An old story I heard once said that the original Intel spec of the 8088 (?) processor included a spec for 1024 hardware interrupts. IBM, of course, in their wisdom, said that the chip was not for a main frame (or something similar), and instead changed to spec to 8 interrupts. Someplace down the road, in the 286 I think(?), they realised the needed more, and expanded to 16 via a second chip. Which is what we have had to deal with ever since.
How can something so well documented get garbled so soon?
The original 8086 design allowed 256 interrupt vectors, hardware and software combined. The hardware has always supported 256 vectors. Intel reserved the first 32 for future use, such as processor exceptions. IBM ignored that and assigned, for instance, INT 5 (which Intel had earmarked for array bounds exceptions) to the BIOS print screen function. Hardware interrupts got shoehorned in at the 0x10-0x1f range.
Why jam so much stuff in low when there were 256 available? Because the original ROM BASIC used the upper 208 or 224 (memory fails) for short-bytecount subroutine calls. It made for a very simple, compact interpreter. A look at a disassembly of the early ones showed that it was basically machine-translated Z80 code!
If you have a floppy drive on your computer, chances are it is still using the ISA bus... So just because you don't have an ISA "slot", don't think that the bus is gone...
No, it's really gone. The floppy connector is just part of the South Bridge chip, along with the keyboard controller, serial and parallel ports, and other low-bandwidth stuff. It hangs off of the PCI bus. These devices are just hard-coded to ACT like the old ISA equivalents, with their addresses, interrupts, DMA, etc. being hard-addigned to the old standard values.
Need more CPU power or more memory? Hot-plug a module into the Infiniband Switch.
Don't gush. Infiniband, or any cabled serial connection, will never be a memory connection worth having. The reason is latency: what matters with memory isn't how many terabytes you can deliver per hour, but how many picoseconds the processor has to wait for that data that's stalling the pipeline right now. Which is very hard to reduce when the address has to be shipped a bit at a time over a serial link (64 bit times @ 10 Gb/s = 6.4 ns), transported over the cable (~50 ps/cm * ~60 cm = 3.0 ns), memory accessed (technology dependent), serialized (another 64 bits, 6.4 ns), shipped back (another 3.0 ns) and finally it gets to the processor. You've added 18.8 ns over and above any protocol overhead (usually much worse) and that's at 10 Gb/s!
One good backhoe accident, and you suddenly have a bunch of intranets. In theory all of the Tier-One operators peer at multiple points, but in practice they route their own traffic through the same facilities. Likewise, in theory the Tier-Two operators are multiconnected through multiple Tier-One providers but in practice thanks to volume contract terms they are single-homed.
Below Tier Two, it really doesn't matter.
Beats me as to who got gobbled up, but I'll bet dollars to donuts that the buyer was EarthLink. They have bought ISPs by the dozen ever since the merger with MindSpring. Look at the number of ISPs they have now, under their banner..
Sprynet
Netcom
Sprint
OneMain
JPS
You can add Primenet (the dialup part of Global Crossing/Frontiernet/Globalcenter) which announced that they were handing over their customers to UrkStink last month.
Interestingly enough, an informal poll suggests that a sizable portion of the customer base hasn't received any notice at all; this may account for less than 100% fleeing ahead of the disaster.
P. J. Plauger reported in (IIRC) Dr. Dobb's Journal on using buddy-system programming at Whitesmith's. The original reason was lack of seats, but the surprising result was better productivity than when both bodies had access to keyboards and screens.
Seems to be a matter of enforced peer review, which any CS (self included) will tell you is both very hard to get in a corporate environment and essential to good software development.
Well, even if RAMBUS loses this case altogether, there is still the competition between DDR and RDRAM. Who will win this battle is completely unknown.
Not really. Servers account for about half of all DRAM shipped, and the server mfgs are unanimous in rejecting Rambus. Likewise video, notebook, and most embedded applications. About the only space that Rambus can come close to competing in is high-end desktop machines.
As a result, DDR is guaranteed to dominate the volume shipments, which runs the economy-of-scale benefit their way. Bottom line: Rambus has to have a killer advantage of some sort to even stay in the game.
