Isn't this basically what the DCMA effectively forces one to do- that is, if you follow it to the letter?
Look, I can't use MY pencil because the RIAA hasn't licensed it to write an opinion about song X from artist, erm label Y. (Yeah, exaggeration, but what the hey..)
The "bad old days" is precisely what large copyright-holders want- It makes control so much easier when it is illegal to create, copy, or use information (which I might point out is the lifeblood of any culture..) without using their hardware or software.
Just imagine what it will (could) be like if we followed the DCMA to the letter =) What fun.
Right.
Yes, but for some things even 4x isn't the speed increase you'd like to see. For example- video compositing. It just can't be effectively done on a PC due to the large amounts of bandwidth taken up by each stream. If you're doing professional video (SDI) and you'd like to deal with the bitstream itself (which would be really nice) it is 240M/s per stream. That is a lot of data. Then you have to consider that you're going in and out of the video capture card, and that you probably would like to fade between two streams, etc. 4x is nice, but an order of magnitude improvement would be more like it.
Basically, for those of you too impatient to read the article, it works by using VSCELs (lasers) through conventional optics, and into a high-speed collector. The lasers can work at up to 1 Ghz, and the processing is done (it seems) in analog by the optics. Acquisition of the date is performed by the collector, which operates at 10 Ghz.
The theory is that optics can perform FFTs, DCTs, etc for you at the speed of light, and there are many applications that need these operatons done. Any other processing, correlation, etc would be done by conventional, low-performance DSPs.
They also say that their current model works at 20 T ops/sec at 20 watts, and list what would be required of typical DSPs, etc down to ASICS.
Seems promising, but it is still a long way away from a nice optical CPU.
Actually, I'd prefer motherboards with NO pci slots. Upgrading to a faster bus would be a good idea.
I'm running up against the bus-bandwidth problem just about every day.
Excuse me?
I deal with high-bandwidth data and cpu-intensive processing for a living, and while I'd be thrilled for a java or java-like language to be able to handle either constraint, the fact is they don't.
I believe that assembly, procedural, functional, and OO programming are all (or should be!) a fundamental part of any good CS curriculum.
In my opinion, they should be taught in the following order: procedural, assembly, functional, OO.
Why?
Procedural is 'natural' without being too high level. It is close to the hardware without sacrificing all generality.
Assembly, well thats obvious. Now that you've built up the student's confidence, its time to introduce them to the innards of the computer. Any Von Neumann architecture will do!
Follow that up with a functional language like LISP. This is different. Its a mold-breaking experience.
Then finish off with OO.
Pascal is great for learning procedural programming.
MIPS is a fine assembly language, but anything will do.
LISP is really the best lan guage for teaching functional programming.
smalltalk is probably the best language for teaching OO.
If the student can't pick up another language in a few days after learning all of this then there is something wrong with them.
And we should believe that threatening a lawsuit costs nothing to either party?
Or how about applying your theory along a different tangent- If the threat of something is meaningless and non-chilling, then why does the DCMA exist? Doesn't it exist solely to counter the _threat_ of copying?
(In reality no, it also exists to counter the threat of competition from independant artists, but that is beside the issue for this post, right?)
I wonder whether or not losses would be
so bad in the house if we did that.
The real metric is: Does converting to high voltage, and then back again lose more power than supplying the low voltage in the first place.
Assuming, of course, that most appliances actually USE low voltage.
(some use step-up transformers... it would be an interesting study, that is for sure...)
I believe that the best solar panel efficiency
right now is 40%, however I don't remember where to point you, or how much more it costs...
IMHO the main advantage of solar power is that
it is relatively low maintenance.
(every 10 yrs or so, replace half the panels,
to compensate for their decreasing efficiency.
Which is another way of saying that there are fewer failure-modes for fibre-pipes vs. standard conductors.
Another advantage of fibre that you didn't mention:
Lower maintanance! For example: They have to check all the cables in an F-16 periodically to look at insulator wear. (because of short-circuits).
Fibre problems are easier to diagnose, because a potential "short circuit" would be incredibly difficult to cause, and because when a connection is degraded, (i.e. fibre is cut/close to cut) it is easy to diagnose without having to open up panels...
At (what is hdtv?) 1900X1200x3x60 we could multicast 243 RAW video streams to each house. (410 Mb/sec/channel)
Compressed, assuming 8:1 compression (a modest compression) would give us 87.5 Gb/s, and 243 channels of HDTV per home. (not counting tha audio, but the audio bandwidth is negligible at these rates. (cuts it down to 86.6 Gb/s assuming cd-quality 5-channel sound/channel of hdtv)
Not bad at all.
So: Multicast is important!!
Re:How should one go about his education then?
on
The IT Labor Shortage
·
· Score: 1
Georgia TECH's co-op program is better known. You say "Georgia" and people thing UGA.
