Janitors at PARC have just announced electronic toilet paper. The revelation of the new electronic TP has shocked the executive staff, on whom the janitorial staff experimented by stocking the executive washroom with prototype rolls of their new invention. A spokesperson for the group revealed that they have now begun working on a digital product that could be deeply embedded, and hope to be through the experimentation stage soon.
> Back in the days of three- to five-year product > cycles
There are still a lot of markets where the life cycle is three to five years. In these markets, copying a competitors product is a viable option. Proprietary software is a barrier to entry that keeps the number of players in the market small.
> If you include MCA and OS/2, IBM probably has a > net loss for it's microcomputer endevors
Get real. IBM raked in the dough for years off the PC, XT and AT. OS/2 was a drop in the bucket. I doubt they lost money on MCA. It was quite popular for a while (it had well over half the high end even after EISA had been out for a year or so).
> [they] just stays in the business so that they > can provide an end-to-end solution.
Yep. IBM PC software was always a loser. Remember PageWriter? First Wordstar beat it, then Word Perfect. They finally put it to bed before M$ Office walked all over everything. They only kept it so long as it was strategicly important to have a "one-stop solution".
> Linux gives IBM a way to... avoiding... the > dismal future of becoming a second-tier > Microsoft solution provider.
IBM has forgotten more about providing solutions that Microsoft will ever know. The reason they're investing in Linux is to distract Microsoft from the real prize, the business enterprise. While Microsoft wastes effort combatting the threat from the low end, IBM gets freer run of the more profitable high end.
Wasn't this one of the aliens on the old classic Star Trek series? One of the Klingons compares Captain Kirk to a Debian slime devil, right before the big bar brawl scene in the Trouble with Tribbles episode.
Some history to support these arguements: 1. Microsoft has publicly stated that they intend to own 100% of the software market. 2. According to someone I worked for, Microsoft sent him an NDA requiring that he give them rights to all intellectual property of his company. 3. Microsoft got 3Com to develop LAN manager on OS/2 to compete with Novell. 3Com took a bath on it, and Microsoft turned around and bundled it as NT Server. 4. Microsoft was reselling DCA's Com Manager SNA gateway product on OS/2. When NT came out, they rolled out their own binary compatible SNA Server product.
Component GUI/Monolithic X
on
The KDE Future
·
· Score: 1
Why are the desktop environments so bloated? Because they have to staticly embed everything. That's why they're adding Corba, so not every applet need to embed a full file browser. It just pops up the file browser applet in place. This *is* the UNIX way of doing things, at the GUI level.
IMNSHO, its X that's the bloated pig - since I only have one machine, why do I need all that networking code in there? Not to mention the code to translate between endianesses, etc, etc.
> An example is the "static branch prediction". > The compiler will evaluate a comparison for a > branch that is always, or never, taken. It'll > pass that hint along to the processor to help > avoid a misprediction penalty. (Of course, > "passing a hint to the processor" means > embedding the likely answer in the code.)
Though the white paper isn't clear on this, they aren't talking about static branch prediction as a chip feature. The improvement comes when the compiler can't tell which path will be taken.
In this case, on a low-end PowerPC, the compiler will guess. For example, for a loop, it always predicts the loop will be taken. For the high end PowerPC, its hint is used as the initial prediction (weakly predict taken or not taken, see below).
The high-end PowerPC does dynamic branch prediction. It has a LRU cache of 2 bit values for the last N branches taken. 00 means strongly predict not taken, 01 is weakly predict not taken, 10 is weakly taken, and 11 is strongly taken. Each time a branch in the cache is hit, its value is incremented if actually taken, decremented if not.
In both of these cases, branch prediction determines the path that will be speculatively executed. I understood what made the IA-64 special was that it executes *both* paths speculatively and then throws away the changes from the path not taken. In other words, it doesn't do branch prediction at all.
> People go around saying things like 'GPL means > you can't link with non-GPL stuff' Now IANAL > but I don't think it does, and the confusion is > not good.
It depends on the non-GPL stuff, but generally speaking, this is true. You're required to link the derived work under the GPL. If the non-GPL stuff is under another license, it will typically prevent you from doing this. Fortunately, you can use the LGPL if you want to allow linking to non-GPL source.
> There appears to be a GPL3 in the works and I > would expect that it will address some of the > issues that RMS complains about. So maybe it > will restrict plug-ins. Is this good? Do I want > RMS to be deciding how my project is > controlled? (Maybe I'll be forced to call it > GNU/Project!:)
Nothing says you have to upgrade to GPL3. Someone else can take your code and GPL3 it, but you can still continue to use GPL2. If they add an enhancement and you want it, you would have to go to GPL3 to take their version.
