Re:EC over IP I have been doing this for years.
on
Replacing TCP?
·
· Score: 1
I mean you take the "PING" program and run ping, you see 90% packet loss with a 6 seconds average latency, I believe I still have the files laying around from that data ping xxx.com > pinglog
Anyhow when you start to study Error corrector/erasure code you quickly realize that at best you can only reduce errors, IE a (DB increase if you will in S/N) By increasing used bandwidth you increase S/N. Shannon's law type of stuff.
Anyhow If can be proven mathematically that no matter how much you add you can not get 100% perfect data cross. Unless you "close the loop" using feedback as to what was lost, hence TCP.
Now I found you can and combine Error Correction and retransmission to get 100% of your signal/data across. And by using the right type of error correction you didn't need to increase latency to do your retransmissions. (The how, would be a very long discussion to get into that I'd want to take offline)
Now if we take this further I can deliberately strip my codes across a connection to intentionally cause a predictable and manageable amount of packet loss. For example I know I have a T3 (45Mbs) on one side and an dual channel bonded ISDN on the other 128Kbps (this shows how far back I had worked on this) I know that the network between my T3 and the ISDN is probably better then 10Mbps at all point up until the "last mile".
Now let say that in 1996 MAE West was experiencing 20% packet loss (I have logs of it) almost every afternoon, but I still needed to get my data to some ISDN line on the other side of this gigabit router that is over saturated with bazillions of TCP connections.
Is this situation what if I use an error correction code set to recover 25% loss but with retransmissions? If I send 128K I'd have 102Kbps arrive. Then there is the overhead of the code itself to deal with. But knowing where my loss if occurring at I can deliberately send 25% more data to that connection. So about 170 Kbps gets sent. 20% lost in the backbone and I am guaranteed that the remaining 5% will be lost at the ISP's ISDN terminal server before being sent down the ISDN line. So I will have a known state of 25% lost data, but 128Kbps going over my connection.
Are you following me? What ever the lost across the backbone is become irrelevant since the lost at the connection will force me to at 25% loss.
Now what If I were to use a perfect code (turn out I found a number of perfect erasure codes) Perfect codes
So start with 128K create 170K, I lost X% across the net lost 25% -X at the ISP and receive 128K. Using a perfect code for 25% loss I recover 100% of my data that needed 128Kbps to get across.
But this only works using retransmissions that are placed back into the error corrected stream! -- YES, Make sure everyone know John Sokol wrote this first since this was the big revelation that made this work!
By doing this it allows you to increase the hamming distance for the next error recovery so you can tolerate future lost data better. While recovering 100% of the data sent first pass.
Lets take the Hamming (7,4) Code it's a perfect code. It has a hamming distance of three. This allow you to correct 1 error and detect 2.
For error you can only correct for 0.5 - 1/2 the hamming distance, but for Erasures you can correct for 1 - the hamming distance.
So I can correct for 2 lost packed out of 7 I send, but only 4 are data with 3 added overhead.
But these are block codes, and are not very efficient or effective. So used brute force and later Genetic algorithms to search for long perfect cyclical codes for Erasures.
The closest thing to the codes I found is called "hagelbarger codes"
Well the Gnutilla seems to be working great, excellent Idea.
Re:EC over IP I have been doing this for years.
on
Replacing TCP?
·
· Score: 1
I'm glad you asked.
Yes I mean just that.
90% packet loss according to ping. (round trip) Didn't have a tool to measure each direction independently at that time. (do now) With almost 10 seconds round trip time for that event!
It all depends on where the bottle neck is and what's behind it. At the time we had characterized the way packets were dropped by differnt routers when the buffers were full and overflowing (dropping packets).
Anyhow since we had 10 Mbps ethernet to the 64K link out of Sri Lanka, then on the other side which was MCI in New York ( I had also managed to place a Co-Lo server very close to that router also) there was only one bottleneck that I didn't have control over. the public internet pipe at 64K traveling over fiber 12,000 miles, I could get the ISP to dissconect the rest of thier coustomers just for me.
You can still get the data through a link like this by having the sender push out almost 600 kbps in packets to get 60Kbs of my packets out on the receive side! This is a RUDE Protocol..
So of 64kbps only 4Kbps packets was other people junk and 60 kbps was mine, this is done by saturating the link with my packets! So does this sound familiar? Sort of like a DOS attack maybe...
Now the real trick was to make the 1/10 of the packets that did make it through turn into the 60Kbps of video data that was trying to send over this link with 0% no loss and low latency! So no retranmissions!
We did this!.
For anyone really willing to put some time into this opensourcing this project I will be happy to go into as much detail as they can handle or need to reimplement into an open and more usefull application.