Right now the only major advantage that Rambus has is that Intel cleverly designed the Pentium IV pipeline around Rambus. Yes, that's right, the P4 is a Rambus-specific processor. Smooth move, Intel.
it doesn't quite shut down the Rambus litigation machine. Besides the totally bogus stuff, they do have a patent on variable CAS latency in the DRAM. One could argue that this was obvious, since everyone and her sister was using controller-side variable CAS latency from 'way back when. Regardless, they have the patent and it does seem to bear directly on SDRAM and DDR.
DDR II, on the other hand, has had it designed out. Dang.
In windows 9x, restart under "DOS command line only" mode, then use an hex editor after copying the.EXE and the protected file to another computer. You can also boot from a FreeDOS or Caldera DOS diskette to do this.
Game Over:
The original data is now useless, essentially 'corrupt', all because you throught you knew what you where doing, and did not. You've just copied some useless encrypted data. Well done.
That's some pretty impressive code, that can operate to corrupt the filesystem even when it's just being passively read from another operating system. How does Zero Knowledge gain such power over flux transitions on write-protected media?
In general, I'm pretty impressed by any scheme that ships the key with the message and expects the data to remain safe from prying eyes. Other issues, such as this requiring that your documents all reside in a common file with write and execute priveledges, and that document recipients be able to take over low-level system functions, are obvious.
Let's see: you take carbon (coal) and oxygen (air) and run a chemical reaction to give you carbon dioxide and energy. Then you add a bacterium and light and get back (Ta Da!) carbon and oxygen.
Either this takes more power for the light than you get from the power plant, or you have a perpetual motion machine. And if your light source (sunlight?) actually does provide more power than the plant produces, why bother with coal?
As far as I can see, all this story does is point out that Federal bureaucrats fund programs completely without technical review.
If I read this correctly, it specifically spikes the kind of claim expansion that Rambus pulled to make their 1990 patent cover the technology that was being developed in 1993-1994.
This kind of "submarine patent" has been abused massively, and was only partly addressed by the recent 20-year-after-first-file change.
For those of you who think that our advances are as important to peoples' lives as our grandparents' advances, I offer the following:
When my grandfather was a child, transportation was horse, train, or boat. When he turned fifty and I was born, it was trains, cars, and airplanes. It's still cars and airplanes.
When he was a child, long-range communications meant paper, at most by telegraph. When he turned fifty and I was born, it meant telephone, radio, and television. It's still mostly telephone, radio, and television.
When he was a child, middle-class families in the USA routinely lost several children to whooping cough, diphtheria, scarlet fever, etc. None of the families I knew as a child did. Logarithms don't count.
My grandfather had smallpox scars. I have a smallpox vaccination scar. My children don't -- because smallpox is gone.
Nobody in his town had indoor plumbing. Nobody in mine didn't.
As a boy, he read about automobiles. Before he died, he watched Armstrong and Aldrin walk on the Moon live on TV. My kids watch reruns of Apollo 13.
I'm busy making some of our wonderful techie toys, and certainly am not complaining about the graphite knee brace that lets me ski. But let's not kid ourselves: my parents and grandparents changed the way we live in utterly profound ways, and it's going to take something on the order of matter transmission to come close.
Microsoft will continue to protect any intellectual property that it embeds as objects in XML wrappers. "We will have proprietary formats to protect our intellectual property," he said.
This is just the real world - as long as MS's standards are open and not proprietry I have no particular problem with it.
Too bad, then, that MS has declared that they intend to use XML as a container for proprietary (read 'closed') data formats. We suspected this for some time, but in a recent interview Ballmer came right out and said so.
It's not the dialup "chickenboner" spammer that causes the big trouble at UUnet. It's the direct-connect account with a T1 or better line. They produce enough revenue that (as the story notes) both Legal and Sales have to be in the loop to get a disconnect.
There are professionals to take care of that. Once the standard Linux system is running, there will be no more time wasted on the M$ cycle of crash, reboot, crash, reboot, crash, reboot...
In your dreams. I'm the guinea pig; I'm the one who fought to be allowed to buy a Dell notebook with Linux. And the terms were that I not only do my own system administration but teach the network admins their way around a Linux system.
And, no, I'm not a professional sysadmin. I'm a professional polygon pusher and part-time electromagnetic egotist.
... that the Frame filters for KWord just got prioritized up. The semantics are similar enough that it should be a lot cleaner than Word, WordPerfect, or whatever.