> I'm not sure I want to be involved with a > project that simultaneously criticizes the BSD > self advertisement clause, and engages in > obvious self advertising with the GNU/Linux > debate.
You are free to use the GPL without otherwise having anything to do with the GNU project.
I noticed a few posts suggesting Linux should do C2. I certainly hope not! C2 is one of the reasons that 16 bit windows code runs so slowly on NT. This is because one of the requirements of C2 is that no process be able to access the memory of any other.
This means that DLLs which share memory between multiple processes are not allowed. Everything you want to communicate to a system service must be sent through a message queue. Thus, a C2 system can guarantee no-one can exploit data from core files to break security.
Lack of C2 and microkernel architecture are, IMO, one of Linux's key strengths. C2 is a feature bullet that everyone pays the price for. Like one of the other posters said, its like ISO - you have it just to say you have it. Don't bloat my OS with it.
If you want to see what people are talking about AFA new kernel development, kernel traffic is a handy summary of discussion on the Linux kernel mailing list which comes out once a week. Its announced right here on/.
This is the first time the "standard" kernel has supported non-intel processors, so for those, its pretty much a no-brainer. For intel, there are devices that are only supported in the 2.2 series. Also, some new software coming out may require 2.2 over 2.0.
This article and alot of the comments assume that software is a finite thing. I beleive its more like Science. The more that you write/know, the more that remains to be written/learned. If you accept this, you discover that the idea of a proprietary operating system puts a limit on software. After all, if your OS is worth $100, you have to expect to pay even more for a full suite of application software. In the eighties, SCO used to charge $500 for Xenix, and $500 for the compiler. Microsoft has reduced that cost somewhat. But if the cost of the OS, development and basic productivity tools were $25, money would be freed up to buy more software. Hopefully, the free software movement will keep raising the bar.
If your going to change the name, you have to keep it shorter than GNU/Linux. Here are some suggestions that use all the letters in GNU and LINUX once.
GLINUX - my personal favourite. LIGNUX - runner up. GNULIX - for people who don't like GNU. GUNLIX - for people who like guns. GNIXUL - pronounced like pixel. GLUNIX - sticky. XIGNUL - German pronounciation of signal. LUGNIX - loonix. GNUXIL - like woozle. NIGULX - niggles.
Unixware has already had one new standard interface for device drivers, the portable driver interface (back when it was AT&T/USL). This seemed to be stillborn, perhaps due to the sale of Unix to Novell. Before that they had DDI/DKI in System V, which is still documented as the standard interface for Solaris. Maybe this one will live as long as the PDI did.
I personally disagree. Those who have released their code under GPL have given a gift to the world. The GPL keeps the code free. Don't look a gift horse in the mouth.
If you want to take advantage of free software, use it in your product, but value add with your own proprietary software. If you're in a vertical market, the open software community shouldn't mind. If you keep your medical, banking, or whatever application code proprietary, but feed back enhancements to the generic portion, you are helping to build the communal base.
This makes sense from a business perspective too. The code which is not a part of your core competence goes back to be maintained by the open source community, and you get to keep the stuff that gives you a business advantage. You can release a product which has a fully functional platform under it without paying royaltees, yet your competitors don't get your business software for free.
The problem with Sun's license is that you have to pay royaltee's to use it. Sun has a right to charge for SW (legal, not moral), but we can choose whether or not to use it. Personally, I'd rather have a GPLed version from GNU. I think RMS has done the world a great service (though I wish he'd get over the Linux name thing).
To those of us running on the SPARC and PPC architectures, this is our only stable version. Then again, I've been running Ultra-Peguin (64-bit/Sun) off the 2.1.X series kernel it came with, for over a month, and it has yet to crash. The only problems have been with the X server. Considering Sun was dragged kicking and screaming to X, the Slowlaris version sure works better.
Slashdot Mirror
Janitors at PARC have just announced electronic toilet paper. The revelation of the new electronic TP has shocked the executive staff, on whom the janitorial staff experimented by stocking the executive washroom with prototype rolls of their new invention. A spokesperson for the group revealed that they have now begun working on a digital product that could be deeply embedded, and hope to be through the experimentation stage soon.
Ethernet is hardly the technology it was 24 years ago. Then again, neither is Unix. For that matter, neither is Windows.
> Back in the days of three- to five-year product
> cycles
There are still a lot of markets where the life cycle is three to five years. In these markets, copying a competitors product is a viable option. Proprietary software is a barrier to entry that keeps the number of players in the market small.
Glad to see you guys finally got this out!
> If you include MCA and OS/2, IBM probably has a
... avoiding ... the
> net loss for it's microcomputer endevors
Get real. IBM raked in the dough for years off the PC, XT and AT. OS/2 was a drop in the bucket. I doubt they lost money on MCA. It was quite popular for a while (it had well over half the high end even after EISA had been out for a year or so).
