It's not that older programmers don't understand OO principles. It's precisely because we do understand them that we have contempt for them.
OO development accomplishes three things:
It makes the program bigger
It makes the program run slower
It makes the program modules more portable, so less programming labor is needed overall
Often OO is simply the wrong way to do things, just as spaghetti code is sometimes far and away the most efficient way to implement an algorithm. When OO doesn't get in the way it's beautiful, but it's easy to lose track of the performance drag all that double-dereferencing causes.
Your older programmer probably started out with a language with weak or no typing (can you say FORTRAN?) on machines where every cycle and every word was a precious resource. He knows that an application of skill to the problem can reduce waste and improve performance, and he's used to doing that. Asking him to willfully waste resources so that the next guy can waste some more by reusing his not-so-optimal code is like asking someone who has known starvation to throw away food.
I am the one and only programmer for a distributor, as opposed to manufacturer, in an industry where very few distributors bother to have in-house software development. As such I have, without really trying, become as close to unfireable as you ever get, which is one reason I hang around the place even though they don't pay as well as larger shops.
One day one of my coworkers was doing a service job and discussing the computer system I'd designed with an employee of the customer. The employee had just finished taking a management course and asked my coworker if there was any backup for me. To which he replied, no, our guy's pretty unique.
The customer employee then demonstrated his grasp of management principles by saying that he'd just been taught that if you have anyone like me on the payroll, you should fire them at once! Sure it will hurt for awhile, but eventually you'll recover and you won't be at their mercy.
So that's what it's about, boys and girls: POWER. We do stuff they don't understand and it scares the shit out of them. The only way they can feel secure is to be sure we can be instantly replaced. Fifteen years of loyalty? Meaningless. Skill and experience? Meaningless. Modern management teaches that the most important thing is staying in control.
Fortunately, my current employer is as out-of-date as I am, and doesn't feel that way. Which is another reason I hang around even though the pay isn't so great.
Once upon a time there was this guy named Linus who had a Better Idea, which was to take this well tested OS built for mainframes but generally thought too complex for the desktop and put it on the desktop. And lo, Linus succeeded (with a lot of help from a lot of his friends), and we have Linux.
We also have desktop machines that are much more powerful than the mainframes on which Unix was first driven. Some of the space-saving strategies that seemed like a Good Idea (tm) then now seem like really Bad Ideas (tm) now. Kinda like what happened to a certain closed-source software company we could mention.
I would also mention that an associate of mine who gets front-updates on what That Closed Source Software Company (tm) is doing has told me that their.NET strategy includes a number of actual intelligent moves. One is the removal of the Variant type from the Visual Studio suite. And if that doesn't seem like a good enough idea to you, then the next will really floor you -- applications will keep their own.DLL's in their own application directories. Just like in the DOS days, you will be able to blow an app completely off your machine by deleting its directory, and version differences will become irrelevant.
Sounds like a good idea to me. Maybe youse linux guys should like thinka implementing something like this before the Redmond Juggernaut gets up another head o' steam.
The Iridium sats are in Low Earth Orbit so that they can get sufficient signal strength for a handheld device to connect. This means that the arctic regions are covered, and also that the network needs a lot of satellites -- which it has. Ping time is probably not good compared to landlinks but is nowhere near the delay you'd expect from geosynchronous.
In 1969 we had a highly charismatic president say "Put men on the Moon.
That was 1963. In 1969, Richard Nixon was Prez and I don't know anybody who would describe him as "highly charismatic" or even as a believer in space exploration.
I meant it to translate "Do you want to sleep with whom tonight?" and I think I got it right, though I admit it's been 20 years since I studied the language and can barely get out je ne parle pas le francais today.
There was an Avengers episode in which whatshisname was strapped in and forced to play an arcade driving game in which he got shocked at increasing voltage whenever he ran off the road. (The game itself was quite farfetched for its day, but hopelessly primitive by today's standards.)