EDA companies have no problem making money in the Linux world -- they're diving into it as fast as they can. The reason is that there's nothing remotely like their products on the horizon in either the speech or beer free camps.
Remember ESR's analysis: commodities tend towards free solutions, while bleeding-edge stuff can command premium pricing. It's just that publishing (word processing, dtp, etc.) are heading towards commodity status.
> They accuse Rambus of subverting the Joint
> Electron Device Engineering Council (JEDEC)
> process when the company kept its patents on
> SDRAM secret while attending JEDEC meetings
> intended to establish an open industry standard.
Can anyone throw more light on this? I haven't hear the jedec part before. Sounds tangential to the patent question, at best, but it's worthy of Bismarck, or at least Nixon, if true.
JEDEC's patent policy is that if a participating company holds IP which covering a proposed standard, the company
Must disclose that fact to the committee, and
Must declare its intention to licence the IP for reasonable and non-discriminatory terms.
JEDEC members acknowledge this policy as a condition of membership, and the policy is announced at the open of each committee meeting.
Rambus actually went quite a bit beyond this, though, since it appears that what they did was attend the meetings, where they found the direction that the industry was going. At that point, they went back and rewrote some abandoned patent applications to cover the proposed SDRAMs.
I've been in shops that Threw The Switch and mail service was at best flaky while they got the Exchange system working. You can probably sell a dual system as a fail-safe compromise.
Exchange actually is better suited as a departmental mail server than as a corporate mail host. So by all means the people who want to use Exchange should have an Exchange server hanging off of the main mail system, and those who don't want to use it can continue to use the Sendmail hosts.
Yes, the bookmarks file is a text file. And yes, I edit it by hand. And no, if Netscape doesn't like what I did it 'fixes' the 'error' for me.
Example: try replacing one of your bookmark descriptions with an <img="file://foo.jpg"> reference to make a graphical index file (REALLY handy if you want to use your bookmarks for your home page.) This used to work, but since (iirc) 4.74 the freaking tags get replaced by ampersand codes. Thanks loads, Netscape.
And if you 'accidentally' delete the "Personal Toolbar Folder" or the "New and Cool" stuff in it, including "What's New," they get thoughtfully replaced despite your best efforts, thus chewing up half a screenful of valuable homepage space for absolutely no value.
.. that garf were removable. Haven't tried the 6.0 flavor yet, but the 4.7x bunch just absolutely will not let you take the preformated sections out of your bookmark file.
There are a few other reasons why I'm probably going to stick with Mozilla at home, but when the admins at work give us a choice I'll probably take it.
You can't combine Open Source (which is a protected term, mind you) with selling the same product for thousands of dollars, as they still claim they will (and deserve to, IMO).
Bzzzzzzt!
Although it doesn't look like that's what Borland will be doing, there is nothing in the GPL to prevent anyone from selling GPL'd software.
What's more, there is nothing in the GPL allowing the source to be passed around without charge. What the GPL does require is that if you sell it, you must make the source available to the purchaser. Which isn't the same as waiving copyright.
Free speech and free beer are totally orthogonal. We're all familiar with
closed both (MS),
open both (Linux),
and closed-source freeware (StarOffice 5.2).
Why does it come as a surprise to remember that up until the 80s commercial software was almost always shipped with source?
In an SDMI world, your soundcard would refuse to play the new.wav because it still has the magic mark of Cain. Likewise, your video display will only display approved images, etc.
Yes, I know: this whole scheme depends on having every single manufacturer of electronic components and systems play along. That's what the DMCA is for (and the recent FCC decision requiring that TV sets do "rights management.") It's Part One of the move to make manufacture of non-SDMI equipment illegal.
Those of you protesting should take consolation: there will always be a place for you. However, you may never realize how fast good software can be written by small teams of top programmers working intensely--and how exhilarating is the experience.
A flat-out run down a double-diamond slope is exhilarating, too. Even with a break to go back up the lift, though, you either take time out for lunch or you're going to be one of my (ski patrol) customers.
Hardware is more my field, despite the CS degree, but we spend our share of wild creative rushes and there's nothing new about the pressure-cooker style of project management (read Tracy Kidder.) Until very recently the Wisdom was that those of us over (30-35-40-45; you pick one) were past our prime and due for replacement by those younger, childless, and more willing to put in the hours.
