I was vegetarian for a year or so, and I wish I could go back to it. I ended up developing a bad intolerance to soy, and moderate problems with beans and grains. Between this, my diet suddenly lacked protein. So I had to go back to some meat in my diet.
That said, after being vegetarian for a while, I'm a lot pickier about whatever meat I do eat, mostly fish and poultry. I tried going back to some of the thing I liked before (like sausage), but I find it rather gross in retrospect.
That said, most of my objection to meat is that it is an unsustainable practice. Unfortunately, "organic" meat is even worse this way. I have been a supporter of lab-grown meat alternatives for a long time, for this reason.
Which succinctly relates the 5 most important numbers in math, e, pi, i, 1 and 0.
Re:Binary computers? How long before base4 compute
on
DNA Goes Binary
·
· Score: 1
Adders aren't constant time, they're O(nlogn) even in hardware, using something better than ripple-carry anyhow.
Switching to a balanced-ternary system would be possible, since ternary works well with a split-rail supply, but any more than 3 states and it starts to act more like an analog computer than a digital. Noise becomes a real problem.
Binary is used because it uses so little area to implement. More complex systems start requiring window comparators, which use very nontrivial area. For a MSI (medium scale integration) system, it could be done, although noise is problematic still. For VLSI, there is just no reason.
Re:I'm Not Convinced
on
DNA Goes Binary
·
· Score: 3, Insightful
It IS pure conjecture, that's the whole point. They are trying to model early life processes, to show that it COULD happen this way. Right now, we have no solid theory on the evolution of DNA, which is the "missing link" in the general theory of evolution.
Remember, these are chemists, not paleobiologists, so they used diaminopurine, presumably because it was easier to artificially create the strands using it. Historical accuracy is not the point, this is a proof of concept.
Let me be the first to thank you on your wonderful viewpoint.
I personally consider myself to be both, simultaneously, pro-choice and pro-life. Abortion is wrong (except in a few specific cases), but adoption is fine.
Then again, my feelings may be skewed, being one of those unfortunate infertile women who still want to have children some day.
It's a reference design. One of the hardest things when you're designing a new motherboard is to get everything to "go" the first time. Once you're running, you can worry about fixing your network card. The board just lets people play around with the hardware.
With any luck there are expansions on the board for custom hardware, but even if there aren't, plenty of chipsets can be moved from PCI form-factor to PCI-on-the-PCB form factor easily enough. Lets you get "your" design running faster.
Oh, and they come with full schematics, which I believe to be royalty-free. Really good for use.
Honestly though, I like lower barrier-of-entry, since I'm mostly a hobbiest, so AMD got my money (the Elan, not the Athlon). Specs and datasheets are free for that chip;)
Funny, there are several things wrong with this statement:
Multiple clocking. Some chips are clocked to rising edge, some to falling edge, and some to both. Clocking to both doubles the EFFECTIVE clock speed.
Differently specced clocks. The AMD Elan I'm working with now can use either 33.000MHz or 33.333MHz as its base clock chip, leading to slight variations in the final frequency, but that's rounding error.
Split clock, anyone? Horrible to work with if you've done any LSI or VLSI, but running the chip on multiple multiples/divisors of the base clock is actually relatively common.
Finally, asynchronous designs, lol, although neither of the systems mentioned are asynchronous. Clockspeed is no longer a true concept anymore.
Reference boards are generally either oversimplified or overfeatured, to give the system designer a quick and dirty overview of how the chip works.
Funny you mention building a laptop from scratch, I'm currently looking at doing just that. To actually do the build, yes, you need the parts and some PCB facilities and a good solder station or a vapor-phase soldering system.
But to design them, you can just take the reference board into schematic capture, netlist to a pcb and move the parts around into your new form factor. Often you will hardly need a redesign.
The chip itself is almost certainly (I haven't worked with the Crusoe, but FPGA's do this) have an external EEPROM or flash ROM to store the configuration information (the code-morph ROM) in. They provide this in binary form, you stick it on the ROM. The system reads it on boot.
They have NOT released specs on code-morph, so you'd be on your own for reverse-engineering it, though.
Lately I've started a little short-run project, a rather small system where fans are impossible, and even heatsinks are not preferred.
So I went looking for competitive parts to use as CPU's. Transmeta looks really nice, unless you look closely. I can't find any way to get even proper tech specs (pinouts, etc) without buying a development board. Honestly, they should be up front with all the docs, and the chips should be buyable in small quantities from their page.
Now, this doesn't help average PC builders, since they are NOT standard pinout compatible. But for embedded hobbiests, it would be a boon. Then again, somebody might adapt some for PC motherboards...