The human body has a fairly high electrical resistance. According to the article the current is limited to 16 mA, which is safe. It's at 20mA and above where a badly routed shock starts to seriously scramble your nervous system.
The voltage it takes to push that 20mA through your skin can vary wildly. Human beings do not conduct electricity very well, and if you play with an ohmmeter you will be able to get skin resistances ranging from a few K-ohms (especially with contacts placed near one another on damp skin) to megohms (with distant contacts on dry skin). It is theoretically possible to electrocute yourself with voltages as low as 40VDC, but it's not easy to do and requires very good electrode placement, skin conductivity, and a weak heart. I am assuming that the game controller uses voltages lower than this so that no matter where you put the electrodes you can't pass enough current through a sensitive enough part of your body to do serious mischief.
I'm sure they wouldn't dare sell this for such a trivial purpose if it weren't lawyer-proof. They must consider people who will put the electrodes on the wrong part of their body for whatever reason. (I predict this will be a hot seller to the S&M folks.)
According to the article the controller uses voltages which aren't painful, so it's not like electroshock learning, where the whole point of getting shocked is that you want to avoid it happening again because it hurts.
The experience of friends who have used TENS and chiropractic electromassage equipment is that this kind of current will leave marks on the skin (and might not work at all) unless you use some fairly messy cream to ensure a good electrical connection.
Also, it seems like the new method of cheating in those fully immersive massively-multiplayer environments will be simple -- don't put the electrodes on.
I think you're talking about Voulez-vous coucher avec moi ce soir.
Of course, you'll see in the post that I correctly quote the first line of Lady Marmalade. I changed the French me (moi) to who (qui).
The line is also interesting for something that can't be done in English. The you in Do you is vous, which is the formal form of address used to address authority figures or strangers. Lovers, friends, or parents talking to children would use tu or toi. Of course a prostitute soliciting business would use vous but coucher, which does imply you will be doing a bit more than sleeping, is exactly the kind of activity you'd normally hear associated with tu. It's probably not even subtle to someone born speaking French but after taking the language in high school I always thought it was a really neat juxtaposition.
Yes, it is, but if you want to get a copy off Napster you will have much better success searching for the first line than the title, because most people don't know the title. Oh wait, it's probably filtered now anyway.
by the group Labelle, not Patti's later solo project.
And this changes the fact that it was actually sung by Patti Labelle how?
I don't know squat about Moulin Rouge the movie but I do now it spawned a truly awful remake of Patti LaBelle's classic disco tune Voulez-Vous Coucher Avec Moi which, for the linguistically and culturally challenged, is French for "Do you want to sleep with me tonight?" The original was one of the better examples of music made back when actual singing and skilled play of musical instruments were involved. LaBelle leered and teased her way through lyrics that said just enough and a tune that alternated between excitement and dark sleaze.
The new version, wherein down in old New Orleans becomes down in old Moulin Rouge, doesn't work. You don't get the feeling that the singers know the meanings of the words they are singing. (All right, I know it's French but it's not that hard to figure out.) The inflections are all wrong. Patti LaBelle managed to sound like a hooker making a come-on when it was appropriate; the new version sounds more like it's trying to sound different from the old one than like it's supposed to make any sense.
This remake also led to one of the most absolutely unbelievable moments in interview history, wherein Patti LaBelle herself claimed that she didn't know what the song's title meant when she recorded it. Uh-huh. Ri-ight.
I was gonna skip this one just because of how they raped the theme song, but now I might have to catch it to see the CGI. Damn.
Is it my imagination, or did we just /. Stanford?
on
The Social Web
·
· Score: 2
While the main page loaded, I can't actually get through to any of the 1,200 individual webpages which are part of the Social Web.
The problem with his line of reasoning is how complex the damn processors are nowadays.
You miss the point. You are looking at the micro scale; I am looking at a higher level of abstraction -- though not much higher.
