We have some real problems in the Open Source world of programming. Microsoft has been cleaning up their Mess Inside.
C# is a win for Microsoft. It's a decent language. We have some real problems in the language end of open source. C is still widely used, despite its age and some terrible design flaws. C++ has become something of a mess, and the C++ standards people are off in template fantasyland, instead of fixing the thing. Java has acquired too much baggage, although the underlying language isn't bad. (Amusingly, Java turned out to be the replacement for COBOL, something its designers never intended.) Perl doesn't scale up well, and Python is too slow. We don't have a good, safe, hard-code-generating language in the open source world. (And don't say that's impossible; it's been done several times.)
Microsoft is better at middleware than the open source world, especially the parts that have to play well together. All Microsoft apps have access to a standard interprocess communication system and a standard database. Neither is great, but they're always there, and they're standardized. The Linux world has about a half dozen interprocess communication schemes, none well-supported by the OS. There's CORBA; Gnome and OpenOffice both use CORBA. Incompatible versions. All of this was horribly ugly prior to ".NET", and the transition was even uglier
("Managed C++"). But the all-C# world is relatively clean.
Build tools in the open source world are still 1970s technology. I was using "make", "sccs", "db", and a version of EMACS in 1978. Now we have "gmake", Subversion, "gdb", and EMACS. That's
rather minor progress. For some reason, Eclipse has never caught on much.
And, as the original author pointed out, Microsoft is way, way ahead on developer documentation. They have people writing articles on how to deal with the ugly parts of their own software.
That's rare in the Open Source world. There's an endless supply of beginner books and recipe books, but quality technical notes on the tough parts are rare.
Much to my annoyance, Anna Eshoo, the local congresswoman for Silicon Valley, is a cosponsor on this turkey. I sent in a letter. This thing needs some work in committee to clean it up, preferably well enough that EPIC is satisfied with it.
E3 is at Barker Hangar, the big WWII metal Quonset hut at Santa Monica Airport? That's moving downscale from the LA Convention Center. Way downscale. That place is used for flea markets and remainder sales.
"The Barker Hangar venue, a short drive from the hotels, will allow participating companies to showcase their games in standardized, turnkey displays areas ranging from 100 square feet to 400 square feet. All display areas will be developed by show management to ensure that the venue is staged efficiently."
The Codex Alimentarius, the international standards body for food, has a standard for chocolate. They require >35% cocoa solids for "chocolate". And they limit other fats:
"The addition of vegetable fats other than cocoa butter shall not exceed 5% of the finished product, after
deduction of the total weight of any other added edible foodstuffs, without reducing the minimum contents of
cocoa materials. Where required by the authorities having jurisdiction, the nature of the vegetable fats
permitted for this purpose may be prescribed in applicable legislation.
Pad2Pad does less than they used to. The original idea of Pad2Pad was that they'd make the blank board, then assemble and solder all the parts, using anything in the Digi-Key catalog. That made it useful, especially since surface mount device soldering really needs to be done in a production environment with the right tooling.
But they couldn't do it profitably. Now they're just another blank PC board maker, of which there are hundreds. It's been routine to send out your board design files and get boards back for
almost twenty years now.
Actually, that was the Mars Pathfinder. It was running VxWorks, and the effect of the priority inversion was that the stall timer would trip and reset the whole system. The problem was that VxWorks, like QNX, lets you turn off "priority inheritance" on a mutex. This is usually a bad decision, but that was done on the Mars Pathfinder, and created the possibility of a livelock.
So they uploaded a patch to change that mutex to "priority inheritance on", and it worked consistently thereafter.
I'm impressed. I did some work on CPU dispatching on a mainframe system in the distant past,
and we were never able to beat O(log n) on an ordered dispatch queue. The obvious implementation is O(n); getting to O(log n) requires a tree, and I want to see how they got to O(1).
This stuff really matters now. If we're going to have 80-CPU multiprocessors, all the main operating system algorithms have to be near O(1), or the scale-up is a lose.