I eventually just settled on a 133MHz AMD Elan. Nice chip, very highly integrated, slower than the Crusoe but does everything i need. Best part? Docs are right on AMD's page, rather easy to find.
Is it just me, or is Transmeta's market surprisingly.... Embedded?
I don't know about you, but when I'm working embedded, I either use Linux, or a proper RTOS (or one of the RTOS flavors of Linux). Windows? What's that?
Since when does slashdot have a new domain for each section? My program refused to send my login cookie, and so I ended up posting AC. ARG I really aught to be more careful. So just for a lark, let me repost my comment:
I would kill for some nice targeted ads... I mean, something from Tektronics or Fluke. Or a chip company, I have enough trouble selecting components, it would be easier if they were advertising them to me.
Instead, I get mortage quote spam (I'm canadian, it doesn't help), car insurance (no car OR driver's licence for that matter), penis enlargement (like WTF?), HGH, and weight loss (I'm on a weight-gain diet, because if I don't gain some mass I'll probably die, stupid anorexia.)
"Sterile, with low self esteem and taught only those things that might sexually please"
Well, except that I've learned many other trades besides (like hacking, hardware or software), it sounds a lot like me. I actually have a little writeup on my webpage of the characters in fiction that remind me of me, and Friday is definately one of them.
Sometimes I don't feel like a real human either. Figures.
I agree that they didn't have it particularly any easier (my first post up there was a mess...). I'm mostly just arguing that they didn't have it a lot worse. The point is that society is supposed to get better.
Some things have improved. Others have worstened, often considerably. All I'm saying is that the cycle will have to be broken before we ever get anywhere with actually improving civilization.
Wow, there have been a lot of good replies to this, so I'm just replying to myself instead of each directly. I wasn't trying to say that industrialization was lowering standard of living, I was trying to say that CORPORATIZED industrialization did.
I personally specialize in automation systems (although I'm still mainly a student), so I certanly don't believe that we should go back to being Amish. What I AM trying to say is that the industrialization has mainly helped the people with the most control over the process, which, despite what we would like to think on slashdot, is not us (although we do get paid a good salary so we won't complain about being used like this).
My ideal system would involve less central control of the industrialized capabilities, but more use of technological methods overall.
Oh, and the issue of mental health is actually quite real. The perception that it is a difference in diagnostic criteria is actually backwards. Evaluated using 1950's standards, something like 60% of children today have an anxeity disorder. Other diagnostic areas (which are non-stress related) have only seen a modest increase (less than 50%).
And I would personally love to see some of this easy money that I'm hearing about here. I live in a small apartment with my SO. One and a half incomes isn't enough to support us (my point above), and I seem to be having some trouble locating a tech job anywhere around here. And our problem is not due to expenditure, of which food is our primary one. The world is not a sunny place for those who are not working in good jobs.
Actually, corporatized industrialization has lowered the quality of life for humans. Starting with the beginnings of the industrial age, where women and children were being forced to work in terrible conditions, to now, when the average work day has not decreased at all (and increased in a lot of professions).
In fact, now we have the problem of North American excesses lowering the quality of life elsewhere in the world, especially in developing countries that can get more money from selling their crops as cattle feed in the US than from selling it to their populace. Guess which they choose?
And, to top it all off, in the past 30 years or so, incidences of stress-related mental illness has increased by something like 500% (I forget which study I read that in, but anyway).
And what do we have to show for it? Do we have more time to spend with our friends and families? No, all we have is a few new toys (although, as a geek myself, I have to admit that they are fun toys). If we see an average person working one day a week and making enough money to support themselves and their families, then that would be a massive improvement in quality of life.
In fact, we have seen the opposite; the two-income family is so common that it has become difficult to be one-income anymore. The quality of life has decreased enough that the average two-income family now lives about the same as an average one-income family in the 1920's.
Remember, those who do not understand history are doomed to repeat it.
This is really starting to sound like certain other operating systems. Every month or two somebody declares Linux dead. While the most obvious is OS/2, that one DID finally die in the end, but took 6 or 7 years to do so. And there is STILL a couple projects to reimplement it, so the death seems to be the fault of closed-source software.
Contrast also with Apple.
Just because Linux gets low numbers of users, it's not the number but the derivative (rate of change) that you have to look at, in order to declare an operating system dead. By this logic, Linux is still kicking, but Windows is dead, since Windows is no longer really increasing in use (they still have sales, but they're almost all "upgrade" sales, hence the attempted change of license methods).
And, somebody please explain, HOW do you kill an open-source work? People like me will always tinker with it, because it's FUN.