I've looked at quite a bit of object code over the years and I am still looking for this compiler that is anywhere near as efficient as a human -- not at stuff like loading the pipelines, but at stuff like figuring out when to use register vs. memory variables and which registers to use (especially important in x86 architecture). Pipeline loading inefficiency is nothing compared to the fact that you used a 16-bit integer variable and did error checking on the subtract operation when the CX register and LOOP would have done the same thing. It is possible to instruct modern compilers to make very tight code but it's almost as much effort as writing the Assembly yourself and very few people know how to do it (or even know that it can be done, or might be worth doing).
Most projects don't warrant being done entirely in Assembly -- my project had to because it was piggyback on a proprietary embedded system and there was no OS and no reasonably fast higher level language to draw on. (Still, about 2,000 of those lines are in the controller's miserably slow BASIC; no use coding user interfaces that don't have to be fast anyway, etc.)
In a more reasonable environment my experience has always been about 80/20, that is 80% of the code in any old high-level language that's available plus 20% really fast assembly, is about as good as 100% assembly. Unfortunately, in Windoze it is a blazing pain in the posterior to use assembly at all.
Which is unfortunate, because I've found that most projects have a few things that happen so often that not doing them in Assembly is really stupid. The engineers on the device I hacked said they had done some statistical analyses and found a few routines that were called so often they bore looking into. One of them was named EVALUATE_KEYWORD. No duh. You don't code the inner loop of an interpreted language in C++; it's only a page or two of Assembly and that will literally give you an order of magnitude improvement. It also helps if you don't use double-precision floating point math throughout regardless of whether it's necessary.
You seem to be concentrating on their mistakes while ignoring their fabulous successes.
I am more than aware of their successes sonny. In fact right now I am going through the fabulous Apollo Lunar Surface Journal, a great record of a time when NASA managed to do a bunch of stuff I'm sure it couldn't do today with all the funding in the world.
Most of what NASA has done since Apollo has been crap.
We only lost one Shuttle -- of four we could afford to build. It was the wrong vehicle for no particular job at all, tries to be everything and isn't very good at anything. With the same money spent on proven Big Dumb Boosters like the SV we'd have had a space station by 1985 and still be making trips to Luna.
The early planetary probes were incredible, holding together far beyond their life expectancies (even if they did leave the parking brakes on in the Pioneers, that was a joke). But we almost lost Pathfinder because of a computer glitch, and then had the string of failures I mentioned in my first post. Again, space travel is about the QC, but the QC has been supplemented by overachieving goals, tight budgets, and an unwillingness to let schedules slip so things can be done right. "Faster, Better, Cheaper" should have been named "Faster, Better, Cheaper -- Pick Two." Haste and thrift are a poor combination.
This is not a troll. I really believe in space exploration; I am sick to this day that Apollo was killed as ignominiously as it was, and that we are stuck with an albatross like the Shuttle and planetary probes whose computers crash, bearings seize, transmitters fail, and can't even manage to navigate past the planet without hitting it. We've had some successes since Apollo, like the Voyagers, Pathfinder, and NEAR. But even Galileo is crippled. I'm still waiting to hear what will fall off of Cassini before it reaches Saturn.
In 1969 we mobilized a force of about 1/2 million people to put two men on the Moon. We could probably do that (right) with today's technology with 50,000 or 100,000 people, a relatively small proportion of our economy -- and support more and more frequent missions. But we are trying to do it with more like 50 or 100 people. All the people who worked on NEAR could fit in a modest banquet hall. The plain fact is that we don't have the commitment any more; NASA knows this and is running scared. And when you run scared, you're looking over your shoulder instead of at your feet. That's why you trip.
(without an atmosphere the moon is a VERY efficient place for solar power and harvesting)
Except that you will also have to store 14 days worth of power somehow to get through the night. We really don't have any efficient ways of doing that yet.
Everything fielded by NASA is really from private industry, let out by contracts. NASA does the specification and QC and flight prep and flight but they don't actually manufacture anything. So this really doesn't change anything; it's like saying Apollo 13 was really a North American Rockwell problem because they built the Command Module.