This seems to be going on forever, but it's not. The end is in sight. SCO can stall, but the process does move onward. Discovery is over; everything significant that's going to come out has come out.
Now we're in the stage where bogus claims get thrown out via summary judgment motions. That phase is well along; summary judgment motions have been made and briefed in both the Novell and IBM cases. Soon the judge will decide them.
The Novell-SCO contract says that Novell retains "all copyrights". If the judge rules that the contract means what it says, that ends the copyright issue.
Then, based on that, the summary judgment motions by IBM against SCO mostly get decided in favor of IBM.
Remember, for IBM, this is no longer a problem. Is it hurting Linux server sales? No. Are customers bothered by it? No. Can IBM afford the legal costs? IBM revenue was $91,000,000,000 in 2006. I doubt this issue gets much management attention in Armonk any more.
Linux really doesn't need a new process scheduler. What it could really do with is I/O prioritisation.
QNX has that, which is essential for real-time work.
QNX has the advantage that I/O, like almost everything else in QNX, is done via inter-process message passing operations. The message passing system uses priority queues, and so requests to file systems and devices get handled in priority order. So resource managers (file systems, device drivers, etc.) don't have to explicitly handle priorities; it's done for them. Some resource managers, like disk handlers, process multiple requests at a time so they can reorder them to optimize access, but network devices and such are FIFO at the resource manager level and priority ordered at the message level.
The end result is that you can compile or surf the web on a system that's also doing real time work without interfering with the real time tasks.
The red state version of this would be Helio's "Buddy Beacon" for NRA members. If you're in trouble, you punch the panic button on your cell phone, and all NRA members within a mile or so get an alert. In two minutes there's enough firepower on site for a small war.
During WWII and the Korea war it was normal to drop hundreds of bombs, flattening several city block or even entire villages, just to hit one bridge.
It's truly amazing how bad navigational technology was during WWII. Bombers had trouble finding the right city. Early in the war, the Germans had trouble finding London. There were cases of aircraft landing in the wrong country. Looking up at the sky with a sextant was state of the art. Radio beams were tried by both sides, but were too easy to jam. Hits were either accidental or achieved by getting really close to the target.
It's a bigger problem than most people realize. Twenty years ago, the original footage shot for a film might be 3x what finally appears on the screen, maybe more for a really big-budget film. Today, not only will there be more raw material, there's far more intermediate work product.
A big project might have twenty layers going into a final frame. During the project, all that stuff is stored. But then what? Where does all that stuff get archived? There will be terabytes of stuff for any major film. Studios would like to keep it around. It might be profitably reused some day. But how?
Not only is there a huge amount of information, it's in a format used by some content management system that probably won't be runnable ten years from now. Worse, each subcontractor will have their own systems. Look at the list of effects companies involved with any major film today. Will they have the intermediates of a project from ten years ago?
And who's going to archive it? Many production companies are ephemeral, lasting only for the lifetime of the project. There's no ongoing operation responsible for the work products. The major studios may be involved with financing and distribution, and may provide some facilities, but they are no longer the organizations who directly make films. Hollywood is a rental business.
It can't be worse than high school. You have to go. You don't get paid. You don't get to pick whom you're with. You're not anonymous. The place is regimented, cliquish, and crowded. Being popular is hard work, and being unpopular is a pain.
Myspace is a relief from all that. The worst day on Myspace is better than a moderately bad day of high school. On Myspace people can annoy you, but you can block them. Unlike real life.
This is the sort of thing unions were created to fix. Crunch time means crunch pay.
Sysadmins need something like the IATSE Hollywood Basic Agreement, which covers most people working on a movie shoot other than the actors. When film productions go into crunch mode, pay goes through the roof. There's overtime. There's double overtime when hours per day go up. Emergency calls in the middle of the night carry a minimum of 4 hours of pay at double the normal rate, plus there's pay for being on-call.