I've played net games at not one, but TWO high-schools. One of them was fairly unofficial, but half the teachers came in and joined us. It was long enough ago we were playing Doom. Then again, in another one, when we were doing Networking class, Doom was played, but because the machines were 486's.
The other thing is that safe-grads often book the computer lab for this very purpose.
OK, I'm complaining about moderation quite a bit
today, it seems... This IS funny though.
Analog computing DOES have a lot of uses, but it
needs vacuum tubes about as much as bicycles need wood.
These (LM741)
do the job in an all-solid state way (you can, of course, use any op-amp
you desire). Connect them with 4066's and you have your
reprogrammable analog computer. Better yet, this can
be built on an IC, and programmed just like an FPGA.
In fact, Lattice and a few others make chips just like this.
Analog computers are terrific for mathematic problems
where an exact numerical answer is not as important
as getting a good answer quickly. It really is too bad
that they've fallen out of favor as of late.
No, but they're terminated in a plain resistor. And
a PROPERLY terminated cable appears as a resistor anyway,
although TV cables are seldom terminated properly. And
the resistive losses in the cables can mess up measurement,
so it really doesn't work well (anything more than about 100ft or
so is not really practical).
Come to think of it, that's probably why they
basically don't do this anymore (they used to do
something similar a while back).
Even a mid-power tube uses no more than a couple
amps of heater current (usually 6.3V if I remember
the standard correctly), which is about maybe 12
Watts. Your CPU uses more like 70 Watts. A little
wussy general-purpose small-signal tube will only
use a couple-hundred milliamps, so the whole thing
would use just a Watt or two. No big deal for heat
either (the wattage directly determines heat).
I'll be the first to say that the absense of a voltage
high enough to do the job is no problem, check
my webpage for an example of a circuit that needed
+12V on a board that was only supplied at +5V (under the
PIC stuff).
I even extended it later so it supplied -10 or
so for an RS232 interface.
A small tube will need
70 to 100V or so, which is not that difficult with
a standard diode-cap cascade voltage multiplier.
The whole circuit would completely lack transformers
and probably fit in a space a centimeter on a side.
The noise issues are NOT from the digital circuits,
although I am a little surprised not to see a metal
box around the tube. Microphonics from the
moving parts in the computer and AF interference from
the motors would be a problem.
I agree with the people who call it a cheap gimmick,
since the tube is almost certainly just a processing
stage before feeding a standard solid-state power amp,
but calling it a hoax outright... That I'm not so
sure about.
It was my understanding that they could read the
impedance on individual lines, and since all TV's
are supposed to be terminated in the same
value, a small DC current can tell you how many
TV's there are.
Now, if you got one of those little cable amplifiers
that Radio Shack sells, then they would only see one terminator.
End of problem. There are many similar tricks
that can be used (impedance matching networks
do a nice job too).
I have the same problem. If I'm on an "ergonomic"
keyboard for any length of time at all, I start
getting shooting pains within 10 minutes. I
personally use an old Honeywell keyboard (back from
when Honeywell made keyboards, I think Keytronics
bought out the keyboard division). It's a pretty
hard feel keyboard, but without the really
obnoxious clicking noises of the old IBM's
(it still resists almost as much). I frequently
go on long coding binges and never have any real
problems with pains.
My experience with ergo's ususally involves doubletyping
as my fingers bounce the keys (followed by the
aforementioned wrist pains).
Then again, after looking at it, I have a strange
slant where my hand attaches to my wrist. The
angle that my hands approach a keyboard from
are very similar to that of somebody normal
on an ergo keyboard. So maybe I'm anomalous (because I
started coding at age 5?)
Oh, there IS more to it, for sure. Perhaps I should
have been more clear.... It's SOI, but it's not
VLSI... With any luck I should be able to make
TTL and other SSI chips when I get the equipment.
VLSI is a whole other ball of wax... I'd need
to tighten up pretty much the ENTIRE system, and
THEN make it work with CMOS. The process is
really optimized for TTL (which is simpler,
and has nice big featuresizes that you can draw
with, for instance, an inkjet head).
Doing the amount of layers involved in a CMOS
process can be quite nasty.
Slashdot Mirror
That said, after being vegetarian for a while, I'm a lot pickier about whatever meat I do eat, mostly fish and poultry. I tried going back to some of the thing I liked before (like sausage), but I find it rather gross in retrospect.
That said, most of my objection to meat is that it is an unsustainable practice. Unfortunately, "organic" meat is even worse this way. I have been a supporter of lab-grown meat alternatives for a long time, for this reason.
Which succinctly relates the 5 most important numbers in math, e, pi, i, 1 and 0.
