That's the going price for a state of the art fab.
Moto can't keep up. IBM doesn't have the userbase. Sun is just keeping their head above water. AMD is constantly hurt by price wars.
Who does that leave? Intel.
Intel is the only processor company that can afford a new cutting edge fab for processor production. That means the others are going to slowly drop out from here on.
(Side issue: since processor speed is becoming less important for many tasks, Intel's largest competitor in the future will likely be TI and Samsung making ARM processors...)
This is me. (X-Force is just a working title, it would be changed.)
Right now this is on the backburner. My main focus right now is Fantasy Realms 3 (something close to, and hopefully better than, Master of Magic).
Why?
Doing a X-Com clone on smartphones is out of the question, the screen is too small. Ditto for low-res Palm PDAs (160x160). I made a mock-up screen for Palm low-res a while back with only 3-level buildings, aliens, soldiers, etc, and everything was so scrunched that you couldn't tell things apart. The Palm 320x320 hi-res screens are better, but physically they are still pretty small. This may be possible on PPCs or Sony's 320x480 units because of the physically bigger screen and landscape mode, but the Sony units are a small niche, and I haven't done any PPC programming yet. Still, things like the amount of scrolling needed for a long shot are a problem. (A stylus is definitely not a mouse, and a control pad isn't nearly as good either.)
I'm likely to do something with the same "feel" as X-Force/X-Com in the near future, but overall these devices have too small of screens to do X-Com justice. (ie, it's possible, I just don't think a clone would end up working well.)
Hmmm, maybe I should have said that they aren't much different in theory instead?
I've got some experience in this, and I think that you are somewhat missing the forest for the trees. (Something that often happens to me when I'm too close to a problem.)
You've got a 3-6(+) month time lag if you need to increase production unexpectedly. That's a long time, but yet it's not a long time, if you know what I mean. Sales now are low. And you're keeping around extra people in case sales go back up (ie, extra capacity).
Many companies have already pruned out this extra capacity, with the two thoughts that they'll make just as much if a shortage happens which lets them increase prices, and also that there is no guarantee that they'll ever need this capacity amount again. ie, if the supply is at an oversupply situation, the next thing to go is the excess capacity, and then that goes all down the supply chain and they all reduce capacity. About the only time this doesn't happen is when it's believed that the drop in orders is a short, temporary thing (or wishful thinking, or if the employers are just nice guys with money to burn...). Basically, you always try to keep a positive cash flow if possible, so you don't pay for anything you don't need. (I can tell that you already know this, but this places the debate at a higher "forest" level.)
With a perfect supply chain, supply directly affects capacity, and vice versa. Of course, like you basically mention, things in the real world aren't perfect...
The reason that the employee credit is a "hidden tax" is because the only difference between it and welfare is basically forcing people to give free labor to companies in order to get their "welfare check". (It doesn't look like that on the surface though, does it?) Of course with GWB borrowing money like crazy, there wouldn't be much of a "tax" here anyway.
Capacity, supply, they're different, but not that much in practice. If you have a lot of excess capacity, you can fill any demand PDQ, so you still don't get an investment/employment increase.
An investment credit can also be a good idea, I personally just don't think that it's a good one right now.
As you say, we've got a temporary situation causing a lot of job loss. ("And in the long run, we're all dead" applies here.) My feeling is that to help fix this, we need to decrease the cost of labor by some amount. Thus an "employee credit" for a larger tax writeoff for the more employees a company hires. Of course this is really just another hidden tax, because someone must pay the bills. (Right? Err, that is right Mr. President? Isn't it? ahh...) The big thing is that it should increase the number of low-end wage earners, who spend most/all their money, which then increases demand. (Worst case it would likely slow down the rate of job loss.)
"And yes, you can rag capitalism, but the fact is: when you cut taxes (to a degree) you increase the amount of taxes collected."
It's a good thing you put "to a degree" there, because this is only true if the extra money is then invested (like you did), and creates more wealth that can then be taxed. Failing that situation, lowering taxes can actually be a bad thing.
Both taxes and investment redistribute wealth, which is nearly always a good thing for an economy. However investment is the better method because it gives money to those who create yet more wealth, and jobs. Note that lowering taxes != increasing investment unless there is a shortage of supply.
During the Reagan years there was a shortage in supply, so cutting taxes was a reasonable idea. Now however, there is a massive glut of supply, so the idea of cutting taxes for investors is crazy, because this money won't be reinvested. What's needed is something like a double tax credit for the first $20K for each employee hired. This would create a lot of low paying jobs that would help people get by a lot better than no jobs at all.
RISC was never about a reduced instruction count. This is just something repeated constantly be those without any computer architecture knowledge. RISC has always been about Reduced complexity Instruction Set Architecture. A number of the earlier papers/reports about it actually called it RCSIC.