A knowledge of statistics would also help since I ended up learning it for this project myself.
John
EC over IP I have been doing this for years.
on
Replacing TCP?
·
· Score: 5, Interesting
ecip.com I call it Error Correcting IP, and used it to stream live video from Sri Lanka in 1997 with Arthur C. Clarke Hal's Birthday
it was a 64K shared line with 90% packet loss, I received 60Kbps for the video stream. ( I have the video to prove it )
We even filled preliminary patents on this back in 1996 but they were never followed through with.
Which is more accurate since UDP doesn't have errors, it either come across 99.999% perfect or not at all.
So there is more information then in an error situation where ever bit is questionable.
What this means almost 1/2 the hamming distance in the codes in needed to correct an errasure verses and error.
Turns out the Error/Erasure correcting scheme it critical and not obvious. I spent almost 5 years working on this part time before it started making some real breakthroughs.
My original system was designed for 25% packet loss (not uncommon in 1996).
In the inital idea we added 1 exored packet for every three data packets, but at 25% packet loss, it turns out that it didn't increase reliablity at all! Working this out with probablities was a major eye opener!
Even when you work the problem out you realize you will still need some retransmissions to make up for lost packets, there is no possible solutions without this.
I have been trying to find people to help opensource this since I have working far too hard just to survive since 2000 to even consider taking on another task.
Anyone interested in my research and carring this forward please see my site and contact me.
At my Enumera Project We used a 100 CPU Beowulf like Cluster to compress Video in realtime using a several different codec technologies including wavelet.
The idea is that TV stations think nothing on spending $100,000 on hardware, but people streaming web video keep trying to use a cheap > $2000 PC. We tried the other approch. The more CPU power a compression algorythem uses the more efficient the compression. This is almost a rule, not counting optimization(this is equal to adding CPU power). There is a limit to this, and the code needs to be able to use this.
We were able to compress 1080i HDTV with H.263l in real time.,Looked decent at 1Mbps. I'm sure doing this with the Dirac codec with this would also work.
Lyons Gate and AMC with their proprietary Digital Theatre Distribution System (DTDS),
is directly going against DCI - Digital Cinema Initiatives that is made up from Disney, Fox, MGM, Paramount, Sony Pictures Entertainment, Universal and Warner Bros.
They are fighting for control and standards for the new Digital Cinema.
AMC's approch was very slick, they started puting low res tv add up, and deploying these digital projectors then very quickly are pushing movies out. I can't find any info on what AMC's resolution or projectors or or the Satellite system used.
DCI is using microspace or Huges for it's system and has standardized on 2K projectors 2048x1080 this is about where HDTV 1080p/24 is 1920x1080.
DCI also supports 4K 4096x2160 , but from my visit at there test bed, the USC, ETC center they were using 1024x768 video to drive everything.
Hey Publishing in JAVA is practicaly Open Source since the code decompiles back to very clean source with original variable names and everything so most JAVA players are derivative works. See Mocha or JAD for decompilers.
The Livecam server we developed in 1995 and dominated the adult industry already did all this and supported more viewers with better quality. We supported Motion JPEG or H.236 in 1999 with GSM audio, with 20Kbps to 70 Kbps streams.
I just love it when someone else come out with it all over again and everyone thinks it's new.
----Original Message----- From: James S Sent: Friday, September 10, 2004 5:18 PM To: sokol@videotechnology.com Cc: Jesse Monroy Subject: Hey These guys just invented the Java player we created in 1999
SpaceShipOne Flights are planed for September 29th, 2004, October 13th, 2004 at the Mojave Airport, Civilian Aerospace Test Center in Mojave, California.
I have taken several 250GB Maxtor hard drives and poured a thick layer of special rubber on them then been able to throw then out a second story window. Then hook them up and have then work fine.
I have also taken several power supply including an antec 480 watt and remove it from it's case and fill it with epoxy and stuff to increase it's thermal conductivity.
And even saw off the heat sink using a band saw from a $500 ATI Radeon x800 xt graphics card and pour epoxy over it. This was the only part that I was feeling queasy over.
Even unsoldering the heatsinks off of the northbridge on motherboards, and cutting off various connectors.
The end result has been a completely fanless, absolute dead silent computers running at 3.4 Ghz.
www.nisvara.com
Mine you many computers were killed in the learning process of how to do this correctly.
From School to work, the world keeps trying to put people into pigeonholes. Expecting individuals to spend 10 years to become adequate at one profession. A resume that shows expertise in Digital and analog hardware design or windows and Linux software tend to turn many prospective employers off. Just add Machine shop, mechanical design, chemistry, optics, number theory, cryptography to that and now they just don't know what to make of you.