Switching to a balanced-ternary system would be possible, since ternary works well with a split-rail supply, but any more than 3 states and it starts to act more like an analog computer than a digital. Noise becomes a real problem.
Binary is used because it uses so little area to implement. More complex systems start requiring window comparators, which use very nontrivial area. For a MSI (medium scale integration) system, it could be done, although noise is problematic still. For VLSI, there is just no reason.
Remember, these are chemists, not paleobiologists, so they used diaminopurine, presumably because it was easier to artificially create the strands using it. Historical accuracy is not the point, this is a proof of concept.
I personally consider myself to be both, simultaneously, pro-choice and pro-life. Abortion is wrong (except in a few specific cases), but adoption is fine.
Then again, my feelings may be skewed, being one of those unfortunate infertile women who still want to have children some day.
With any luck there are expansions on the board for custom hardware, but even if there aren't, plenty of chipsets can be moved from PCI form-factor to PCI-on-the-PCB form factor easily enough. Lets you get "your" design running faster.
Oh, and they come with full schematics, which I believe to be royalty-free. Really good for use.
Honestly though, I like lower barrier-of-entry, since I'm mostly a hobbiest, so AMD got my money (the Elan, not the Athlon). Specs and datasheets are free for that chip ;)
Multiple clocking. Some chips are clocked to rising edge, some to falling edge, and some to both. Clocking to both doubles the EFFECTIVE clock speed.
Differently specced clocks. The AMD Elan I'm working with now can use either 33.000MHz or 33.333MHz as its base clock chip, leading to slight variations in the final frequency, but that's rounding error.
Split clock, anyone? Horrible to work with if you've done any LSI or VLSI, but running the chip on multiple multiples/divisors of the base clock is actually relatively common.
Finally, asynchronous designs, lol, although neither of the systems mentioned are asynchronous. Clockspeed is no longer a true concept anymore.
Funny you mention building a laptop from scratch, I'm currently looking at doing just that. To actually do the build, yes, you need the parts and some PCB facilities and a good solder station or a vapor-phase soldering system.
But to design them, you can just take the reference board into schematic capture, netlist to a pcb and move the parts around into your new form factor. Often you will hardly need a redesign.
The chip itself is almost certainly (I haven't worked with the Crusoe, but FPGA's do this) have an external EEPROM or flash ROM to store the configuration information (the code-morph ROM) in. They provide this in binary form, you stick it on the ROM. The system reads it on boot.
They have NOT released specs on code-morph, so you'd be on your own for reverse-engineering it, though.
So I went looking for competitive parts to use as CPU's. Transmeta looks really nice, unless you look closely. I can't find any way to get even proper tech specs (pinouts, etc) without buying a development board. Honestly, they should be up front with all the docs, and the chips should be buyable in small quantities from their page.
Now, this doesn't help average PC builders, since they are NOT standard pinout compatible. But for embedded hobbiests, it would be a boon. Then again, somebody might adapt some for PC motherboards...
I eventually just settled on a 133MHz AMD Elan. Nice chip, very highly integrated, slower than the Crusoe but does everything i need. Best part? Docs are right on AMD's page, rather easy to find.
I don't know about you, but when I'm working embedded, I either use Linux, or a proper RTOS (or one of the RTOS flavors of Linux). Windows? What's that?
I would kill for some nice targeted ads... I mean, something from Tektronics or Fluke. Or a chip company, I have enough trouble selecting components, it would be easier if they were advertising them to me.
Instead, I get mortage quote spam (I'm canadian, it doesn't help), car insurance (no car OR driver's licence for that matter), penis enlargement (like WTF?), HGH, and weight loss (I'm on a weight-gain diet, because if I don't gain some mass I'll probably die, stupid anorexia.)
Is that the Gor I keep hearing about in the BDSM world? Quite frankly what I've heard scares me... Oh well.
Well, except that I've learned many other trades besides (like hacking, hardware or software), it sounds a lot like me. I actually have a little writeup on my webpage of the characters in fiction that remind me of me, and Friday is definately one of them.
Sometimes I don't feel like a real human either. Figures.
Some things have improved. Others have worstened, often considerably. All I'm saying is that the cycle will have to be broken before we ever get anywhere with actually improving civilization.
I personally specialize in automation systems (although I'm still mainly a student), so I certanly don't believe that we should go back to being Amish. What I AM trying to say is that the industrialization has mainly helped the people with the most control over the process, which, despite what we would like to think on slashdot, is not us (although we do get paid a good salary so we won't complain about being used like this).
My ideal system would involve less central control of the industrialized capabilities, but more use of technological methods overall.
