This was spelled out in the writer's guide that they would give to people who wanted to write episodes for TOS. The phaser was a three piece modular system. The basic phaser was a small box roughly the size and shape of the communicator and was to be used concealed in missions where they didn't expect to need weapons. If that wasn't enough then you would mount the basic phaser on a pistol grip which made aiming easier and increased firepower by adding its own power supply. Take a careful look at a picture of the pistol model and you will see what you thought was a communicator attached to the top of it. The third option was a rifle style piece to which you attached the pistol version and again increased aim and power.
Since we are being picky, the 68000 actually had two simple 16 bit ALUs and one complex 16 bit ALU. The first was directly coupled to the high bits of the data and address registers and the second to the low bits of the address registers. The complex ALU was tightly coupled to the low bits of the data registers. The middle section of the processor (low address bits) could be connected to the other two when needed, but it was possible for the processor to handle 48 bits at a time in some common situations!
See "Microprogrammed Implementation of a Single Chip Microprocessor" by Skip Stritter and Nick Tredennick in SIGMICRO Newsletter Volume 9, Number 4, December 1978.
Given that we haven't heard more about this since then I am not taking it too seriously. Here in Brazil we have lots of similar articles where some government official is interviewed and claims "Brazil rejects OLPC". But that official is not involved in the decision process and those who are continue to be very positive about it.
Damn, I hate the academic world for letting Microsoft create and patent Singularity! Why didn't MIT, CMU or Berkeley do this research first and put the results in the public domain!
In one of their papers they explicitly contrast the adiabatic approach (which seeks the lowest possible power per operation at the cost of slower operations) with their own solution (which seeks to lower power even while operating at the highest possible speeds). So for the best possible battery life you would want adiabatic logic, but for multiple gigahertz operation Multigig seems like a nice option.
I am pretty sure the QNX microkernel took up less than 1KB unless you counted the interrupt table and stubs. But I am thinking of the pre 2.X versions which didn't include networking and distributed computing. Even with these extensions my guess is that the microkernel was still less than 3KB.
It would be interesting to know what the size of recent versions is.
I don't know why some people think Self is a very different language than Smalltalk-80. With the optional parser and GNU Smalltalk classes you can even file in Smalltalk-80 code and run it.
I don't remember who said "Self is like Smalltalk, only more so" but that is a great definition. To avoid having to repeat this discussion all the time I have renamed my Self/R project as Neo Smalltalk.
Mr. "don't mode me in!" is also the guy who invented the cut/copy/paste editing where you always replace the selection that is used by everybody these days.
There are two different services in that web site: hosting video which can be downloaded and searching through external video (this has been available for some time now). There isn't much local content yet so you are unlikely to run across it in a random search. Online videos have a little triangle beside the title that you can click on to watch the movie. Try one of their suggested searches, like "capoeira" and you will see that they haven't taken anything offline.
When are open-sourced based CPUs going to be available? Does anyone know of any available? I don't *feel* free with the current processor offerings available.
Here you go. Check out the OpenRISC 1000 - I am guessing that it will be particularly interesting for you since the 1200 version has been used to demonstrate Linux. Of course the MIPS and Sparc clones can do so as well.
Transmeta was the closest, since Linus worked for them way back when.
Given that the native instruction set was top secret, I would say it was the least open source processor of all.
Given that T'Pau appears in two fourth season Enterprise episodes (I haven't seen them - they are still showing the third season down here), I would say that both the parent and grand-parent's theories are wrong.
> But is it POSIX? The government shouldn't be in > the business of dictating implementations, but > interfaces. Links?
No POSIX, sorry. In fact, my microprocessor will probably never even have a C compiler for it so this wouldn't make much of a difference anyway. But it is not possible to make a really modern OS if you demand POSIX compatibility - you will just end up with a Unix with a very odd (and irrelevant) kernel or exo-kernel or whatever.
Compatibility with the Internet standard can get you pretty far these days, no matter what your computer looks like on the inside.