A smaller number of instructions has never really gained anyone anything. However RISCs reduced complexity made many things much easier to implement.
I'm working on such a game. (I'm NOT going to post my website up here. The bandwidth fees would kill me...)
It's on Palm only. (I don't have a team, just me, and working on the Palm is limited enough that one person can do it.) At Handango (www.handango.com) look up Fantasy Realms 2, and then picture this with a map, spell & artifact research, a HOMM type building style, resources, multiple planes, etc. (It will be a minimum of 6 more months before this is out.)
The 2600 was nothing like a 400/800. The 5200 was the game machine conversion of the 400/800.
The Atari computers had several specialized chips (gtia, antic, pokey) for doing things like full display list control and 8/16 bit sound. They were probably the most complex general-purpose 8-bit computers sold back then.
CPLDs can use SRAM, and FPGAs can use flash. The difference is in how routing is done.
CPLDs have a pretty firm definition of having several large blocks, holding many macrocells, with feedback to a global (and often also local) switch matrix.
FPGAs don't have a firm definition, but Xilinx ones use SRAM based LUTs connected in a grid pattern with manhatten routing. Most of Altera's FPGAs have a more CPLD bent, while Actel uses antifuse and flash technology with more of a "sea of gates" look to the developer.
"True, however the core applications, such as synthesis or timing engines, are not written in TCL, they're written in C or C++."
Huh? I've never heard of *any* eda tool that was only written in tcl. All of them are using tcl as a scripting front-end to make calls into the c/c++ methods. It also makes hooking in tk easy for the gui.
To add more, perl, not tcl, is becoming the prefered high level scripting language for automation. tcl usually sits in the "middle" between the perl script and the c++ code.
"However, trying to write a real application in Tcl is an excercise in futility."
You must not know about the EDA industry then. Anymore practically everything uses tcl and tk. Some of these GUIs would be impressive as just a native Windows program, but yet they are portable.
Games? Linux? WiFi? Video? Heavy computation? Get a PPC. The Axims give the best price/performance of any device, but the new HP device is much smaller and portable. (The new Sony Palm devices are a close second here.)
PIM? Long battery life? Portability? Get a Palm. I suggest anymore the treo 90. (see below)
In between? Nothing really good here, best bets are probably a Tungsten T or the new thin HP PDA.
I'm personally using a treo 90 now. It's the first PDA I've owned that is small enough that I actually carry it everywhere, and the battery lasts long enough that it isn't tied to the recharger.
Closed source, sold software has a strong feedback loop. The developer puts it out, customers complain about it, developer makes it better, customers start buying and give more feedback, developer makes it better yet, more customers start buying, etc, etc.
Money creates a strong feedback loop, which creates a program that fits better with the demands of the customers. (It also lowers support costs. It breaks down in a monopoly situation...)
OSS doesn't have a strong feedback loop. That's why nearly all of the truely successful OSS projects have truely expert programmers, which somewhat make up for not having this feedback loop. Still, there isn't any significant pressure on them to make the product closer to what customers want, instead of what they want, leading to "usability" problems for others.
Palm still has fundamental market advantages compared to PPC. But Palm's marketing & sales divisions are so !!!!!!! bad that they may kill the company.
(I'm a Palm developer now looking at cross platform development because of this. Palm/PPC/Symbian.)
Palm is a standard. What firmware issues? In all honesty, why should these people care about OSS? (There is also a huge amount of free/OSS stuff out for Palms. What does a simputer have?) It's more rugged, but will likely still break if you drop it screen first. With 10 old Palms for this thing's price, they could play baseball with one, use another to prop up a desk leg, take two apart to look at the insides, and give one to the village klutz. And still have 5 left over...
Graffiti is a point, but aren't these suppose to be illiterates?
Used Palms go for $20-$50 on eBay. You could get possibly 10 Palm PDAs for the price of one of these. So can anyone tell me why this is better than flooding the country with a bunch of old and obsolete Palms?
"Oh yes, and the Tungstun slidely thing is silly. I'm going to have to spend my entire time opening and retracting it as
I use grafitti a lot - which means it'll break quickly."
Reportedly, this was tested for 100,000 open/close cycles. That doesn't sound fragile to me.
"32 meg is peanuts!"
That's why you have SD cards. 1G ones will be coming out soon.
This is wrong. You can compress truely random data. However you can't be guaranteed that any algorithm (or combination of algorithms) you use will give you a positive compression ratio in every single case.
Slashdot Mirror
Yah, but for how much longer?
$3B fabs.
That's the going price for a state of the art fab.
Moto can't keep up. IBM doesn't have the userbase. Sun is just keeping their head above water. AMD is constantly hurt by price wars.
Who does that leave? Intel.
Intel is the only processor company that can afford a new cutting edge fab for processor production. That means the others are going to slowly drop out from here on.