Ga Tech's CS program is relatively well respected. Its ratings have been consistantly going up in the school ratings for a few years now. (and people say that those ratings are typically 4 or so years behind the actual changes).
I will be graduating from Tech with a CS degree in May, and I think I have a degree which is worth something. (I go to interiews and when I mention "Ga Tech" people say: Oh thats a good school)
I don't know about IllinoisTech, so you'll have to look into that.
If you are planning on going into grad-school, Ga Tech is a nice place to be (right now). Luck!
I do video-processing. Meaning, that I run 8*3*640*480*30 bits through the bus every second. Thats 221184000 bits/second, or 210 MB/sec.
Memory is a serious constraint for me, as most of the images tend to reside in memory before I get to process them. (and besides, the results go back to a different location in memory...)
memory BANDWIDTH IS my problem. Latency doesn't effect me much, because its nearly completely sequential reads/writes, nevertheless, this is something that the cache does not cope well with.
(And I really want to process 4 video-streams at a time, think 840 MB/second...)
From a technical standpoint this would be icky icky. First of all, ethernet doesn't allow synchronous data xferrs, like both Firewire and USB, nor does it allow isosynchronous transmission... Nope, ethernet is asynchronous, and that isn't good for a bus that needs synchronous capabilities! (The ability to do all three is what makes firewire and USB neat.) And USB-2 will be around 400MB/sec...(supposedly, I believe that it is still a draft-standard) (And this throughput isn't effected nearly as much by traffic as Ethernet is.. Expect a max utilization of around 40% with ethernet (thats pushing it!) USB (being mostly synchronous) does better.) That being said, USB is like early mainframe protocols in that it is asymetric with respect to control of the medium.. Ethernet is not.. (which is why ethernet is good for peer-peer connections, while USB wouldn't handle more than one master well (it'd break. So sorry. =) ))
Slashdot Mirror
lol.
Isn't this basically what the DCMA effectively forces one to do- that is, if you follow it to the letter?
Look, I can't use MY pencil because the RIAA hasn't licensed it to write an opinion about song X from artist, erm label Y. (Yeah, exaggeration, but what the hey..)
The "bad old days" is precisely what large copyright-holders want- It makes control so much easier when it is illegal to create, copy, or use information (which I might point out is the lifeblood of any culture..) without using their hardware or software.
Just imagine what it will (could) be like if we followed the DCMA to the letter =) What fun.
Right.
Yes, but for some things even 4x isn't the speed increase you'd like to see. For example- video compositing. It just can't be effectively done on a PC due to the large amounts of bandwidth taken up by each stream. If you're doing professional video (SDI) and you'd like to deal with the bitstream itself (which would be really nice) it is 240M/s per stream. That is a lot of data. Then you have to consider that you're going in and out of the video capture card, and that you probably would like to fade between two streams, etc. 4x is nice, but an order of magnitude improvement would be more like it.
Ahh. thanks! =)
Basically, for those of you too impatient to read the article, it works by using VSCELs (lasers) through conventional optics, and into a high-speed collector. The lasers can work at up to 1 Ghz, and the processing is done (it seems) in analog by the optics. Acquisition of the date is performed by the collector, which operates at 10 Ghz.
The theory is that optics can perform FFTs, DCTs, etc for you at the speed of light, and there are many applications that need these operatons done. Any other processing, correlation, etc would be done by conventional, low-performance DSPs.
They also say that their current model works at 20 T ops/sec at 20 watts, and list what would be required of typical DSPs, etc down to ASICS.
Seems promising, but it is still a long way away from a nice optical CPU.
Actually, I'd prefer motherboards with NO pci slots. Upgrading to a faster bus would be a good idea.
I'm running up against the bus-bandwidth problem just about every day.
Excuse me?
I deal with high-bandwidth data and cpu-intensive processing for a living, and while I'd be thrilled for a java or java-like language to be able to handle either constraint, the fact is they don't.
I believe that assembly, procedural, functional, and OO programming are all (or should be!) a fundamental part of any good CS curriculum.
In my opinion, they should be taught in the following order: procedural, assembly, functional, OO.
Why?
Procedural is 'natural' without being too high level. It is close to the hardware without sacrificing all generality.
Assembly, well thats obvious. Now that you've built up the student's confidence, its time to introduce them to the innards of the computer. Any Von Neumann architecture will do!
Follow that up with a functional language like LISP. This is different. Its a mold-breaking experience.
Then finish off with OO.
Pascal is great for learning procedural programming.
MIPS is a fine assembly language, but anything will do.
LISP is really the best lan guage for teaching functional programming.
smalltalk is probably the best language for teaching OO.
If the student can't pick up another language in a few days after learning all of this then there is something wrong with them.