> [they] just stays in the business so that they
> can provide an end-to-end solution.
Yep. IBM PC software was always a loser. Remember PageWriter? First Wordstar beat it, then Word Perfect. They finally put it to bed before M$ Office walked all over everything. They only kept it so long as it was strategicly important to have a "one-stop solution".
> Linux gives IBM a way to
> dismal future of becoming a second-tier
> Microsoft solution provider.
IBM has forgotten more about providing solutions that Microsoft will ever know. The reason they're investing in Linux is to distract Microsoft from the real prize, the business enterprise. While Microsoft wastes effort combatting the threat from the low end, IBM gets freer run of the more profitable high end.
Cruftmiester
Wasn't this one of the aliens on the old classic
Star Trek series? One of the Klingons compares
Captain Kirk to a Debian slime devil, right before
the big bar brawl scene in the Trouble with Tribbles episode.
Some history to support these arguements:
1. Microsoft has publicly stated that they intend
to own 100% of the software market.
2. According to someone I worked for, Microsoft
sent him an NDA requiring that he give them
rights to all intellectual property of his
company.
3. Microsoft got 3Com to develop LAN manager on
OS/2 to compete with Novell. 3Com took a bath
on it, and Microsoft turned around and bundled
it as NT Server.
4. Microsoft was reselling DCA's Com Manager SNA
gateway product on OS/2. When NT came out,
they rolled out their own binary compatible
SNA Server product.
Why are the desktop environments so bloated? Because they have to staticly embed everything. That's why they're adding Corba, so not every applet need to embed a full file browser. It just pops up the file browser applet in place. This *is* the UNIX way of doing things, at the GUI level.
IMNSHO, its X that's the bloated pig - since I only have one machine, why do I need all that networking code in there? Not to mention the code to translate between endianesses, etc, etc.
> An example is the "static branch prediction".
> The compiler will evaluate a comparison for a
> branch that is always, or never, taken. It'll
> pass that hint along to the processor to help
> avoid a misprediction penalty. (Of course,
> "passing a hint to the processor" means
> embedding the likely answer in the code.)
Though the white paper isn't clear on this, they aren't talking about static branch prediction as a chip feature. The improvement comes when the compiler can't tell which path will be taken.
In this case, on a low-end PowerPC, the compiler will guess. For example, for a loop, it always predicts the loop will be taken. For the high end PowerPC, its hint is used as the initial prediction (weakly predict taken or not taken, see below).
The high-end PowerPC does dynamic branch prediction. It has a LRU cache of 2 bit values for the last N branches taken. 00 means strongly predict not taken, 01 is weakly predict not taken, 10 is weakly taken, and 11 is strongly taken. Each time a branch in the cache is hit, its value is incremented if actually taken, decremented if not.
In both of these cases, branch prediction determines the path that will be speculatively executed. I understood what made the IA-64 special was that it executes *both* paths speculatively and then throws away the changes from the path not taken. In other words, it doesn't do branch prediction at all.
> People go around saying things like 'GPL means
:)
> you can't link with non-GPL stuff' Now IANAL
> but I don't think it does, and the confusion is
> not good.
It depends on the non-GPL stuff, but generally speaking, this is true. You're required to link the derived work under the GPL. If the non-GPL stuff is under another license, it will typically prevent you from doing this. Fortunately, you can use the LGPL if you want to allow linking to non-GPL source.
> There appears to be a GPL3 in the works and I
> would expect that it will address some of the
> issues that RMS complains about. So maybe it
> will restrict plug-ins. Is this good? Do I want
> RMS to be deciding how my project is
> controlled? (Maybe I'll be forced to call it
> GNU/Project!
Nothing says you have to upgrade to GPL3. Someone else can take your code and GPL3 it, but you can still continue to use GPL2. If they add an enhancement and you want it, you would have to go to GPL3 to take their version.
> I'm not sure I want to be involved with a
> project that simultaneously criticizes the BSD
> self advertisement clause, and engages in
> obvious self advertising with the GNU/Linux
> debate.
You are free to use the GPL without otherwise having anything to do with the GNU project.
I noticed a few posts suggesting Linux should do C2. I certainly hope not! C2 is one of the reasons that 16 bit windows code runs so slowly on NT. This is because one of the requirements of C2 is that no process be able to access the memory of any other.
This means that DLLs which share memory between multiple processes are not allowed. Everything you want to communicate to a system service must be sent through a message queue. Thus, a C2 system can guarantee no-one can exploit data from core files to break security.
Lack of C2 and microkernel architecture are, IMO, one of Linux's key strengths. C2 is a feature bullet that everyone pays the price for. Like one of the other posters said, its like ISO - you have it just to say you have it. Don't bloat my OS with it.