(Side issue: since processor speed is becoming less important for many tasks, Intel's largest competitor in the future will likely be TI and Samsung making ARM processors...)
Yup. The first ataris (400/800) had 4 joystick ports, and they ripped out 2 of them in the later models. (You Bastards!)
There were several early games like MULE that used all 4 joysticks, but these pretty much disappeared when the 800XLs were massed produced.
This is me. (X-Force is just a working title, it would be changed.)
Right now this is on the backburner. My main focus right now is Fantasy Realms 3 (something close to, and hopefully better than, Master of Magic).
Why?
Doing a X-Com clone on smartphones is out of the question, the screen is too small. Ditto for low-res Palm PDAs (160x160). I made a mock-up screen for Palm low-res a while back with only 3-level buildings, aliens, soldiers, etc, and everything was so scrunched that you couldn't tell things apart. The Palm 320x320 hi-res screens are better, but physically they are still pretty small. This may be possible on PPCs or Sony's 320x480 units because of the physically bigger screen and landscape mode, but the Sony units are a small niche, and I haven't done any PPC programming yet. Still, things like the amount of scrolling needed for a long shot are a problem. (A stylus is definitely not a mouse, and a control pad isn't nearly as good either.)
I'm likely to do something with the same "feel" as X-Force/X-Com in the near future, but overall these devices have too small of screens to do X-Com justice. (ie, it's possible, I just don't think a clone would end up working well.)
Hmmm, maybe I should have said that they aren't much different in theory instead?
I've got some experience in this, and I think that you are somewhat missing the forest for the trees. (Something that often happens to me when I'm too close to a problem.)
You've got a 3-6(+) month time lag if you need to increase production unexpectedly. That's a long time, but yet it's not a long time, if you know what I mean. Sales now are low. And you're keeping around extra people in case sales go back up (ie, extra capacity).
Many companies have already pruned out this extra capacity, with the two thoughts that they'll make just as much if a shortage happens which lets them increase prices, and also that there is no guarantee that they'll ever need this capacity amount again. ie, if the supply is at an oversupply situation, the next thing to go is the excess capacity, and then that goes all down the supply chain and they all reduce capacity. About the only time this doesn't happen is when it's believed that the drop in orders is a short, temporary thing (or wishful thinking, or if the employers are just nice guys with money to burn...). Basically, you always try to keep a positive cash flow if possible, so you don't pay for anything you don't need. (I can tell that you already know this, but this places the debate at a higher "forest" level.)
With a perfect supply chain, supply directly affects capacity, and vice versa. Of course, like you basically mention, things in the real world aren't perfect...
The reason that the employee credit is a "hidden tax" is because the only difference between it and welfare is basically forcing people to give free labor to companies in order to get their "welfare check". (It doesn't look like that on the surface though, does it?) Of course with GWB borrowing money like crazy, there wouldn't be much of a "tax" here anyway.
Capacity, supply, they're different, but not that much in practice. If you have a lot of excess capacity, you can fill any demand PDQ, so you still don't get an investment/employment increase.
An investment credit can also be a good idea, I personally just don't think that it's a good one right now.
As you say, we've got a temporary situation causing a lot of job loss. ("And in the long run, we're all dead" applies here.) My feeling is that to help fix this, we need to decrease the cost of labor by some amount. Thus an "employee credit" for a larger tax writeoff for the more employees a company hires. Of course this is really just another hidden tax, because someone must pay the bills. (Right? Err, that is right Mr. President? Isn't it? ahh...) The big thing is that it should increase the number of low-end wage earners, who spend most/all their money, which then increases demand. (Worst case it would likely slow down the rate of job loss.)
"And yes, you can rag capitalism, but the fact is: when you cut taxes (to a degree) you increase the amount of taxes collected."
It's a good thing you put "to a degree" there, because this is only true if the extra money is then invested (like you did), and creates more wealth that can then be taxed. Failing that situation, lowering taxes can actually be a bad thing.
Both taxes and investment redistribute wealth, which is nearly always a good thing for an economy. However investment is the better method because it gives money to those who create yet more wealth, and jobs. Note that lowering taxes != increasing investment unless there is a shortage of supply.
During the Reagan years there was a shortage in supply, so cutting taxes was a reasonable idea. Now however, there is a massive glut of supply, so the idea of cutting taxes for investors is crazy, because this money won't be reinvested. What's needed is something like a double tax credit for the first $20K for each employee hired. This would create a lot of low paying jobs that would help people get by a lot better than no jobs at all.
RISC was never about a reduced instruction count. This is just something repeated constantly be those without any computer architecture knowledge. RISC has always been about Reduced complexity Instruction Set Architecture. A number of the earlier papers/reports about it actually called it RCSIC.
A smaller number of instructions has never really gained anyone anything. However RISCs reduced complexity made many things much easier to implement.