A funny thing happened. The industry seems to have woken up one morning and realized that after 30 years of steering projects through rocks (and all too often running into them) we Old Farts just might know some things that the fresh faces don't. Such as that the biggest risks to a project aren't technological, they're human and that the people on your project are the same version 1.0 types who built the Pyramids.
PG might want to seriously consider that his software-management style absolutely depends not on taking time to do things right but always being able to do them over. When your production is bits that might sound reasonable, but when it's measured in kilotonnes of steel or requires a half-million in mask tooling you had damn well better get it right the first try more often than not. For that, I want people who are rested, awake, and not already 95% gone on to their next Artistic Creation.
Slashdot Mirror
An old story I heard once said that the original Intel spec of the 8088 (?) processor included a spec for 1024 hardware interrupts. IBM, of course, in their wisdom, said that the chip was not for a main frame (or something similar), and instead changed to spec to 8 interrupts. Someplace down the road, in the 286 I think(?), they realised the needed more, and expanded to 16 via a second chip. Which is what we have had to deal with ever since.
How can something so well documented get garbled so soon?
The original 8086 design allowed 256 interrupt vectors, hardware and software combined. The hardware has always supported 256 vectors. Intel reserved the first 32 for future use, such as processor exceptions. IBM ignored that and assigned, for instance, INT 5 (which Intel had earmarked for array bounds exceptions) to the BIOS print screen function. Hardware interrupts got shoehorned in at the 0x10-0x1f range.
Why jam so much stuff in low when there were 256 available? Because the original ROM BASIC used the upper 208 or 224 (memory fails) for short-bytecount subroutine calls. It made for a very simple, compact interpreter. A look at a disassembly of the early ones showed that it was basically machine-translated Z80 code!
If you have a floppy drive on your computer, chances are it is still using the ISA bus... So just because you don't have an ISA "slot", don't think that the bus is gone...
No, it's really gone. The floppy connector is just part of the South Bridge chip, along with the keyboard controller, serial and parallel ports, and other low-bandwidth stuff. It hangs off of the PCI bus. These devices are just hard-coded to ACT like the old ISA equivalents, with their addresses, interrupts, DMA, etc. being hard-addigned to the old standard values.
Need more CPU power or more memory? Hot-plug a module into the Infiniband Switch.
Don't gush. Infiniband, or any cabled serial connection, will never be a memory connection worth having. The reason is latency: what matters with memory isn't how many terabytes you can deliver per hour, but how many picoseconds the processor has to wait for that data that's stalling the pipeline right now. Which is very hard to reduce when the address has to be shipped a bit at a time over a serial link (64 bit times @ 10 Gb/s = 6.4 ns), transported over the cable (~50 ps/cm * ~60 cm = 3.0 ns), memory accessed (technology dependent), serialized (another 64 bits, 6.4 ns), shipped back (another 3.0 ns) and finally it gets to the processor. You've added 18.8 ns over and above any protocol overhead (usually much worse) and that's at 10 Gb/s!
Not gonna happen.
One good backhoe accident, and you suddenly have a bunch of intranets. In theory all of the Tier-One operators peer at multiple points, but in practice they route their own traffic through the same facilities. Likewise, in theory the Tier-Two operators are multiconnected through multiple Tier-One providers but in practice thanks to volume contract terms they are single-homed.
Below Tier Two, it really doesn't matter.
Beats me as to who got gobbled up, but I'll bet dollars to donuts that the buyer was EarthLink. They have bought ISPs by the dozen ever since the merger with MindSpring. Look at the number of ISPs they have now, under their banner..
Sprynet
Netcom
Sprint
OneMain
JPS
You can add Primenet (the dialup part of Global Crossing/Frontiernet/Globalcenter) which announced that they were handing over their customers to UrkStink last month.
Interestingly enough, an informal poll suggests that a sizable portion of the customer base hasn't received any notice at all; this may account for less than 100% fleeing ahead of the disaster.
P. J. Plauger reported in (IIRC) Dr. Dobb's Journal on using buddy-system programming at Whitesmith's. The original reason was lack of seats, but the surprising result was better productivity than when both bodies had access to keyboards and screens.
Seems to be a matter of enforced peer review, which any CS (self included) will tell you is both very hard to get in a corporate environment and essential to good software development.
