, I think they should have stuck with 2.95.2, warts and all, but it was a judgement call for them, and the path they took is not without some justification.
(BTW, the FreeBSD folks are so disgusted with 2.95.2 that they're considering making a similar move.)
This is a hardware, not an OS issue. And sure enough, Alpha (like Pentia) has an atomic test-and-set instruction! Good thing, too -- you need this sort of instruction to implement mutexes and other locks both in kernel and user space. There is no efficient software workaround for the lack of such an instruction.
Cool; perhaps you did. Nobody I knew did, but I'm hardly omniscient. However, what it was called has little to do with my original point: the "Internet" of the day (whatever it was called) isn't what we call the "Internet" today.
Perhaps a bit of residue left over from the changeover of the early '90's will convince you. Network 10 (addresses where the first octet is "10") are not routed over the Internet. Addresses in this range are used for internal networks, for testing, and what-not. Well, guess what the first octet of all ARPANET addresses was? Yup, it was 10.
Back when I first started working for them, RAND's address on the ARPANET was 10.0.0.7 (and not coincidentally, they were about the seventh node to join the ARPANET). At the time of the MILNET split, RAND received the new address 192.5.14.* (along with a few other class C's) which it has today; needless to say, the old address, like the ARPANET itself, no longer exists.
My original point remains: several acts of Congress opened up and funded the Internet as we know it today, legislation that came out of Gore's office. You can love him or hate him for other reasons, but he at least deserves credit for that.
The terms "internet" and "internetworking" have existed from the beginning as names for the technology. But until the MILNET split, no one called the network which used those technologies "The Internet." It was "The ARPANET."
Well, the most technically astute president in the last several decades (and perhaps ever) was Jimmy Carter, who was a nuclear engineer. He was also the last man who had a credible claim to being a "Washington outsider," and one of the most honest men who has ever occupied the White House. And he was a mediocre president, at best.
Thus, I'm not necessarily faulting Gore -- he at least knew who to listen to, whether he himself was technically savvy or not. So if I were voting just on the basis of who would most likely be net-friendly (and not just follow corporate interests and the wishes of the national security apparatus), I'd give him the nod. But it's real hard to get much enthusiasm over either of the major candidates.
In the late '80s the military insisted on cracking down on "unofficial use." Many schools and contractors wound up having to justify their access, and came up short. Efforts to start member-supported networks, such as CSNET, got a lukewarm reception since they (1) cost too much (no more 100% government subsidy) and (2) did too little (email and limited file transfer only). It took an act of congress to get these folks hooked up again: the ARPANET was split into the MILNET and NSFNET (NSF == National Science Foundation), with the latter offering subsidized access to any academic institution that wanted it and paid its share. And it was after this split that users starting refering to it as just "The Internet" (though from time to time I heard ARPANET referred to as "The ARPA Internet" back in the '80s).
Guess who wrote the bill that funded NSFNET? Or the National Supercomputing Initiative which built its backbone? Or the one somehwat later that lifted restrictions on commercial use? Yup, Smilin' Al (or, rather, one of his staffers wrote the bill and he sponsored it).
When you look at it from a technical perspective, saying that he "created" the Internet is a clear absurdity. But from a legislative and public policy perspective, it's hardly an exaggeration. I'm not convinced that Al "gets" the Internet any more than W does. But he did manage to listen to someone on his staff who "got" it. (And I've been racking my brain to remember the guy's name--I met him briefly while I was working at The RAND Corporation.) It's too bad that so few technophiles go to work as congressional staffers; it leaves me with some doubts as to whether Gore wasn't just "lucky" to be involved with the Internet.
Wow; after 200 posts on the subject yesterday and today, this is the first post I've seen mentioning this simple fact. I'd hope that /.'ers would be a little more
statistics-savvy (they usually are).
I understand that, but he probably signed away the right to make that choice. My point is that he didn't need to. Perhaps he could have agreed not to assign inventions to anyone else or to patent them himself, thus reserving the right to make the choice he now wants to make. But that's not what he agreed to, and now he wants to take back that agreement. That's usually not possible...
They can't make him, but they probably can sue him! Although there are limits on how far employment contracts can reach beyond employment, they don't become null and void just because you end employment.
I'm no lawyer, but I've consulted with a lawyer every time I've signed an employment contract, sometimes negotiating changes in clauses I found offensive. Given his beliefs, I suppect this fellow should have done this, too.
Those pieces of paper people sign when they become employees aren't just for show. I'm wondering why this fellow signed an agreement to assign inventions to the company if he didn't plan to keep it. He should have asked to have the offending language removed or changed. Granted, there is a good chance he'd be shown the door. Or he might have had to make concessions on compensation or in other areas. It depends upon the company, and how much his services are worth to the company. But taking a moral stand cannot be an afterthought.