Let's see, first they ran Mars Climate Orbiter into its destination. They lost Mars Polar Explorer and can't even find it with Mars Global Surveyor (which had a few odd problems of its own; one of the solar panels had a movement problem which radically lengthened the aerobraking stage of the mission). They dissed Tito, and by extension anyone else who wants casual space tourism to become a reality. The Canadarm has a bad bearing. And now the rocket blew up their fancy new airplane.
Maybe that casual space tourism thing should wait until we have a better... but then, he didn't take the shuttle up he got some people who knew what they were doing to do the job.
And nobody seems to have interest in assembly language.
Too true. It's not really that much harder to code a sizeable project in Assembly, if you have a bit of discipline about it. The hardest thing is defining your data structures and sticking to them. You build the "language" as you need it in the form of purpose-built subroutines. I just got finished doing a 12,000 line project for an embedded controller. I socially engineered the manufacturer into giving me a few hooks -- which added about 2 pages to the firmware source code I'm told -- and then supplemented their well-designed but miserably slow BASIC variant with blazingly fast background Assembly. Now this gizmo which can only execute 100 lines or so of BASIC per second is weighing and sorting 150 pieces per minute, doing true weight conversions at 60Hz (the firmware only manages 10Hz due to the use of floating-point math) and doing accurate 60Hz timing which BASIC cannot do.
As one of the engineers said after seeing the video, "Well, I guess you've been telling us this was possible since 1995."
This controller uses a 20 MHz 80186. Its replacement introduced last year uses a 40MHz 80386DX, and my code still runs an order of magnitude faster on the old hardware than BASIC code does on the new board. Another thing I told them back in 1995 -- you may spend a lot of time to write software, but you only have to write it once. When you up the CPU specification it increases the cost of every unit you produce.
You could go to B. Dalton and slam down twelve buks and get a manual which told you everything you needed to know about the C64 to write professional grade applications -- every hardware register (and there were a lot of them for a machine of its era), the entire memory map, every plug and socket pinout, details of the 1541 disk drive interface, and quite a few useful ROM hooks.
Meanwhile, back in the present, I've probably spent more than a thousand bucks on software and PC documentation over the last 10 years without ever getting anything that resembles a complete description of the hardware.
I have hated Microsoft since a day when few people had even heard of it -- before IBM introduced the PC, BillG was chugging out slow, crappy, bug-ridden BASIC interpreters which were the mainstay of home computer do-it-yourself design in the late 1970's, and like many self-taught programmers I was learning assembly language to get around the limitations of those systems.
MicroSlothed has continued to chug out slow, crappy, bug-ridden products that never quite get finished ever since -- a text-based operating system that was almost right by version 6.11, then abandoned; a 16-bit windowing system that was almost finished at 3.11; a 32-over-16 system that was almost finished at 98 SE; and a secure corporate system that was almost useful by 4.0, now also abandoned. The jury is still out on 2000 but I'm not holding my breath; this zebra's stripes are engraved.
So why is BillG one of the 1.5 richest people in the world? Because he got in the game early. First post, not best post.
Meatspace moderation isn't measured in karma, it's measured in dollars -- and it isn't capped.
Slashdot Mirror
OO development accomplishes three things:
- It makes the program bigger
- It makes the program run slower
- It makes the program modules more portable, so less programming labor is needed overall
Often OO is simply the wrong way to do things, just as spaghetti code is sometimes far and away the most efficient way to implement an algorithm. When OO doesn't get in the way it's beautiful, but it's easy to lose track of the performance drag all that double-dereferencing causes.Your older programmer probably started out with a language with weak or no typing (can you say FORTRAN?) on machines where every cycle and every word was a precious resource. He knows that an application of skill to the problem can reduce waste and improve performance, and he's used to doing that. Asking him to willfully waste resources so that the next guy can waste some more by reusing his not-so-optimal code is like asking someone who has known starvation to throw away food.