Which is why producers and studios spend time on doing good film scheduling. When management screws up, management pays the bills.
Maybe the problem with the thing is that it takes significant maintenance support. Xerox copiers and printers back in the selenium drum era were leased, because they cost so much and required considerable skilled maintenance. If this new technology has that problem, a lease-only approach at introduction makes sense.
Most home door locks are terrible. The standard for them specifies that they should resist opening for 15 seconds with a screwdriver. Really.
The US Department of Housing and Urban Development used to have good standards for doors and locks in their housing projects. Every unit had a steel-sheathed fire door with a steel frame and locks that could resist serious abuse. In a building with interior walls of reinforced concrete, this provided quite good security. Which was needed.
I once saw a news video where some cops were raiding an apartment in a housing project. They show up at the door with a two-person battering ram, and bang away for a while. After about thirty seconds of banging, the cops are exhausted, and they try yelling through the door at the occupant to open the door. From inside, a sleepy voice answers "I can't. You broke the lock". The door held until they sent out for power saws.
You cannot record anyone without permission unless you have a court order.
No, some states prohibit that, and some don't. Federal law does not. For the states that prohibit it, it only applies to calls where both ends are in the state, since across state lines, federal law applies to telecom.
Perhaps one (arguably) good opportunity for the platform has passed.
UNIX/Linux on the desktop has failed to go mainstream so many times now. In the early 1980s there was an AT&T PC that ran UNIX. That went nowhere. The whole UNIX workstation industry has come and gone. (If you'd like a high-end SGI or Sun workstation, Wierd Stuff Warehouse in Sunnyvale, California has a pile of them at low, low prices. Hurry before they rust.) Linux laptops are no longer available from any major vendor, and the minor vendors charge more for a Linux laptop than the same system costs with Windows.
I have hope for the OLPC initiative, though. That's the first major rethink in desktops in a decade.
"The time for mass consumer sales of Linux on the desktop has not yet come."
No. It's past.
Linux missed the window. 2003-2004 was when Linux on the desktop might have made it.
WalMart, Dell, and HP all had supported Linux offerings. Microsoft XP was stale and buggy, while Longhorn/Vista was stalled. Firefox worked great, OpenOffice worked OK, and things looked good for desktop Linux. That was the big opportunity.
The window closed when playing DRM-protected video content on PCs became important to the average consumer. Linux couldn't keep up in that area, because the proprietary codec vendors didn't all support Linux. As Microsoft and Apple took control of video formats, Linux was squeezed out.
Neither iTunes nor Microsoft video codecs are available for Linux.
I had real hopes for Linux on the desktop, but it's not going to happen now. Focus on the OLPC software; that might make it.
Microsoft needs low-wage customer-paid armies of second-tier people to keep their software mess going. It's in their interest that computer science education be dumbed down.
That project took way too long. I remember people working on it when I went through Stanford in the mid-1980s. It was something of a boondoggle; it mostly produced students, not flight hardware. I'm glad to hear it finally worked, though.
Incredibly dumb move. It just draws attention to some material that probably sucks.
Back in 1989, during the 2 Live Crew censorship controversy, I remarked, after listening to their stuff, that without the censorship, they would have never made it off the South Florida club circuit. With the censorship controversy, they hit #29 on the Billboard 200.
But by 1991, they'd peaked, and broke up around 1992. There was a "New 2 Live Crew", which went nowhere and broke up in 1995, a reunion in 1998, and another reunion in 2005. Nobody cared much. Wikipedia says "The two core members are still popular within the Miami Bass community and Dancehall goers".
Or, in other words, they're back on the South Florida club circuit where they belong. Censorship gave them their 15 minutes of fame.
Slashdot Mirror
We have some real problems in the Open Source world of programming. Microsoft has been cleaning up their Mess Inside.