I'm working on such a game. (I'm NOT going to post my website up here. The bandwidth fees would kill me...)
It's on Palm only. (I don't have a team, just me, and working on the Palm is limited enough that one person can do it.) At Handango (www.handango.com) look up Fantasy Realms 2, and then picture this with a map, spell & artifact research, a HOMM type building style, resources, multiple planes, etc. (It will be a minimum of 6 more months before this is out.)
The 2600 was nothing like a 400/800. The 5200 was the game machine conversion of the 400/800.
The Atari computers had several specialized chips (gtia, antic, pokey) for doing things like full display list control and 8/16 bit sound. They were probably the most complex general-purpose 8-bit computers sold back then.
CPLDs can use SRAM, and FPGAs can use flash. The difference is in how routing is done.
CPLDs have a pretty firm definition of having several large blocks, holding many macrocells, with feedback to a global (and often also local) switch matrix.
FPGAs don't have a firm definition, but Xilinx ones use SRAM based LUTs connected in a grid pattern with manhatten routing. Most of Altera's FPGAs have a more CPLD bent, while Actel uses antifuse and flash technology with more of a "sea of gates" look to the developer.
FPGAs are typically bigger, slower, and use a lot more power. The only advantage they give is flexibility.
"True, however the core applications, such as synthesis or timing engines, are not written in TCL, they're written in C or C++."
Huh? I've never heard of *any* eda tool that was only written in tcl. All of them are using tcl as a scripting front-end to make calls into the c/c++ methods. It also makes hooking in tk easy for the gui.
To add more, perl, not tcl, is becoming the prefered high level scripting language for automation. tcl usually sits in the "middle" between the perl script and the c++ code.
"However, trying to write a real application in Tcl is an excercise in futility."
You must not know about the EDA industry then. Anymore practically everything uses tcl and tk. Some of these GUIs would be impressive as just a native Windows program, but yet they are portable.
Depends on what you want to do.
Games? Linux? WiFi? Video? Heavy computation? Get a PPC. The Axims give the best price/performance of any device, but the new HP device is much smaller and portable. (The new Sony Palm devices are a close second here.)
PIM? Long battery life? Portability? Get a Palm. I suggest anymore the treo 90. (see below)
In between? Nothing really good here, best bets are probably a Tungsten T or the new thin HP PDA.
I'm personally using a treo 90 now. It's the first PDA I've owned that is small enough that I actually carry it everywhere, and the battery lasts long enough that it isn't tied to the recharger.
Real cheating is using the artifact cheat to put flamestrike X4 on every offensive item you make for free...
Closed source, sold software has a strong feedback loop. The developer puts it out, customers complain about it, developer makes it better, customers start buying and give more feedback, developer makes it better yet, more customers start buying, etc, etc.
Money creates a strong feedback loop, which creates a program that fits better with the demands of the customers. (It also lowers support costs. It breaks down in a monopoly situation...)
OSS doesn't have a strong feedback loop. That's why nearly all of the truely successful OSS projects have truely expert programmers, which somewhat make up for not having this feedback loop. Still, there isn't any significant pressure on them to make the product closer to what customers want, instead of what they want, leading to "usability" problems for others.
Palm still has fundamental market advantages compared to PPC. But Palm's marketing & sales divisions are so !!!!!!! bad that they may kill the company.
(I'm a Palm developer now looking at cross platform development because of this. Palm/PPC/Symbian.)
Palm is a standard. What firmware issues? In all honesty, why should these people care about OSS? (There is also a huge amount of free/OSS stuff out for Palms. What does a simputer have?) It's more rugged, but will likely still break if you drop it screen first. With 10 old Palms for this thing's price, they could play baseball with one, use another to prop up a desk leg, take two apart to look at the insides, and give one to the village klutz. And still have 5 left over...
Graffiti is a point, but aren't these suppose to be illiterates?
Used Palms go for $20-$50 on eBay. You could get possibly 10 Palm PDAs for the price of one of these. So can anyone tell me why this is better than flooding the country with a bunch of old and obsolete Palms?
Cost is the major factor.
Right now, PDA sized greyscale screens go for $5-$15, and color ones for $50-$100.
If all the other specs are decent, and they can cut the cost of the screens by half, then they have a winner.
"Oh yes, and the Tungstun slidely thing is silly. I'm going to have to spend my entire time opening and retracting it as
I use grafitti a lot - which means it'll break quickly."
Reportedly, this was tested for 100,000 open/close cycles. That doesn't sound fragile to me.
"32 meg is peanuts!"
That's why you have SD cards. 1G ones will be coming out soon.
Some of the latest Sparc chips have async parts that take more than one clock cycle to execute.
They probably use technology from lifting bodies.
This is wrong. You can compress truely random data. However you can't be guaranteed that any algorithm (or combination of algorithms) you use will give you a positive compression ratio in every single case.