One day one of my coworkers was doing a service job and discussing the computer system I'd designed with an employee of the customer. The employee had just finished taking a management course and asked my coworker if there was any backup for me. To which he replied, no, our guy's pretty unique.
The customer employee then demonstrated his grasp of management principles by saying that he'd just been taught that if you have anyone like me on the payroll, you should fire them at once! Sure it will hurt for awhile, but eventually you'll recover and you won't be at their mercy.
So that's what it's about, boys and girls: POWER. We do stuff they don't understand and it scares the shit out of them. The only way they can feel secure is to be sure we can be instantly replaced. Fifteen years of loyalty? Meaningless. Skill and experience? Meaningless. Modern management teaches that the most important thing is staying in control.
Fortunately, my current employer is as out-of-date as I am, and doesn't feel that way. Which is another reason I hang around even though the pay isn't so great.
We also have desktop machines that are much more powerful than the mainframes on which Unix was first driven. Some of the space-saving strategies that seemed like a Good Idea (tm) then now seem like really Bad Ideas (tm) now. Kinda like what happened to a certain closed-source software company we could mention.
I would also mention that an associate of mine who gets front-updates on what That Closed Source Software Company (tm) is doing has told me that their .NET strategy includes a number of actual intelligent moves. One is the removal of the Variant type from the Visual Studio suite. And if that doesn't seem like a good enough idea to you, then the next will really floor you -- applications will keep their own .DLL's in their own application directories. Just like in the DOS days, you will be able to blow an app completely off your machine by deleting its directory, and version differences will become irrelevant.
Sounds like a good idea to me. Maybe youse linux guys should like thinka implementing something like this before the Redmond Juggernaut gets up another head o' steam.
The Iridium sats are in Low Earth Orbit so that they can get sufficient signal strength for a handheld device to connect. This means that the arctic regions are covered, and also that the network needs a lot of satellites -- which it has. Ping time is probably not good compared to landlinks but is nowhere near the delay you'd expect from geosynchronous.
That was 1963. In 1969, Richard Nixon was Prez and I don't know anybody who would describe him as "highly charismatic" or even as a believer in space exploration.
Better?
At almost 150mm square and 50mm thick and weighing almsot 1 Kg, "rat" would seem more appropriate.
I meant it to translate "Do you want to sleep with whom tonight?" and I think I got it right, though I admit it's been 20 years since I studied the language and can barely get out je ne parle pas le francais today.
There was an Avengers episode in which whatshisname was strapped in and forced to play an arcade driving game in which he got shocked at increasing voltage whenever he ran off the road. (The game itself was quite farfetched for its day, but hopelessly primitive by today's standards.)
news://alt.binaries.pictures.erotica.torture
The voltage it takes to push that 20mA through your skin can vary wildly. Human beings do not conduct electricity very well, and if you play with an ohmmeter you will be able to get skin resistances ranging from a few K-ohms (especially with contacts placed near one another on damp skin) to megohms (with distant contacts on dry skin). It is theoretically possible to electrocute yourself with voltages as low as 40VDC, but it's not easy to do and requires very good electrode placement, skin conductivity, and a weak heart. I am assuming that the game controller uses voltages lower than this so that no matter where you put the electrodes you can't pass enough current through a sensitive enough part of your body to do serious mischief.
I'm sure they wouldn't dare sell this for such a trivial purpose if it weren't lawyer-proof. They must consider people who will put the electrodes on the wrong part of their body for whatever reason. (I predict this will be a hot seller to the S&M folks.)
The experience of friends who have used TENS and chiropractic electromassage equipment is that this kind of current will leave marks on the skin (and might not work at all) unless you use some fairly messy cream to ensure a good electrical connection.
Also, it seems like the new method of cheating in those fully immersive massively-multiplayer environments will be simple -- don't put the electrodes on.