McCabe's Cyclomatic Complexity can be misleading, especially for assemblers, simulators, and other software which must represent a model of hardware. The same is true for compilers, interpreters, emulators, and other such software.
In the case of an assembler, dividing opcode dispatch or address mode decoding across dozens of metric <10 functions can make the correctness and completeness of the assembler harder, not easier, to determine. (However, it is appropriate that a thorough knowledge of the architecture of the machine the assembler is for be required for such understanding and determination.) It's the nature of the beast, though it is possible to build or obtain tools that allow for a more direct representation of the target hardware (e.g. GAS uses such a method).
Complexity metrics have the flaw that they encourage (or in some shops, enforce) the diffusion of complex but highly regular logic which is better represented as a whole. It may be impossible to rate the actual comprehensibility and maintainability of a module based on the number of edges in its control flow, as McCabe's metric attempts to do. The predicates which generate those edges may model a regular structure such that drawing them together into a single module makes it easier to understand, even though its McCabe metric is high. Conversely, a control flow with a few edges generated by predicates with obscure interactions and dependencies, such as would be required by the division of certain problems into submodules, might yield low McCabe metrics and be nearly impossible to understand. Inappropriately applied, metrics like McCabe's can encourage the generation of artifacts based not on the inherent structure of the system being modeled, but on the satisfaction of blind rules.
In some application areas (business logic, for instance) metric like McCabe's work well. They can be useful even with parts of systems where they are otherwise inappropriate. But there are, indeed, cases where they are not only inappropriate, but dangerously misleading.
I'm not arguing whether or not NASM is good code. But I am arguing that use of McCabe's metric is probably inappropriate in this case.
No, they are actually doing the opposite from that: instead of multiple bits/symbol, they are using one bit/symbol. The bit is initially encoded by a change in duty cycle of a square(-ish) wave. They then modulate a carrier with this signal, remove the carrier and a sideband, finally filtering the resulting sideband with extremely sharp, patented-technology filters.
I'm real suspicious... the initial square wave would have to be several MHz, and it seems that the subsequent processing would either wind up stripping out the information or result in a multi-MHz bandwidth in the resulting signal. I could be all wrong, but until I saw the math that shows exactly how the bits are encoded and extracted, I'd be skeptical.
It was hardly evident at the time that X would be the dominant non-MS windowing system. Sun itself had SunTools/SunView followed by its own Display PostScript-like system, NeWS. The latter predates NeXT's use of DPS by a couple of years. It flopped; Sun ultimately abandoned it, after bleeding much effort and money.
And I have to disagree on the hardware angle. NeXT should have dropped the proprietary hardware angle long before it did; as it stands Jobs' insistance on an all-proprietary platform bled the company to the brink of bankruptcy, sacrificing efforts on the OS at a time when Microsoft had yet to obtain its total lock on the OS market. NeXTStep could have been a Windows-killer if the company had put all its effort into improving and marketing it (and providing more assistance to developers).
We'll never know. Jobs didn't see that the window of opportunity for an all-proprietary platform that Apple had entered was closed -- closed by the accelerating market in commodity PC's on the low end, and a multitude of established Unix workstation vendors on the high end.
The "turbo packet support" feature has just appeared in test versions of pre-2.4 Linux in the past week or two. That's hardly "always had it," as our troll claims (as if a system that isn't even released yet could "have" anything). So it's no further along than the FreeBSD equivalent, and despite the claims of our troll, is just as likely to be a copy of the BSD feature as vice-versa (in other words, unlikely). But of course, the reality is that some other OSes have had zero-copy for quite a long time. So who is copying whom?
(You'll have to search the "linux-net" list, not the "linux-kernel" list, for a discussion of "turbo packet support" and not "turbo sockets." Not only is our troll such a Linux ignoramous that he doesn't know where to look, he doesn't even know what to call it.)
It's always a sign of an immature wanna-be when such bare claims of BSD or Linux superiority are made; it's pretty obvious why our troll has decided to remain anonymous. The fact is that Linux and BSD each help make the other better, with friendly competition among the actual developers and a free flow of information between them. With rare exceptions, the attitude between workers in the two camps is one of mutual respect and even occasionally admiration. (Far more dissing goes on within the groups themselves than between them -- it's the camp followers who make all the us-vs-them noises.)
The Linux vs. BSD lamers simply don't understand what free software development is all about; unlike commercial software, win/win situations are the rule, and not the exception.
Well, I don't agree. Remembering someone's life upon occasion of their death is hardly "wallowing." If it seems that way to you then I have to wonder just how much meaning life has for you.
When someone famous dies, it is an opportunity for us to reflect on our own accomplishments and mortality. From such reflection derives wisdom. If you don't know or care enough about Alec Guinness' life to reflect on where it touches you own, that's fine. Just show a bit of respect for those of us who do.