Well, even if RAMBUS loses this case altogether, there is still the competition between DDR and RDRAM. Who will win this battle is completely unknown.
Not really. Servers account for about half of all DRAM shipped, and the server mfgs are unanimous in rejecting Rambus. Likewise video, notebook, and most embedded applications. About the only space that Rambus can come close to competing in is high-end desktop machines.
As a result, DDR is guaranteed to dominate the volume shipments, which runs the economy-of-scale benefit their way. Bottom line: Rambus has to have a killer advantage of some sort to even stay in the game.
Right now the only major advantage that Rambus has is that Intel cleverly designed the Pentium IV pipeline around Rambus. Yes, that's right, the P4 is a Rambus-specific processor. Smooth move, Intel.
it doesn't quite shut down the Rambus litigation machine. Besides the totally bogus stuff, they do have a patent on variable CAS latency in the DRAM. One could argue that this was obvious, since everyone and her sister was using controller-side variable CAS latency from 'way back when. Regardless, they have the patent and it does seem to bear directly on SDRAM and DDR.
DDR II, on the other hand, has had it designed out. Dang.
That's some pretty impressive code, that can operate to corrupt the filesystem even when it's just being passively read from another operating system. How does Zero Knowledge gain such power over flux transitions on write-protected media?
In general, I'm pretty impressed by any scheme that ships the key with the message and expects the data to remain safe from prying eyes. Other issues, such as this requiring that your documents all reside in a common file with write and execute priveledges, and that document recipients be able to take over low-level system functions, are obvious.
Let's see: you take carbon (coal) and oxygen (air) and run a chemical reaction to give you carbon dioxide and energy. Then you add a bacterium and light and get back (Ta Da!) carbon and oxygen.
Either this takes more power for the light than you get from the power plant, or you have a perpetual motion machine. And if your light source (sunlight?) actually does provide more power than the plant produces, why bother with coal?
As far as I can see, all this story does is point out that Federal bureaucrats fund programs completely without technical review.
If I read this correctly, it specifically spikes the kind of claim expansion that Rambus pulled to make their 1990 patent cover the technology that was being developed in 1993-1994.
This kind of "submarine patent" has been abused massively, and was only partly addressed by the recent 20-year-after-first-file change.
I'm busy making some of our wonderful techie toys, and certainly am not complaining about the graphite knee brace that lets me ski. But let's not kid ourselves: my parents and grandparents changed the way we live in utterly profound ways, and it's going to take something on the order of matter transmission to come close.
Glad to oblige. From "Microsoft's Ballmer: Sun has no clue.":
Microsoft will continue to protect any intellectual property that it embeds as objects in XML wrappers. "We will have proprietary formats to protect our intellectual property," he said.
This is just the real world - as long as MS's standards are open and not proprietry I have no particular problem with it.
Too bad, then, that MS has declared that they intend to use XML as a container for proprietary (read 'closed') data formats. We suspected this for some time, but in a recent interview Ballmer came right out and said so.
It's not the dialup "chickenboner" spammer that causes the big trouble at UUnet. It's the direct-connect account with a T1 or better line. They produce enough revenue that (as the story notes) both Legal and Sales have to be in the loop to get a disconnect.
There are professionals to take care of that. Once the standard Linux system is running, there will be no more time wasted on the M$ cycle of crash, reboot, crash, reboot, crash, reboot...
In your dreams. I'm the guinea pig; I'm the one who fought to be allowed to buy a Dell notebook with Linux. And the terms were that I not only do my own system administration but teach the network admins their way around a Linux system.
And, no, I'm not a professional sysadmin. I'm a professional polygon pusher and part-time electromagnetic egotist.
... that the Frame filters for KWord just got prioritized up. The semantics are similar enough that it should be a lot cleaner than Word, WordPerfect, or whatever.
EDA companies have no problem making money in the Linux world -- they're diving into it as fast as they can. The reason is that there's nothing remotely like their products on the horizon in either the speech or beer free camps.
Remember ESR's analysis: commodities tend towards free solutions, while bleeding-edge stuff can command premium pricing. It's just that publishing (word processing, dtp, etc.) are heading towards commodity status.
> Electron Device Engineering Council (JEDEC)
> process when the company kept its patents on
> SDRAM secret while attending JEDEC meetings
> intended to establish an open industry standard.