Of course, you'll see in the post that I correctly quote the first line of Lady Marmalade. I changed the French me (moi) to who (qui).
The line is also interesting for something that can't be done in English. The you in Do you is vous, which is the formal form of address used to address authority figures or strangers. Lovers, friends, or parents talking to children would use tu or toi. Of course a prostitute soliciting business would use vous but coucher, which does imply you will be doing a bit more than sleeping, is exactly the kind of activity you'd normally hear associated with tu. It's probably not even subtle to someone born speaking French but after taking the language in high school I always thought it was a really neat juxtaposition.
Yes, it is, but if you want to get a copy off Napster you will have much better success searching for the first line than the title, because most people don't know the title. Oh wait, it's probably filtered now anyway.
by the group Labelle, not Patti's later solo project.
And this changes the fact that it was actually sung by Patti Labelle how?
Maybe it's an effect Katz has on people. Sheesh.
The new version, wherein down in old New Orleans becomes down in old Moulin Rouge, doesn't work. You don't get the feeling that the singers know the meanings of the words they are singing. (All right, I know it's French but it's not that hard to figure out.) The inflections are all wrong. Patti LaBelle managed to sound like a hooker making a come-on when it was appropriate; the new version sounds more like it's trying to sound different from the old one than like it's supposed to make any sense.
This remake also led to one of the most absolutely unbelievable moments in interview history, wherein Patti LaBelle herself claimed that she didn't know what the song's title meant when she recorded it. Uh-huh. Ri-ight.
I was gonna skip this one just because of how they raped the theme song, but now I might have to catch it to see the CGI. Damn.
While the main page loaded, I can't actually get through to any of the 1,200 individual webpages which are part of the Social Web.
You miss the point. You are looking at the micro scale; I am looking at a higher level of abstraction -- though not much higher.
I've looked at quite a bit of object code over the years and I am still looking for this compiler that is anywhere near as efficient as a human -- not at stuff like loading the pipelines, but at stuff like figuring out when to use register vs. memory variables and which registers to use (especially important in x86 architecture). Pipeline loading inefficiency is nothing compared to the fact that you used a 16-bit integer variable and did error checking on the subtract operation when the CX register and LOOP would have done the same thing. It is possible to instruct modern compilers to make very tight code but it's almost as much effort as writing the Assembly yourself and very few people know how to do it (or even know that it can be done, or might be worth doing).
Most projects don't warrant being done entirely in Assembly -- my project had to because it was piggyback on a proprietary embedded system and there was no OS and no reasonably fast higher level language to draw on. (Still, about 2,000 of those lines are in the controller's miserably slow BASIC; no use coding user interfaces that don't have to be fast anyway, etc.)
In a more reasonable environment my experience has always been about 80/20, that is 80% of the code in any old high-level language that's available plus 20% really fast assembly, is about as good as 100% assembly. Unfortunately, in Windoze it is a blazing pain in the posterior to use assembly at all.
Which is unfortunate, because I've found that most projects have a few things that happen so often that not doing them in Assembly is really stupid. The engineers on the device I hacked said they had done some statistical analyses and found a few routines that were called so often they bore looking into. One of them was named EVALUATE_KEYWORD. No duh. You don't code the inner loop of an interpreted language in C++; it's only a page or two of Assembly and that will literally give you an order of magnitude improvement. It also helps if you don't use double-precision floating point math throughout regardless of whether it's necessary.
I am more than aware of their successes sonny. In fact right now I am going through the fabulous Apollo Lunar Surface Journal, a great record of a time when NASA managed to do a bunch of stuff I'm sure it couldn't do today with all the funding in the world.
Most of what NASA has done since Apollo has been crap.
We only lost one Shuttle -- of four we could afford to build. It was the wrong vehicle for no particular job at all, tries to be everything and isn't very good at anything. With the same money spent on proven Big Dumb Boosters like the SV we'd have had a space station by 1985 and still be making trips to Luna.