Sir Alec earned his fame, and you might find it worthwhile someday to investigate how and why. Until then, leave those who wish to celebrate and memorialize his life and its connection to our own lives free to do so without your churlish comments.
Documenting the exploit isn't the same as posting ready-to-build-and-run source code. Describe the exploit on a packet-by-packet level, and forward it to Cisco, CERT, and other appropriate security fora.
Once you've given good, solid information that anyone with good knowledge of TCP/IP can understand and verify with a moderate amount of effort, you've done your part. It isn't your job to "prove" anything; if you get told "put-up-or-shut-up" then just move on.
Yup, though there likely was a reason why he did it: all the other system calls at that time were five letters or less. Rather than make "creat" break the rule (and perhaps mess up those nice neat columns of assembler code -- you did read enough of the article to know that Unix was originally written in assembler, hmmm?) he dropped off the "e".
Typing on those old ASR-33's got to be painful after a while; you had to press pretty hard on those keys and then klunk! it would press in and bottom out. There is a reason all those ancient Unix commands were so short!
Good grief, all you did was repeat half of what I wrote, but with five times the words. You write as if I were trying to defend CLI's. I'm not. They are an arcane and essentially primitive use of human language abilities. My point was that GUI's aren't any further advanced in the use of the human capacity for visual expression than CLI's are of the human capacity for verbal expression. Their advantage is that they allow for recognition to replace recall, and in a relatively compact way. That's a huge advance in "usability" over a CLI. It's a damn shame it's pretty much the only one.
Your example shows the power of narrative and recognition (e.g. landmarks like the Wal-Mart), not the power of visualization. An example of the latter would be a simple map. And that's what I want in a UI.
Why can't I draw a simple map for my computer and say "do this?" Instead, I'm forced to click on this, then this, then this. If I want to do the same thing to a different set of objects, I have to go through the same series of clicks. Do GUI's let me draw a series of steps and then apply them to the objects I choose? No, they force me to script -- in other words, back to the g*d d*mn CLI again!
That's primitive!
No; most people are verbally-oriented. The ability to visualize and juggle spatial concepts is less common than the ability to describe, narrate, and request things. The special skill of the mathematician is one of expressing in symbols things that ordinarily can only be visualized. In other words, mathematicians (and good programmers) have the ability to formulate and utilize abstractions.
Nonetheless, there is a lot that could be done to make GUI's more useful. Take a look at how limited most GUI's are in terms of their use of visual metaphor: click, drag, drop, click. Things that wouldn't tax a 3-year-old. You would communicate with a real visual interface by drawing, by gesturing, by manipulating a visual representation in much more interesting ways than merely clicking and selecting.
No, today's GUI's are just more compact versions of the menu interfaces of yore, with nice pictures instead of words but with little else that utilizes visual intelligence. We use them because we don't know any better. We use them because, like earlier text menus, they take advantage of the fact that the human brain is much better at recognition than blind recall. We use them even though they make little use of the visual realm.
MS Windows and MacOS have done the same thing to the GUI world that Unix did to the CLI world: they are "good enough" that no one has managed to step back and start from scratch, questioning the assumptions that were made 30-odd years ago at Xerox (for GUI's) and Bell Labs (for CLI's). It would be a monumental effort at this point, given the enormous inertia and mental stenosis that affects the industry. We should salute anyone with enough guts to try and take UI's beyond the mental world of the 3-year-old.
The difference between reselling that Metallica CD and sharing it with a few thousand of your closest net-friends is that in the former case there still is only the one CD, while in the latter case thousands of copies have been made and distributed. These are two very different situations from the perspective of those who assert rights to the recording (and profit from those rights). That's not to say that there is anything "natural" about a one-medium, one-copy rule. But that is pretty much the condition under which the current system of music distribution has evolved, for good or for ill.
The problem with a purely free-for-the-downloading approach to music is that there is no money in it for those who make the recordings. Recording music would be pretty much a hobby activity, or at best one subsidized by musical performances. And making good recordings isn't cheap, with $1500 microphones, $40,000 mixing consoles, and so on. True, digital technology is going to reduce some of these costs a bit, through hard-disk recording and the like, and some forms of music (computer- based electronica, for instance) are especially cheap to produce this way. But recording live performances is always going to be labor- and gear-intensive, with at least a certain amount of expensive equipment and personnel required to produce a quality result. All-free, all-the-time isn't good for music or musicians, or ultimately, music-listeners. (And those luddites who feel that such a situation will benefit live music enough so as to be a positive trade-off -- save your arguments for later.)