My company's link is included in the header of all my posts, right? But though even more outdated, my pages from the 1990s were better organized. Yuck - the main page is trashed:-(
I have been keeping track of the "PC Conectado" project as closely as I can, though I haven't heard much lately. The original specification from last year was very strict: 2.4GHz Celeron processor (AMD need not apply?), 128MB of memory, disk drive, CD-ROM drive, keyboard and a 15-inch monitor. The software was supposed to be Linux and a list of some 10 application programs.
What the article was implying is that this might change so a reduced version of Windows could be offered as an alternative. Perhaps there is some relation to AMD's PIC project which uses WinCE?
The real distinguishing feature of the machines sold through this program is that the government will help out each buyer who can prove he is poor with $70 (they have set aside a fund for one million buyers). So it isn't like people can't choose Windows, just that it will cost them an extra $70 beyond whatever Microsoft adds to the bargain.
My own "computer for the poor" project will be hurt just as much as MS by this since though my OS is Free Software it isn't Linux. Not nice at all, but hardly the end of the world.
BTW, 70% of the computers sold down here are the so called "gray boxes". So the largest manufacturer of totally legal machines, Positivo, only made 100K machines in 2004. Getting 1M will require convencing at least several different manufacturers. Actually, these would rather have a tax break since 40% of the price of computers to the end users is taxes (which the "Frankenstein" machines mostly avoid). But the government argued that such a move would help rich people as well and they must make sure that only the poor participate.
Interesting that you and the previous poster got into computing around 1985 or so. You missed the dreaded "don't forget ctrl-C before removing the floppy or you will ruin it" curse in CP/M. Fixing this was the one great improvement MS-DOS 1.0 had over the older system.
This was done by changing all disk operations from "write back" to "write through" (to use cache terms). Unfortunately, the cost is a reduction of several times in disk performance as the head constantly moves back and forth between the middle of the floppy (where the file is being accessed) and the first tracks (where the FATs are). This allows you to yank out the disk at any time with a very low probability of damage and also makes it likely that you will still have your data after a power failure.
For the floppy-only 128KB original Mac it is likely that this loss in performance would have been unacceptable. So Apple selected a "write back" scheme and prevented you from removing the disk without telling the software first so it could save all of its buffers. For the rarer case of a power failure the file system stored redundant information which the built-in disk repair utility could use to make up for any unsaved data.
I am not a North American, if that is what you meant (I am from Brazil).
QNX was my desktop OS from 1985 to 1988, but I have been watching it closely since so...
... I have a pretty good idea of how much it is used in the various markets.
I would have loved to use QNX again when I needed a PC desktop OS in 1994, but due to cost / source access / community size issues I adopted the inferior Linux instead.
As recently as 2001 I have helped out companies with QNX-based projects.
I'll admit that QNX has been far more popular than any of the others in my list, but I still stand by my statement that it will ultimately be just a footnote.
It's a hypothetical exercise to ask whether they should have written their own OS and designed their own chips. It's sort of a retroactive attempt to kill the goose laying the golden industry. Sure, they could have done it, but no, it wouldn't have helped them, and certainly not us.
Exactly! This is like all those silly articles that tell us that Linus would be extremely rich if he had sold Linux instead of giving it away. No he wouldn't - Linux would just have been a forgotten footnote in the history of computing along with Coherent, Idris, Unos, Regulos, Uniflex, OS-9, QNX, Plan 9, etc...
You can't change one aspect of history and expect all the rest to remain the same.
Exactly! And it is hard to learn math in a vacuum, yet it is what is being attempted.
Consider that in both the foreign language and math classes the kids:
don't have any use for the stuff other than getting grades
don't see their parents using it, or even the teacher except as needed by the class. And asking the parents for help with the homework makes this impression stronger because they seem to have forgotten absolutely everything about it
So Papert suggested that we create a "Mathland" where these things wouldn't happen. He did a great job with Logo, but sadly many people didn't understand the ideas and perverted it into the world's most boring drawing program.
I got to visit several mainframe "computer centers" in the early 1980s and typically about 30% of the staff were women, both operators and programmers. As another poster mentioned, nearly all of the "key punchs" were women. I had a friend who was studying computer science in one of the top schools in the country (Brazil) and women made up over 60% of his class.
By the mid 1980s, however, the drastic decline in female participation was quite visible. It is interesting to speculate why this is and what, if anything, should be done about it.