And we should believe that threatening a lawsuit costs nothing to either party? Or how about applying your theory along a different tangent- If the threat of something is meaningless and non-chilling, then why does the DCMA exist? Doesn't it exist solely to counter the _threat_ of copying? (In reality no, it also exists to counter the threat of competition from independant artists, but that is beside the issue for this post, right?)
I wonder whether or not losses would be
so bad in the house if we did that.
The real metric is: Does converting to high voltage, and then back again lose more power than supplying the low voltage in the first place.
Assuming, of course, that most appliances actually USE low voltage.
(some use step-up transformers... it would be an interesting study, that is for sure...)
I believe that the best solar panel efficiency
right now is 40%, however I don't remember where to point you, or how much more it costs...
IMHO the main advantage of solar power is that
it is relatively low maintenance.
(every 10 yrs or so, replace half the panels,
to compensate for their decreasing efficiency.
If that is the only reason that they stayed with that sofware, then they acted like fools.
I would've imagine that they'd have heard about sunk costs!
"Pike in hand"...
Seems more than a coincidence..
control is money.
Which is another way of saying that there are fewer failure-modes for fibre-pipes vs. standard
conductors.
Another advantage of fibre that you didn't mention:
Lower maintanance!
For example: They have to check all the cables in
an F-16 periodically to look at insulator wear.
(because of short-circuits).
Fibre problems are easier to diagnose, because a potential "short circuit" would be incredibly difficult to cause, and because when a connection is degraded, (i.e. fibre is cut/close to cut) it is easy to diagnose without having to open up panels...
Which is why multicast is important.
At (what is hdtv?) 1900X1200x3x60 we could
multicast 243 RAW video streams to each house.
(410 Mb/sec/channel)
Compressed, assuming 8:1 compression (a modest compression) would give us 87.5 Gb/s, and 243
channels of HDTV per home. (not counting tha audio, but the audio bandwidth is negligible at these rates. (cuts it down to 86.6 Gb/s assuming cd-quality 5-channel sound/channel of hdtv)
Not bad at all.
So: Multicast is important!!
Georgia TECH's co-op program is better known.
You say "Georgia" and people thing UGA.
Ga Tech's CS program is relatively well respected.
Its ratings have been consistantly going up in the school ratings for a few years now. (and people say that those ratings are typically 4 or so years behind the actual changes).
I will be graduating from Tech with a CS degree in May, and I think I have a degree which is worth something. (I go to interiews and when I mention "Ga Tech" people say: Oh thats a good school)
I don't know about IllinoisTech, so you'll have to look into that.
If you are planning on going into grad-school, Ga Tech is a nice place to be (right now).
Luck!
a->d is very very slow.
If you perform a->d inside your cpu, its going to
be dog slow (since it sounds like there will be an a->d inside your ALU).
Looks like the universal from universal serial bus
is missing =)
Next, you realize, there will be OpenUSB, and then NetUSB...
Heh.
I do video-processing.
Meaning, that I run 8*3*640*480*30 bits through the bus every second. Thats 221184000 bits/second,
or 210 MB/sec.
Memory is a serious constraint for me, as most of the images tend to reside in memory before I get to process them. (and besides, the results go back to a different location in memory...)
memory BANDWIDTH IS my problem. Latency doesn't effect me much, because its nearly completely sequential reads/writes, nevertheless, this is something that the cache does not cope well with.
(And I really want to process 4 video-streams at a time, think 840 MB/second...)
I live in an apartment.
Its tough to wire up one of those...
(Management gets peeved =) )
From a technical standpoint this would be icky icky. First of all, ethernet doesn't allow synchronous data xferrs, like both Firewire and USB, nor does it allow isosynchronous transmission... Nope, ethernet is asynchronous, and that isn't good for a bus that needs synchronous capabilities! (The ability to do all three is what makes firewire and USB neat.) And USB-2 will be around 400MB/sec...(supposedly, I believe that it is still a draft-standard) (And this throughput isn't effected nearly as much by traffic as Ethernet is.. Expect a max utilization of around 40% with ethernet (thats pushing it!) USB (being mostly synchronous) does better.) That being said, USB is like early mainframe protocols in that it is asymetric with respect to control of the medium.. Ethernet is not.. (which is why ethernet is good for peer-peer connections, while USB wouldn't handle more than one master well (it'd break. So sorry. =) ))
Even under DMCA it is legal to reverse-engineer for reasons of interoperability.
(as per my understanding)
Lucky for us, we have a really friggen BIG asteroid (trojan, maybe, moon, maybe) "orbiting" the earth a little ways out.
I wonder when we'll mine that sucker, its bound to have a LOT of good stuff in it.
The artemis society has plans for all of this.
And they are using current, available technology (and its been researched by AEs and other (some NASA) types)
Visit their page.
And yea I meant to (mis)spell it that way =)