Can anyone throw more light on this? I haven't hear the jedec part before. Sounds tangential to the patent question, at best, but it's worthy of Bismarck, or at least Nixon, if true.
JEDEC's patent policy is that if a participating company holds IP which covering a proposed standard, the company
- Must disclose that fact to the committee, and
- Must declare its intention to licence the IP for reasonable and non-discriminatory terms.
JEDEC members acknowledge this policy as a condition of membership, and the policy is announced at the open of each committee meeting.Rambus actually went quite a bit beyond this, though, since it appears that what they did was attend the meetings, where they found the direction that the industry was going. At that point, they went back and rewrote some abandoned patent applications to cover the proposed SDRAMs.
What's with the Holy War stuff, anyway?
I've been in shops that Threw The Switch and mail service was at best flaky while they got the Exchange system working. You can probably sell a dual system as a fail-safe compromise.
Exchange actually is better suited as a departmental mail server than as a corporate mail host. So by all means the people who want to use Exchange should have an Exchange server hanging off of the main mail system, and those who don't want to use it can continue to use the Sendmail hosts.
What a concept.
Yes, the bookmarks file is a text file. And yes, I edit it by hand. And no, if Netscape doesn't like what I did it 'fixes' the 'error' for me.
Example: try replacing one of your bookmark descriptions with an <img="file://foo.jpg"> reference to make a graphical index file (REALLY handy if you want to use your bookmarks for your home page.) This used to work, but since (iirc) 4.74 the freaking tags get replaced by ampersand codes. Thanks loads, Netscape.
And if you 'accidentally' delete the "Personal Toolbar Folder" or the "New and Cool" stuff in it, including "What's New," they get thoughtfully replaced despite your best efforts, thus chewing up half a screenful of valuable homepage space for absolutely no value.
.. that garf were removable. Haven't tried the 6.0 flavor yet, but the 4.7x bunch just absolutely will not let you take the preformated sections out of your bookmark file.
There are a few other reasons why I'm probably going to stick with Mozilla at home, but when the admins at work give us a choice I'll probably take it.
Bzzzzzzt!
Although it doesn't look like that's what Borland will be doing, there is nothing in the GPL to prevent anyone from selling GPL'd software.
What's more, there is nothing in the GPL allowing the source to be passed around without charge. What the GPL does require is that if you sell it, you must make the source available to the purchaser. Which isn't the same as waiving copyright.
Free speech and free beer are totally orthogonal. We're all familiar with
- closed both (MS),
- open both (Linux),
- and closed-source freeware (StarOffice 5.2).
Why does it come as a surprise to remember that up until the 80s commercial software was almost always shipped with source?In an SDMI world, your soundcard would refuse to play the new .wav because it still has the magic mark of Cain. Likewise, your video display will only display approved images, etc.
Yes, I know: this whole scheme depends on having every single manufacturer of electronic components and systems play along. That's what the DMCA is for (and the recent FCC decision requiring that TV sets do "rights management.") It's Part One of the move to make manufacture of non-SDMI equipment illegal.
Those of you protesting should take consolation: there will always be a place for you. However, you may never realize how fast good software can be written by small teams of top programmers working intensely--and how exhilarating is the experience.
A flat-out run down a double-diamond slope is exhilarating, too. Even with a break to go back up the lift, though, you either take time out for lunch or you're going to be one of my (ski patrol) customers.
Hardware is more my field, despite the CS degree, but we spend our share of wild creative rushes and there's nothing new about the pressure-cooker style of project management (read Tracy Kidder.) Until very recently the Wisdom was that those of us over (30-35-40-45; you pick one) were past our prime and due for replacement by those younger, childless, and more willing to put in the hours.
A funny thing happened. The industry seems to have woken up one morning and realized that after 30 years of steering projects through rocks (and all too often running into them) we Old Farts just might know some things that the fresh faces don't. Such as that the biggest risks to a project aren't technological, they're human and that the people on your project are the same version 1.0 types who built the Pyramids.
PG might want to seriously consider that his software-management style absolutely depends not on taking time to do things right but always being able to do them over. When your production is bits that might sound reasonable, but when it's measured in kilotonnes of steel or requires a half-million in mask tooling you had damn well better get it right the first try more often than not. For that, I want people who are rested, awake, and not already 95% gone on to their next Artistic Creation.