The early planetary probes were incredible, holding together far beyond their life expectancies (even if they did leave the parking brakes on in the Pioneers, that was a joke). But we almost lost Pathfinder because of a computer glitch, and then had the string of failures I mentioned in my first post. Again, space travel is about the QC, but the QC has been supplemented by overachieving goals, tight budgets, and an unwillingness to let schedules slip so things can be done right. "Faster, Better, Cheaper" should have been named "Faster, Better, Cheaper -- Pick Two." Haste and thrift are a poor combination.
This is not a troll. I really believe in space exploration; I am sick to this day that Apollo was killed as ignominiously as it was, and that we are stuck with an albatross like the Shuttle and planetary probes whose computers crash, bearings seize, transmitters fail, and can't even manage to navigate past the planet without hitting it. We've had some successes since Apollo, like the Voyagers, Pathfinder, and NEAR. But even Galileo is crippled. I'm still waiting to hear what will fall off of Cassini before it reaches Saturn.
In 1969 we mobilized a force of about 1/2 million people to put two men on the Moon. We could probably do that (right) with today's technology with 50,000 or 100,000 people, a relatively small proportion of our economy -- and support more and more frequent missions. But we are trying to do it with more like 50 or 100 people. All the people who worked on NEAR could fit in a modest banquet hall. The plain fact is that we don't have the commitment any more; NASA knows this and is running scared. And when you run scared, you're looking over your shoulder instead of at your feet. That's why you trip.
Except that you will also have to store 14 days worth of power somehow to get through the night. We really don't have any efficient ways of doing that yet.
Everything fielded by NASA is really from private industry, let out by contracts. NASA does the specification and QC and flight prep and flight but they don't actually manufacture anything. So this really doesn't change anything; it's like saying Apollo 13 was really a North American Rockwell problem because they built the Command Module.
Maybe that casual space tourism thing should wait until we have a better ... but then, he didn't take the shuttle up he got some people who knew what they were doing to do the job.
Too true. It's not really that much harder to code a sizeable project in Assembly, if you have a bit of discipline about it. The hardest thing is defining your data structures and sticking to them. You build the "language" as you need it in the form of purpose-built subroutines. I just got finished doing a 12,000 line project for an embedded controller. I socially engineered the manufacturer into giving me a few hooks -- which added about 2 pages to the firmware source code I'm told -- and then supplemented their well-designed but miserably slow BASIC variant with blazingly fast background Assembly. Now this gizmo which can only execute 100 lines or so of BASIC per second is weighing and sorting 150 pieces per minute, doing true weight conversions at 60Hz (the firmware only manages 10Hz due to the use of floating-point math) and doing accurate 60Hz timing which BASIC cannot do.
As one of the engineers said after seeing the video, "Well, I guess you've been telling us this was possible since 1995."
This controller uses a 20 MHz 80186. Its replacement introduced last year uses a 40MHz 80386DX, and my code still runs an order of magnitude faster on the old hardware than BASIC code does on the new board. Another thing I told them back in 1995 -- you may spend a lot of time to write software, but you only have to write it once. When you up the CPU specification it increases the cost of every unit you produce.
Meanwhile, back in the present, I've probably spent more than a thousand bucks on software and PC documentation over the last 10 years without ever getting anything that resembles a complete description of the hardware.
Only 9 comments and the site is already slashdotted.
MicroSlothed has continued to chug out slow, crappy, bug-ridden products that never quite get finished ever since -- a text-based operating system that was almost right by version 6.11, then abandoned; a 16-bit windowing system that was almost finished at 3.11; a 32-over-16 system that was almost finished at 98 SE; and a secure corporate system that was almost useful by 4.0, now also abandoned. The jury is still out on 2000 but I'm not holding my breath; this zebra's stripes are engraved.
So why is BillG one of the 1.5 richest people in the world? Because he got in the game early. First post, not best post.
Meatspace moderation isn't measured in karma, it's measured in dollars -- and it isn't capped.