For myself and many people who I have met like me that can rise above average levels , It is a challenge. There are many people who are threatened and territorial, and others who try to box you into one thing or control you. Or, if you are not Buzzword compliant will pass you bye.
I have never had the time/patients to deal with college and with no formal education find that I have constantly made a much higher income then most of my college graduate peers.
Income is about social interactions, and relating with other people, be it business owner or employee. So if your really smart in all areas and income is what your after this is the area for study. It took me a while to realize that better mousetraps don't bring money, selling them does.
With diverse interests and intelligence, you just need to focus yourself on profitable projects and avoid the interesting money pits. Things like making a telescope, robotics, model railroads, hot rods, and prime number theory unless you plan these for profit (in a realistic way) from the start, are bad. Focus only things that you can see potential profit in the near future, are good.
Another problem is what ever you pick and start doing for other people can get boring really fast. Work like fixing things, programming, or after creating some brilliant invention they expect you to be the full time engineer/scientist thereafter and forego all other interests are a real hazard.
I find being an all around expert in everything from Video Compression, Physics, Thermodynamics, mathematics, mechanical design, electronics, material science, keeps me constantly stimulated.
So I have many small companies that I work for or have founded, (little profit centers) that I just come in and do the interesting-creative-brilliant parts, pass on my knowledge to others and move on.
They key for anything is a long-term consistent effort. The odds of instant success are very low. Usually I find the technical part to be the most straightforward part but success is anything but instant. Even when creating lots of diverse things, I can't just go around starting projects then losing interest once I know the answers. I may start many things, but also make sure that I have others that I can then pass it over to when the interesting parts are worked out. And I must still make myself available so that they can stay on track and follow through.
Virus/firewall (Trend Micro) cygwin Hummingbird exceed putty lemmy Visual Studio Mozilla winzip vnc divx quicktime player Windows media player trillian
Meshcast Live Peer 2 peer audio
on
Usenet Audio
·
· Score: 1
I had comeup with a way to do live streaming audio over a P2P network (without file sharing) and tried to raise funding, but P2P had such a bad name at the time I was forced to drop it,
those iterested can see more at www.livecamserver.com and www.ecip.com
There is a group I was talking to under NDA that has an secure electronic cash system patent pending. Their same scheme would allow for secure electronic voting. And would allow for 100% total electronic audits that would be almost impossible to cheat their system.
There site is www.decash.com , but there is very little information on the site right now.
If Microsoft posted the code then in a few years they can sue the Linux companies like SCO is doing claming that someone contaiminated the Linux source tree with it.
The responsiblity should be one the mail package writers. Not the novices who attempt to run a mailserver off there dsl and the experts who don't really have this problem.
If all mail packages reported a sucessful relay for any connection but then would either just toss the relayed message into/dev/null (delete it) or attempt to E-mail back a failure message, SPAMMERS whould have a hard time findind a relay. Or would have to expose valid E-mail address when testing for relays, there by allowing them to be tracked down and/dev/null'ed;).
But it's even harder then this. Many spammers own there relays and operate them from off shore in asia/russia/south america and other places where there is little authority or desire on the part of the local authorities to shut these people down. The burden is then on us or the incomming fiber/backbone providers/ and or ISP's to block the IP of spammers. Attempting to get individual Mail server operators to track and block hundreads of thousands of spam servers from off shore just isn't practical...
One word, Linux wouldn't be what it is, or even the entire internet for that matter, if it were not for BSD. BSD was until reciently been considered a research OS and the TCP/IP stack was prototyped there.
Much of early linux borrowed from Net2 and 386BSD and later replaced this code.
Linux also had similar growing pains but lack the copyright problems that held back BSD allowing Linux to capture the lions share of the market before BSD was able to be officialy free of AT&T.
Linux has also has similar infighting but is was more obscure at that time, and so most people aren't aware of that. While the BSD was very visible while its fight were occuring.
And by most accounts BSD has alway been more stable then Linux. Why Else whould Apple choose this for the base of there new OS?
It didn't strike me till this third movie that this was just a modernized version of Tron.
A lot was altered but overall had much of the same concepts.
Did Neo really die in this one? it wasn't clear, and how did the machine know what was really going on?
With agent smith essentilay taking over all man kind how would this effect the power output for the machine city? I would think that power output would be it's only concern...
Sun had there Burritotool and Pizzatool back in 1991 There is some discussion here. http://slashdot.org/articles/02/08/22/17322 01.shtm l?tid=154
From most of the Sun office, Sun employiess could run this tool on your workstation and it would fax an order to the appropriate restaurant that would deliver it right to your office. Very good for keeping programmers from being distracted by things like lunch, dinner, or going home even.