If you want to see what people are talking about AFA new kernel development, kernel traffic is a handy summary of discussion on the Linux kernel mailing list which comes out once a week. Its announced right here on /.
This is the first time the "standard" kernel has supported non-intel processors, so for those, its pretty much a no-brainer. For intel, there are devices that are only supported in the 2.2 series. Also, some new software coming out may require 2.2 over 2.0.
This article and alot of the comments assume that software is a finite thing. I beleive its more like Science. The more that you write/know, the more that remains to be written/learned.
If you accept this, you discover that the idea of a proprietary operating system puts a limit on software. After all, if your OS is worth $100, you have to expect to pay even more for a full suite of application software.
In the eighties, SCO used to charge $500 for Xenix, and $500 for the compiler. Microsoft has reduced that cost somewhat. But if the cost of the OS, development and basic productivity tools were $25, money would be freed up to buy more software. Hopefully, the free software movement will keep raising the bar.
If your going to change the name, you have to keep it shorter than GNU/Linux. Here are some suggestions that use all the letters in GNU and LINUX once.
GLINUX - my personal favourite.
LIGNUX - runner up.
GNULIX - for people who don't like GNU.
GUNLIX - for people who like guns.
GNIXUL - pronounced like pixel.
GLUNIX - sticky.
XIGNUL - German pronounciation of signal.
LUGNIX - loonix.
GNUXIL - like woozle.
NIGULX - niggles.
Unixware has already had one new standard interface for device drivers, the portable driver interface (back when it was AT&T/USL). This seemed to be stillborn, perhaps due to the sale of Unix to Novell. Before that they had DDI/DKI in System V, which is still documented as the standard interface for Solaris. Maybe this one will live as long as the PDI did.
I personally disagree. Those who have released their code under GPL have given a gift to the world. The GPL keeps the code free. Don't look a gift horse in the mouth.
If you want to take advantage of free software, use it in your product, but value add with your own proprietary software. If you're in a vertical market, the open software community shouldn't mind. If you keep your medical, banking, or whatever application code proprietary, but feed back enhancements to the generic portion, you are helping to build the communal base.
This makes sense from a business perspective too. The code which is not a part of your core competence goes back to be maintained by the open source community, and you get to keep the stuff that gives you a business advantage. You can release a product which has a fully functional platform under it without paying royaltees, yet your competitors don't get your business software for free.
The problem with Sun's license is that you have to pay royaltee's to use it. Sun has a right to charge for SW (legal, not moral), but we can choose whether or not to use it. Personally, I'd rather have a GPLed version from GNU. I think RMS has done the world a great service (though I wish he'd get over the Linux name thing).
To those of us running on the SPARC and PPC architectures, this is our only stable version. Then again, I've been running Ultra-Peguin (64-bit/Sun) off the 2.1.X series kernel it came with, for over a month, and it has yet to crash. The only problems have been with the X server. Considering Sun was dragged kicking and screaming to X, the Slowlaris version sure works better.
> The proper name for Linux is GNU/Linux.
M$ agreeing with RMS?
> if all software were free, how would they eat?
He well knows free is not in the monetary sense. FUD.
> When pushed, the Free Software Foundation
> defines "free" as a matter of liberty, not of
> price.
Really have to push hard to get this definition out of them - NOT. More FUD.
> Xwindows, which looks amazingly like Windows.
After all, M$ did invent the GUI, didn't they. Oh, I forgot, they stole it from Apple who copied it from Xerox.
> A big practical disadvantage of Linux is that
> there isn't much application software for it.
Desktop application software, maybe.
> But that's because so few people use it.
And yet Slate is wasting oh so valuable space on their site (running on *NIX?) and your time over it.
> Linux has a utility program that helped me
> [repartition] (because it knows you're going to
> want to keep running Windows too).
Oh yes, exactly why.
> On my machine I can claim only a week of
> running without restarting.
Missed blue screen and had to reboot to see one.
> If you hate Microsoft, you can use Linux.
I sure hate this FUD.
> Do all those software developers writing open
> source code for Linux have the incentive to fix
> problems as they arise?
Yes (just read Kernal Traffic). FUD, FUD, FUD.
> To keep old code running?
Something M$ is well known for doing.
> Perhaps the greatest technological feature that
> Windows possesses is that it can handle
> programs as old as the first DOS applications.
Bullsh*t.
> Linux will never do that.
Thankfully, but it will run good Unix SW which predates MS-DOG.
> Some critics say that Linux will fracture into > a dozen different incompatible versions, just
> as Unix did.
Stinking FUD.
> This is the ultimate problem with Open Source
> development: not enough formal engaged testing.
But M$ has definately got this one licked. Yeah right.
Please delete/moderate this inflammatory and offensive post into the oblivion it deserves.