My point is that however strange these pictures might seem today, they were actually representative of the time when this book was written.
For some "Math is hard". Just as there are people who in this day and age can't read. We all aren't wired the same, but our learning methods sure act as though we are.
As Seymour Papert (one of the inventors of the Logo programming language) likes to point out, when French was regularly taught in USA schools hardly anybody learned it. Yet a few did, probably due to being "wired differently" as you put it. The interesting thing, however, is that in France 100% of the children learned it perfectly. So it is obvious that the learning method can make a difference.
When kids can't read or do math, how can we be sure that with some very different system they wouldn't have learned perfectly?
Don't think of a fancy scientific 3D Java applets (though those can be nice). Think of the lowly spreadsheet. Like Visicalc and Lotus 123 which were the reason businesses started using micros in the first place.
Now go to a modern business and watch people using Excel. Very pretty tables and the numbers are automatically converted into pie charts to be included in PowerPoint presentations. But if you change one of the numbers, nothing happens! It is not a living simulation of how their business is doing, but a dead approximation of a paper report.
Some people are using spreadsheets properly and they are getting good value out of their computers. Most aren't and don't even know what they are missing.
Ended up taking about 1.5 gal in the tractor to grow enough corn to produce 1 gal of alcohol.
If you use corn you do get these negative results, but here in Brazil we use sugar cane. The alcohol program, started in the 1970s, produced millions of cars (many of which are still running) until a shortage in the early 1990s scare most consumers away. It is making a major comeback since the introduction of "flex power" cars about a year ago. These work with either gasoline or pure alcohol so the buyer doesnt have to worry about future supply problems.
At about $0.23 per liter (multiply by 4 for gallons) vs $0.57 for gasoline, alcohol is the current choice for everyone who can use it here even with up to a 20% loss in mileage.
Starting the car in very cold days has proved to be the only real problem in nearly three decades of continous use. This isnt a big worry in Brazil, but probably would be in other countries.
Slashdot Mirror
If hundreds of millions of children actually do get Linux computers in the next few years it will probably change things significantly.
This was spelled out in the writer's guide that they would give to people who wanted to write episodes for TOS. The phaser was a three piece modular system. The basic phaser was a small box roughly the size and shape of the communicator and was to be used concealed in missions where they didn't expect to need weapons. If that wasn't enough then you would mount the basic phaser on a pistol grip which made aiming easier and increased firepower by adding its own power supply. Take a careful look at a picture of the pistol model and you will see what you thought was a communicator attached to the top of it. The third option was a rifle style piece to which you attached the pistol version and again increased aim and power.
Just another step in the well known Wheel of Reincarnation. At least well known to all three of us who don't completely ignore computer history ;-)
Since we are being picky, the 68000 actually had two simple 16 bit ALUs and one complex 16 bit ALU. The first was directly coupled to the high bits of the data and address registers and the second to the low bits of the address registers. The complex ALU was tightly coupled to the low bits of the data registers. The middle section of the processor (low address bits) could be connected to the other two when needed, but it was possible for the processor to handle 48 bits at a time in some common situations!
See "Microprogrammed Implementation of a Single Chip Microprocessor" by Skip Stritter and Nick Tredennick in SIGMICRO Newsletter Volume 9, Number 4, December 1978.
Given that we haven't heard more about this since then I am not taking it too seriously. Here in Brazil we have lots of similar articles where some government official is interviewed and claims "Brazil rejects OLPC". But that official is not involved in the decision process and those who are continue to be very positive about it.
Damn, I hate the academic world for letting Microsoft create and patent Singularity! Why didn't MIT, CMU or Berkeley do this research first and put the results in the public domain!
_ Systems)) is a great example of prior art, if that is what you are worried about.
While it was a commercial product, Burroughs' Master Control Program (see http://en.wikipedia.org/wiki/MCP_(Burroughs_Large
In one of their papers they explicitly contrast the adiabatic approach (which seeks the lowest possible power per operation at the cost of slower operations) with their own solution (which seeks to lower power even while operating at the highest possible speeds). So for the best possible battery life you would want adiabatic logic, but for multiple gigahertz operation Multigig seems like a nice option.