Anyhow wouldn't this be location specific services and information?
Or for that matter wouldn't yp.yahoo.com also predate the MS patent?
I have seen a Unix Blow dryers in Korea, I really wanted to bring one home but didn't have the time to.
Here is the Unix brand UN-1421 blow dryer. Goto http://www.lgeshop.com and search for UN-1421 and click on the image of the blow dryer that come up. Because of there Java Server pages a direct link doesn't work.
Or
A copy of the image on my server. http://www.unixprogram.com/images/1092924 _N1.jpg
You really need to see the box this thing is shipped in.
I have also a copy of the Linux brand paper towels one line. http://www.unixprogram.com/images/linux.jpg
The grid being a distributed system of power generators. Each power plant is independent and is governed by a simple set of rulls, but collectivly the overall system gives rise to complex and unexpected behavior. Stephan Wolfram goes into this sort of thing in a New kind of science, but I have also seen this in Peer to Peer network or even the early studies of TCP trafic flows causing ocilations. Also highway trafic simulations exibit this collective behavior that is now well understood. The problem here stems from power plants haveing inadequate monitoring and control to allow then map the power loads and to trigger small local rolling blackouts and predict internal and external power loads before they get out of hand.
Currently the whole power generator kicks off line! causing them to take hours to restart. This make sense if there is a crowbar short in some of the lines directly out of the plant, but not if it's an external surge triggered from another grid.
If it's just memory like the CMOS or something like the little flash on the ethernet cards that hold the mac then is still would be accessable.
Only a locked fpga or microcontroller or antifuse proms like atmel sells would be difficult but it's based on a secret, that would be what? A DES key or something, it only took 2 year to crack the des key for DVD so I don't think it would really work in practice. There will alway be the option to write a bogus sound driver the dumps raw samples to disk or a video over lay driver or direct show filter that does the same.
I mean you just can't lock someone out of a PC with an endless varity of reverse engineering tools available. The only way is for the media(video audio) to pass through the computer without decryption or decodeing straight to a closed output device that did all the decoding internaly and not on the host processor and never passing data across the host processor.
I have done this several times in the past and what happens is you then have an application trapped on a non-standard system. Any system that it runs that application will need your kernal patch with the new calls.
This gets even worse when the OS keeps upgrading and you are forced to migrate your changes up to be able to use current hardware. (this is need when all of the supported hardware is no longer available new)
In practice it turn into a major undertaking everytime a new OS release came out.
It's is a cool thing to have fun with. But think twice before you base a product or application on a kernel change. (unless you can get the main development tree to adopt it.) Or have the resources to maintain your own OS development team.
Adding new syscontrols and sockopts are also great fun.
At one point we had a versions of FreeBSD that could run DES encrypted Binaries, access the hard driver serial numbers and Mount a CD from HTTP or FTP connection, transmit Datagrams masquerading as TCP connections and be able to process Router Alert packets.
Slashdot Mirror
I mean you take the "PING" program and run ping, you see 90% packet loss with a 6 seconds average latency, I believe I still have the files laying around from that data ping xxx.com > pinglog
L ec tures/forhtml/node3.html
Anyhow when you start to study Error corrector/erasure code you quickly realize that at best you can only reduce errors, IE a (DB increase if you will in S/N) By increasing used bandwidth you increase S/N. Shannon's law type of stuff.
Anyhow If can be proven mathematically that no matter how much you add you can not get 100% perfect data cross. Unless you "close the loop" using feedback as to what was lost, hence TCP.
Now I found you can and combine Error Correction and retransmission to get 100% of your signal/data across. And by using the right type of error correction you didn't need to increase latency to do your retransmissions. (The how, would be a very long discussion to get into that I'd want to take offline)
Now if we take this further I can deliberately strip my codes across a connection to intentionally cause a predictable and manageable amount of packet loss.
For example I know I have a T3 (45Mbs) on one side and an dual channel bonded ISDN on the other 128Kbps (this shows how far back I had worked on this) I know that the network between my T3 and the ISDN is probably better then 10Mbps at all point up until the "last mile".
Now let say that in 1996 MAE West was experiencing 20% packet loss (I have logs of it) almost every afternoon, but I still needed to get my data to some ISDN line on the other side of this gigabit router that is over saturated with bazillions of TCP connections.