C# is a win for Microsoft. It's a decent language. We have some real problems in the language end of open source. C is still widely used, despite its age and some terrible design flaws. C++ has become something of a mess, and the C++ standards people are off in template fantasyland, instead of fixing the thing. Java has acquired too much baggage, although the underlying language isn't bad. (Amusingly, Java turned out to be the replacement for COBOL, something its designers never intended.) Perl doesn't scale up well, and Python is too slow. We don't have a good, safe, hard-code-generating language in the open source world. (And don't say that's impossible; it's been done several times.)
Microsoft is better at middleware than the open source world, especially the parts that have to play well together. All Microsoft apps have access to a standard interprocess communication system and a standard database. Neither is great, but they're always there, and they're standardized. The Linux world has about a half dozen interprocess communication schemes, none well-supported by the OS. There's CORBA; Gnome and OpenOffice both use CORBA. Incompatible versions. All of this was horribly ugly prior to ".NET", and the transition was even uglier ("Managed C++"). But the all-C# world is relatively clean.
Build tools in the open source world are still 1970s technology. I was using "make", "sccs", "db", and a version of EMACS in 1978. Now we have "gmake", Subversion, "gdb", and EMACS. That's rather minor progress. For some reason, Eclipse has never caught on much.
And, as the original author pointed out, Microsoft is way, way ahead on developer documentation. They have people writing articles on how to deal with the ugly parts of their own software. That's rare in the Open Source world. There's an endless supply of beginner books and recipe books, but quality technical notes on the tough parts are rare.
I don't like it either.
Much to my annoyance, Anna Eshoo, the local congresswoman for Silicon Valley, is a cosponsor on this turkey. I sent in a letter. This thing needs some work in committee to clean it up, preferably well enough that EPIC is satisfied with it.
E3 is at Barker Hangar, the big WWII metal Quonset hut at Santa Monica Airport? That's moving downscale from the LA Convention Center. Way downscale. That place is used for flea markets and remainder sales.
"The Barker Hangar venue, a short drive from the hotels, will allow participating companies to showcase their games in standardized, turnkey displays areas ranging from 100 square feet to 400 square feet. All display areas will be developed by show management to ensure that the venue is staged efficiently."
The Codex Alimentarius, the international standards body for food, has a standard for chocolate. They require >35% cocoa solids for "chocolate". And they limit other fats:
"The addition of vegetable fats other than cocoa butter shall not exceed 5% of the finished product, after deduction of the total weight of any other added edible foodstuffs, without reducing the minimum contents of cocoa materials. Where required by the authorities having jurisdiction, the nature of the vegetable fats permitted for this purpose may be prescribed in applicable legislation.
What are the numbers in the FDA proposal?
Pad2Pad does less than they used to. The original idea of Pad2Pad was that they'd make the blank board, then assemble and solder all the parts, using anything in the Digi-Key catalog. That made it useful, especially since surface mount device soldering really needs to be done in a production environment with the right tooling.
But they couldn't do it profitably. Now they're just another blank PC board maker, of which there are hundreds. It's been routine to send out your board design files and get boards back for almost twenty years now.
Drivers can, and bugs in the OS can. User-run programs can only (accidentally) trigger one of those...
If only. The way Microsoft does DRM, much of audio and video processing takes place in kernel space.
Actually, that was the Mars Pathfinder. It was running VxWorks, and the effect of the priority inversion was that the stall timer would trip and reset the whole system. The problem was that VxWorks, like QNX, lets you turn off "priority inheritance" on a mutex. This is usually a bad decision, but that was done on the Mars Pathfinder, and created the possibility of a livelock.
So they uploaded a patch to change that mutex to "priority inheritance on", and it worked consistently thereafter.
I'm impressed. I did some work on CPU dispatching on a mainframe system in the distant past, and we were never able to beat O(log n) on an ordered dispatch queue. The obvious implementation is O(n); getting to O(log n) requires a tree, and I want to see how they got to O(1).
This stuff really matters now. If we're going to have 80-CPU multiprocessors, all the main operating system algorithms have to be near O(1), or the scale-up is a lose.