I am pretty sure the QNX microkernel took up less than 1KB unless you counted the interrupt table and stubs. But I am thinking of the pre 2.X versions which didn't include networking and distributed computing. Even with these extensions my guess is that the microkernel was still less than 3KB.
It would be interesting to know what the size of recent versions is.
You guessed wrong:
http://www.apple.com/trailers/disney/rovingmars/
I don't know why some people think Self is a very different language than Smalltalk-80. With the optional parser and GNU Smalltalk classes you can even file in Smalltalk-80 code and run it.
http://research.sun.com/self/papers/smalltalk.pdf
I don't remember who said "Self is like Smalltalk, only more so" but that is a great definition. To avoid having to repeat this discussion all the time I have renamed my Self/R project as Neo Smalltalk.
Mr. "don't mode me in!" is also the guy who invented the cut/copy/paste editing where you always replace the selection that is used by everybody these days.
How much is a guy like that worth?
There are two different services in that web site: hosting video which can be downloaded and searching through external video (this has been available for some time now). There isn't much local content yet so you are unlikely to run across it in a random search. Online videos have a little triangle beside the title that you can click on to watch the movie. Try one of their suggested searches, like "capoeira" and you will see that they haven't taken anything offline.
When are open-sourced based CPUs going to be available? Does anyone know of any available? I don't *feel* free with the current processor offerings available.
Here you go. Check out the OpenRISC 1000 - I am guessing that it will be particularly interesting for you since the 1200 version has been used to demonstrate Linux. Of course the MIPS and Sparc clones can do so as well.
Transmeta was the closest, since Linus worked for them way back when.
Given that the native instruction set was top secret, I would say it was the least open source processor of all.
Given that T'Pau appears in two fourth season Enterprise episodes (I haven't seen them - they are still showing the third season down here), I would say that both the parent and grand-parent's theories are wrong.
> But is it POSIX? The government shouldn't be in
:-(
> the business of dictating implementations, but
> interfaces. Links?
No POSIX, sorry. In fact, my microprocessor will probably never even have a C compiler for it so this wouldn't make much of a difference anyway. But it is not possible to make a really modern OS if you demand POSIX compatibility - you will just end up with a Unix with a very odd (and irrelevant) kernel or exo-kernel or whatever.
Compatibility with the Internet standard can get you pretty far these days, no matter what your computer looks like on the inside.
My company's link is included in the header of all my posts, right? But though even more outdated, my pages from the 1990s were better organized. Yuck - the main page is trashed
I have been keeping track of the "PC Conectado" project as closely as I can, though I haven't heard much lately. The original specification from last year was very strict: 2.4GHz Celeron processor (AMD need not apply?), 128MB of memory, disk drive, CD-ROM drive, keyboard and a 15-inch monitor. The software was supposed to be Linux and a list of some 10 application programs.
What the article was implying is that this might change so a reduced version of Windows could be offered as an alternative. Perhaps there is some relation to AMD's PIC project which uses WinCE?
The real distinguishing feature of the machines sold through this program is that the government will help out each buyer who can prove he is poor with $70 (they have set aside a fund for one million buyers). So it isn't like people can't choose Windows, just that it will cost them an extra $70 beyond whatever Microsoft adds to the bargain.
My own "computer for the poor" project will be hurt just as much as MS by this since though my OS is Free Software it isn't Linux. Not nice at all, but hardly the end of the world.
BTW, 70% of the computers sold down here are the so called "gray boxes". So the largest manufacturer of totally legal machines, Positivo, only made 100K machines in 2004. Getting 1M will require convencing at least several different manufacturers. Actually, these would rather have a tax break since 40% of the price of computers to the end users is taxes (which the "Frankenstein" machines mostly avoid). But the government argued that such a move would help rich people as well and they must make sure that only the poor participate.
Interesting that you and the previous poster got into computing around 1985 or so. You missed the dreaded "don't forget ctrl-C before removing the floppy or you will ruin it" curse in CP/M. Fixing this was the one great improvement MS-DOS 1.0 had over the older system.
This was done by changing all disk operations from "write back" to "write through" (to use cache terms). Unfortunately, the cost is a reduction of several times in disk performance as the head constantly moves back and forth between the middle of the floppy (where the file is being accessed) and the first tracks (where the FATs are). This allows you to yank out the disk at any time with a very low probability of damage and also makes it likely that you will still have your data after a power failure.