Oh, and the issue of mental health is actually quite real. The perception that it is a difference in diagnostic criteria is actually backwards. Evaluated using 1950's standards, something like 60% of children today have an anxeity disorder. Other diagnostic areas (which are non-stress related) have only seen a modest increase (less than 50%).
And I would personally love to see some of this easy money that I'm hearing about here. I live in a small apartment with my SO. One and a half incomes isn't enough to support us (my point above), and I seem to be having some trouble locating a tech job anywhere around here. And our problem is not due to expenditure, of which food is our primary one. The world is not a sunny place for those who are not working in good jobs.
And, to top it all off, in the past 30 years or so, incidences of stress-related mental illness has increased by something like 500% (I forget which study I read that in, but anyway).
And what do we have to show for it? Do we have more time to spend with our friends and families? No, all we have is a few new toys (although, as a geek myself, I have to admit that they are fun toys). If we see an average person working one day a week and making enough money to support themselves and their families, then that would be a massive improvement in quality of life.
In fact, we have seen the opposite; the two-income family is so common that it has become difficult to be one-income anymore. The quality of life has decreased enough that the average two-income family now lives about the same as an average one-income family in the 1920's.
Remember, those who do not understand history are doomed to repeat it.
Contrast also with Apple.
Just because Linux gets low numbers of users, it's not the number but the derivative (rate of change) that you have to look at, in order to declare an operating system dead. By this logic, Linux is still kicking, but Windows is dead, since Windows is no longer really increasing in use (they still have sales, but they're almost all "upgrade" sales, hence the attempted change of license methods).
And, somebody please explain, HOW do you kill an open-source work? People like me will always tinker with it, because it's FUN.
I'm 20 and have been programming for 15 years... Does that mean I don't exist?
(And I promptly vanish in a puff of smoke).
One of them was fairly unofficial, but half the
teachers came in and joined us. It was long enough
ago we were playing Doom. Then again, in another
one, when we were doing Networking class, Doom was
played, but because the machines were 486's.
The other thing is that safe-grads often book the
computer lab for this very purpose.
Analog computing DOES have a lot of uses, but it needs vacuum tubes about as much as bicycles need wood.
These (LM741) do the job in an all-solid state way (you can, of course, use any op-amp you desire). Connect them with 4066's and you have your reprogrammable analog computer. Better yet, this can be built on an IC, and programmed just like an FPGA. In fact, Lattice and a few others make chips just like this.
Analog computers are terrific for mathematic problems where an exact numerical answer is not as important as getting a good answer quickly. It really is too bad that they've fallen out of favor as of late.
Come to think of it, that's probably why they basically don't do this anymore (they used to do something similar a while back).
Even a mid-power tube uses no more than a couple amps of heater current (usually 6.3V if I remember the standard correctly), which is about maybe 12 Watts. Your CPU uses more like 70 Watts. A little wussy general-purpose small-signal tube will only use a couple-hundred milliamps, so the whole thing would use just a Watt or two. No big deal for heat either (the wattage directly determines heat).
I'll be the first to say that the absense of a voltage high enough to do the job is no problem, check my webpage for an example of a circuit that needed +12V on a board that was only supplied at +5V (under the PIC stuff). I even extended it later so it supplied -10 or so for an RS232 interface.
A small tube will need 70 to 100V or so, which is not that difficult with a standard diode-cap cascade voltage multiplier. The whole circuit would completely lack transformers and probably fit in a space a centimeter on a side.
The noise issues are NOT from the digital circuits, although I am a little surprised not to see a metal box around the tube. Microphonics from the moving parts in the computer and AF interference from the motors would be a problem.
I agree with the people who call it a cheap gimmick, since the tube is almost certainly just a processing stage before feeding a standard solid-state power amp, but calling it a hoax outright... That I'm not so sure about.
Now, if you got one of those little cable amplifiers that Radio Shack sells, then they would only see one terminator. End of problem. There are many similar tricks that can be used (impedance matching networks do a nice job too).
My experience with ergo's ususally involves doubletyping as my fingers bounce the keys (followed by the aforementioned wrist pains).
Then again, after looking at it, I have a strange slant where my hand attaches to my wrist. The angle that my hands approach a keyboard from are very similar to that of somebody normal on an ergo keyboard. So maybe I'm anomalous (because I started coding at age 5?)
VLSI is a whole other ball of wax... I'd need to tighten up pretty much the ENTIRE system, and THEN make it work with CMOS. The process is really optimized for TTL (which is simpler, and has nice big featuresizes that you can draw with, for instance, an inkjet head).
Doing the amount of layers involved in a CMOS process can be quite nasty.