This seems to be going on forever, but it's not. The end is in sight. SCO can stall, but the process does move onward. Discovery is over; everything significant that's going to come out has come out. Now we're in the stage where bogus claims get thrown out via summary judgment motions. That phase is well along; summary judgment motions have been made and briefed in both the Novell and IBM cases. Soon the judge will decide them.
The Novell-SCO contract says that Novell retains "all copyrights". If the judge rules that the contract means what it says, that ends the copyright issue.
Then, based on that, the summary judgment motions by IBM against SCO mostly get decided in favor of IBM.
Remember, for IBM, this is no longer a problem. Is it hurting Linux server sales? No. Are customers bothered by it? No. Can IBM afford the legal costs? IBM revenue was $91,000,000,000 in 2006. I doubt this issue gets much management attention in Armonk any more.
Linux really doesn't need a new process scheduler. What it could really do with is I/O prioritisation.
QNX has that, which is essential for real-time work.
QNX has the advantage that I/O, like almost everything else in QNX, is done via inter-process message passing operations. The message passing system uses priority queues, and so requests to file systems and devices get handled in priority order. So resource managers (file systems, device drivers, etc.) don't have to explicitly handle priorities; it's done for them. Some resource managers, like disk handlers, process multiple requests at a time so they can reorder them to optimize access, but network devices and such are FIFO at the resource manager level and priority ordered at the message level.
The end result is that you can compile or surf the web on a system that's also doing real time work without interfering with the real time tasks.
The red state version of this would be Helio's "Buddy Beacon" for NRA members. If you're in trouble, you punch the panic button on your cell phone, and all NRA members within a mile or so get an alert. In two minutes there's enough firepower on site for a small war.
During WWII and the Korea war it was normal to drop hundreds of bombs, flattening several city block or even entire villages, just to hit one bridge.
It's truly amazing how bad navigational technology was during WWII. Bombers had trouble finding the right city. Early in the war, the Germans had trouble finding London. There were cases of aircraft landing in the wrong country. Looking up at the sky with a sextant was state of the art. Radio beams were tried by both sides, but were too easy to jam. Hits were either accidental or achieved by getting really close to the target.
It's a bigger problem than most people realize. Twenty years ago, the original footage shot for a film might be 3x what finally appears on the screen, maybe more for a really big-budget film. Today, not only will there be more raw material, there's far more intermediate work product. A big project might have twenty layers going into a final frame. During the project, all that stuff is stored. But then what? Where does all that stuff get archived? There will be terabytes of stuff for any major film. Studios would like to keep it around. It might be profitably reused some day. But how?
Not only is there a huge amount of information, it's in a format used by some content management system that probably won't be runnable ten years from now. Worse, each subcontractor will have their own systems. Look at the list of effects companies involved with any major film today. Will they have the intermediates of a project from ten years ago?
And who's going to archive it? Many production companies are ephemeral, lasting only for the lifetime of the project. There's no ongoing operation responsible for the work products. The major studios may be involved with financing and distribution, and may provide some facilities, but they are no longer the organizations who directly make films. Hollywood is a rental business.
It can't be worse than high school. You have to go. You don't get paid. You don't get to pick whom you're with. You're not anonymous. The place is regimented, cliquish, and crowded. Being popular is hard work, and being unpopular is a pain.
Myspace is a relief from all that. The worst day on Myspace is better than a moderately bad day of high school. On Myspace people can annoy you, but you can block them. Unlike real life.
This is the sort of thing unions were created to fix. Crunch time means crunch pay.
Sysadmins need something like the IATSE Hollywood Basic Agreement, which covers most people working on a movie shoot other than the actors. When film productions go into crunch mode, pay goes through the roof. There's overtime. There's double overtime when hours per day go up. Emergency calls in the middle of the night carry a minimum of 4 hours of pay at double the normal rate, plus there's pay for being on-call.
Which is why producers and studios spend time on doing good film scheduling. When management screws up, management pays the bills.