For the floppy-only 128KB original Mac it is likely that this loss in performance would have been unacceptable. So Apple selected a "write back" scheme and prevented you from removing the disk without telling the software first so it could save all of its buffers. For the rarer case of a power failure the file system stored redundant information which the built-in disk repair utility could use to make up for any unsaved data.
There was also one laptop made, but Jef himself has it and he's not giving it up (or at least wasn't when I asked him about it a few years ago).
/ index.html
I wouldn't give it up either, if I were him. But it seems he is willing to share a picture and loan a prototype: http://www.digibarn.org/collections/systems/swyft
I'll admit that QNX has been far more popular than any of the others in my list, but I still stand by my statement that it will ultimately be just a footnote.
It's a hypothetical exercise to ask whether they should have written their own OS and designed their own chips. It's sort of a retroactive attempt to kill the goose laying the golden industry. Sure, they could have done it, but no, it wouldn't have helped them, and certainly not us.
Exactly! This is like all those silly articles that tell us that Linus would be extremely rich if he had sold Linux instead of giving it away. No he wouldn't - Linux would just have been a forgotten footnote in the history of computing along with Coherent, Idris, Unos, Regulos, Uniflex, OS-9, QNX, Plan 9, etc...
You can't change one aspect of history and expect all the rest to remain the same.
Exactly! And it is hard to learn math in a vacuum, yet it is what is being attempted.
Consider that in both the foreign language and math classes the kids:
So Papert suggested that we create a "Mathland" where these things wouldn't happen. He did a great job with Logo, but sadly many people didn't understand the ideas and perverted it into the world's most boring drawing program.
I got to visit several mainframe "computer centers" in the early 1980s and typically about 30% of the staff were women, both operators and programmers. As another poster mentioned, nearly all of the "key punchs" were women. I had a friend who was studying computer science in one of the top schools in the country (Brazil) and women made up over 60% of his class.
By the mid 1980s, however, the drastic decline in female participation was quite visible. It is interesting to speculate why this is and what, if anything, should be done about it.
My point is that however strange these pictures might seem today, they were actually representative of the time when this book was written.
For some "Math is hard". Just as there are people who in this day and age can't read. We all aren't wired the same, but our learning methods sure act as though we are.
As Seymour Papert (one of the inventors of the Logo programming language) likes to point out, when French was regularly taught in USA schools hardly anybody learned it. Yet a few did, probably due to being "wired differently" as you put it. The interesting thing, however, is that in France 100% of the children learned it perfectly. So it is obvious that the learning method can make a difference.
When kids can't read or do math, how can we be sure that with some very different system they wouldn't have learned perfectly?
"Like what?" you and others are asking.
Alan answered this in the interview: simulation.
Don't think of a fancy scientific 3D Java applets (though those can be nice). Think of the lowly spreadsheet. Like Visicalc and Lotus 123 which were the reason businesses started using micros in the first place.
Now go to a modern business and watch people using Excel. Very pretty tables and the numbers are automatically converted into pie charts to be included in PowerPoint presentations. But if you change one of the numbers, nothing happens! It is not a living simulation of how their business is doing, but a dead approximation of a paper report.
Some people are using spreadsheets properly and they are getting good value out of their computers. Most aren't and don't even know what they are missing.
Ended up taking about 1.5 gal in the tractor to grow enough corn to produce 1 gal of alcohol.
If you use corn you do get these negative results, but here in Brazil we use sugar cane. The alcohol program, started in the 1970s, produced millions of cars (many of which are still running) until a shortage in the early 1990s scare most consumers away. It is making a major comeback since the introduction of "flex power" cars about a year ago. These work with either gasoline or pure alcohol so the buyer doesnt have to worry about future supply problems.
At about $0.23 per liter (multiply by 4 for gallons) vs $0.57 for gasoline, alcohol is the current choice for everyone who can use it here even with up to a 20% loss in mileage.
Starting the car in very cold days has proved to be the only real problem in nearly three decades of continous use. This isnt a big worry in Brazil, but probably would be in other countries.