So the question boils down to this: how can music recording make enough money to be self-sustaining when anyone can make copies at will, for free? The fact is, it can't. And unlike free software, where services to, with, and about a free product provide ample economic opportunities, a recording musician can't make much money with a free recording.
The RIAA members may grossly overcharge for their services, but at least sometimes musicians get recorded and paid. The 'net provides an alternative to what the RIAA does in manufacturing and distributing recorded media, and at much less cost; they want to be able to continue to make fat profits on a system that is rapidly becoming antiquated until they find a way to fill their money pipeline via electronic distribution. You and I don't expect, or want, that to happen. But the fact remains that free distribution won't pay for artists and it won't pay for recordings to be made. What will replace the RIAA members, and how will it make its money? Answer the latter question, and the former will take care of itself.
I personally don't think that having individuals rent music is a particularly viable solution to financing recordings -- I just don't see any enabling technology on the horizon that the average music listener would accept. And that's what this discussion boils down to. But I could be wrong: this could be a very fruitful discussion. It's not going to go anywhere if we keep focused on RIAA vs. Napster, however -- I don't see either of them as representing viable solutions.
And the leased lines it used were damned expensive, too; the cheaper leased lines used by some leaf nodes were actually as slow as 8Kbps. This was true as little as 18 years ago, back when TCP/IP was just being invented. (Betcha didn't know that the ARPANET didn't always use TCP/IP.) So it's silly to say that TCP/IP wasn't designed for such low datarates -- at the time there wasn't much that was faster.
"Affirmative action" only affects companies who contract with the government, and is hardly universal even then. It's usually been fairly toothless, with a few notable exceptions. You may be referring to antidiscrimination laws that affect almost all companies above a certain size, but these merely prohibit hiring/training/promoting based on specific things other than skills and experience -- a different thing entirely than what you imply.
Large companies actually pay proportionately less in taxes than small companies and individuals, leading to such absurd situations as one of the largest and most profitable companies in the US, General Electric, paying no income taxes at all.
I'd agree with you if service reps only had to work in the semi-office environment of the typical machine room. But data and telephone infrastructure is hardly limited to environments such as yours; they can be every bit as dangerous as construction zones -- and frequently they are construction zones.
As for the larger issue, until you've worked on a line and seen the sorts of things employers try to get away with in terms of safety and working conditions, you've no ground on which to criticise the bulk of what OSHA does. Visit a data-entry facility if you want to see just why some rules might be required for keyboard workers -- they are veritable sweatshops with computers. It may seem a little silly when they get applied to software developers -- but then again, having known several with work-related repetitive-stress injury, perhaps the rules aren't quite so "arcane" for us either.
You're pretty young, aren't you? Union clout these days is miniscule compared to what it was twenty or thirty years ago, with about a third the members today as then and whole industries de-unionized. Unions currently are weaker in the US than in any other industrialized nation, without doubt.
There's this thing called the US Constitution that requires Congress to do this. But I've not noticed this little fact putting much of a crimp in the sails of multinational corporations lately, have you?
In brief, I don't know what broadsheet you copied the above "facts" from, but they show a pretty short-sighted view of the situation. Study the history of US industry from the late 1800's to the start of the depression and you'll see an eerie picture of what happens when corporations increasingly rule the US. Perhaps you'll gain a little appreciation for why the government is involved in the affairs of corporations today.
Actually, Peter Wemm is core; he just happens to work at Yahoo!. Jayanth's last name is "Vijayaraghavan." And I was there for the first day.
Actually, Microsoft doesn't need to have specs at all. It just needs to carry along a bunch of legacy code that gets glued into successive versions, perhaps with some API modifications or a compatability layer. "Conformance" to such a non-spec can be determined by regression testing.
A surprising number of software projects cook along for many years and through many revisions without ever having complete specs. And though the lack of specs may be bad, code re-use is usually a Good Thing, specs or no.
I agree; any marketing droid will tell you "Ogg" is a horrible name. Of course, they'd have told you the exact same thing about "Yahoo!"...
This, in my experience, is a genuine trend, and is likely one of the unspoken reasons the RIAA folks have their knickers in a twist. Even if piracy ultimately promotes sales, it screws up the big label's marketing machines since it lets people expose themselves to a wider variety of music. Think about it: the fewer titles a label needs to produce and market to sell a given volume of CDs, the lower its costs for studio, production setup, unsold units, and so on. But if people buy a greater variety of music (because they are exposed to that variety on-line, legally or illegally), their costs go up, and thus their profits go down.
This is, I think, a major reason why the recording industry appears so clueless in dealing with the on-line world. The natural progression would be in the direction of greater choice--and they don't want you to have any more choices than they want you to. Of course, the major labels' practices also mean that most artists wind up on the sidelines, except for their chosen few. They want their audiences to fall in large, well-defined (and well-controlled) groups.