Another great Web 2.0 concept.
Some days I think the Web peaked at HTML 3.2.
Maybe the problem with the thing is that it takes significant maintenance support. Xerox copiers and printers back in the selenium drum era were leased, because they cost so much and required considerable skilled maintenance. If this new technology has that problem, a lease-only approach at introduction makes sense.
Most home door locks are terrible. The standard for them specifies that they should resist opening for 15 seconds with a screwdriver. Really.
The US Department of Housing and Urban Development used to have good standards for doors and locks in their housing projects. Every unit had a steel-sheathed fire door with a steel frame and locks that could resist serious abuse. In a building with interior walls of reinforced concrete, this provided quite good security. Which was needed.
I once saw a news video where some cops were raiding an apartment in a housing project. They show up at the door with a two-person battering ram, and bang away for a while. After about thirty seconds of banging, the cops are exhausted, and they try yelling through the door at the occupant to open the door. From inside, a sleepy voice answers "I can't. You broke the lock". The door held until they sent out for power saws.
Now that's how security should work.
You cannot record anyone without permission unless you have a court order.
No, some states prohibit that, and some don't. Federal law does not. For the states that prohibit it, it only applies to calls where both ends are in the state, since across state lines, federal law applies to telecom.
Perhaps one (arguably) good opportunity for the platform has passed.
UNIX/Linux on the desktop has failed to go mainstream so many times now. In the early 1980s there was an AT&T PC that ran UNIX. That went nowhere. The whole UNIX workstation industry has come and gone. (If you'd like a high-end SGI or Sun workstation, Wierd Stuff Warehouse in Sunnyvale, California has a pile of them at low, low prices. Hurry before they rust.) Linux laptops are no longer available from any major vendor, and the minor vendors charge more for a Linux laptop than the same system costs with Windows.
I have hope for the OLPC initiative, though. That's the first major rethink in desktops in a decade.
"The time for mass consumer sales of Linux on the desktop has not yet come."
No. It's past.
Linux missed the window. 2003-2004 was when Linux on the desktop might have made it. WalMart, Dell, and HP all had supported Linux offerings. Microsoft XP was stale and buggy, while Longhorn/Vista was stalled. Firefox worked great, OpenOffice worked OK, and things looked good for desktop Linux. That was the big opportunity.
The window closed when playing DRM-protected video content on PCs became important to the average consumer. Linux couldn't keep up in that area, because the proprietary codec vendors didn't all support Linux. As Microsoft and Apple took control of video formats, Linux was squeezed out. Neither iTunes nor Microsoft video codecs are available for Linux.
I had real hopes for Linux on the desktop, but it's not going to happen now. Focus on the OLPC software; that might make it.
The National Center for Women and Technology is funded by Microsoft and Avaya (which "specializes in call center technology", according to Wikipedia). This sounds like a keep-wages-down lobbying effort, like Microsoft's lobbying for more H1B visas.
Microsoft needs low-wage customer-paid armies of second-tier people to keep their software mess going. It's in their interest that computer science education be dumbed down.
That project took way too long. I remember people working on it when I went through Stanford in the mid-1980s. It was something of a boondoggle; it mostly produced students, not flight hardware. I'm glad to hear it finally worked, though.
Incredibly dumb move. It just draws attention to some material that probably sucks.
Back in 1989, during the 2 Live Crew censorship controversy, I remarked, after listening to their stuff, that without the censorship, they would have never made it off the South Florida club circuit. With the censorship controversy, they hit #29 on the Billboard 200. But by 1991, they'd peaked, and broke up around 1992. There was a "New 2 Live Crew", which went nowhere and broke up in 1995, a reunion in 1998, and another reunion in 2005. Nobody cared much. Wikipedia says "The two core members are still popular within the Miami Bass community and Dancehall goers".
Or, in other words, they're back on the South Florida club circuit where they belong. Censorship gave them their 15 minutes of fame.