Is this situation what if I use an error correction code set to recover 25% loss but with retransmissions? If I send 128K I'd have 102Kbps arrive. Then there is the overhead of the code itself to deal with. But knowing where my loss if occurring at I can deliberately send 25% more data to that connection. So about 170 Kbps gets sent. 20% lost in the backbone and I am guaranteed that the remaining 5% will be lost at the ISP's ISDN terminal server before being sent down the ISDN line. So I will have a known state of 25% lost data, but 128Kbps going over my connection.
Are you following me? What ever the lost across the backbone is become irrelevant since the lost at the connection will force me to at 25% loss.
Now what If I were to use a perfect code (turn out I found a number of perfect erasure codes)
Perfect codes
So start with 128K create 170K, I lost X% across the net lost 25% -X at the ISP and receive 128K. Using a perfect code for 25% loss I recover 100% of my data that needed 128Kbps to get across.
But this only works using retransmissions that are placed back into the error corrected stream! -- YES, Make sure everyone know John Sokol wrote this first since this was the big revelation that made this work!
By doing this it allows you to increase the hamming distance for the next error recovery so you can tolerate future lost data better. While recovering 100% of the data sent first pass.
Lets take the Hamming (7,4) Code it's a perfect code. It has a hamming distance of three. This allow you to correct 1 error and detect 2.
http://www.dcs.st-and.ac.uk/~sal/school/CS3010/
For error you can only correct for 0.5 - 1/2 the hamming distance, but for Erasures you can correct for 1 - the hamming distance.
So I can correct for 2 lost packed out of 7 I send, but only 4 are data with 3 added overhead.
But these are block codes, and are not very efficient or effective. So used brute force and later Genetic algorithms to search for long perfect cyclical codes for Erasures.
The closest thing to the codes I found is called "hagelbarger codes"
Well the Gnutilla seems to be working great, excellent Idea.
I'm glad you asked.
Yes I mean just that.
90% packet loss according to ping. (round trip)
Didn't have a tool to measure each direction independently at that time. (do now) With almost 10 seconds round trip time for that event!
It all depends on where the bottle neck is and what's behind it. At the time we had characterized the way packets were dropped by differnt routers when the buffers were full and overflowing (dropping packets).
Anyhow since we had 10 Mbps ethernet to the 64K link out of Sri Lanka, then on the other side which was MCI in New York ( I had also managed to place a Co-Lo server very close to that router also) there was only one bottleneck that I didn't have control over. the public internet pipe at 64K traveling over fiber 12,000 miles, I could get the ISP to dissconect the rest of thier coustomers just for me.
You can still get the data through a link like this by having the sender push out almost 600 kbps in packets to get 60Kbs of my packets out on the receive side! This is a RUDE Protocol..
So of 64kbps only 4Kbps packets was other people junk and 60 kbps was mine, this is done by saturating the link with my packets! So does this sound familiar? Sort of like a DOS attack maybe...
Now the real trick was to make the 1/10 of the packets that did make it through turn into the 60Kbps of video data that was trying to send over this link with 0% no loss and low latency! So no retranmissions!
We did this!.
For anyone really willing to put some time into this opensourcing this project I will be happy to go into as much detail as they can handle or need to reimplement into an open and more usefull application.
A knowledge of statistics would also help since I ended up learning it for this project myself.
John
ecip.com I call it Error Correcting IP, and used it to stream live video from Sri Lanka in 1997 with Arthur C. Clarke Hal's Birthday
it was a 64K shared line with 90% packet loss, I received 60Kbps for the video stream. ( I have the video to prove it )
We even filled preliminary patents on this back in 1996 but they were never followed through with.
Luigi Rizzo (now head of the FreeBSD project)also did some excellent work on this also. http://info.iet.unipi.it/~luigi/fec.html
He calls it Erasure codes.
Which is more accurate since UDP doesn't have errors, it either come across 99.999% perfect or not at all.
So there is more information then in an error situation where ever bit is questionable.
What this means almost 1/2 the hamming distance in the codes in needed to correct an errasure verses and error.
Turns out the Error/Erasure correcting scheme it critical and not obvious. I spent almost 5 years working on this part time before it started making some real breakthroughs.
My original system was designed for 25% packet loss (not uncommon in 1996).
In the inital idea we added 1 exored packet for every three data packets, but at 25% packet loss, it turns out that it didn't increase reliablity at all! Working this out with probablities was a major eye opener!
Even when you work the problem out you realize you will still need some retransmissions to make up for lost packets, there is no possible solutions without this.
I have been trying to find people to help opensource this since I have working far too hard just to survive since 2000 to even consider taking on another task.
Anyone interested in my research and carring this forward please see my site and contact me.
John L. Sokol
At my Enumera Project We used a 100 CPU Beowulf like Cluster to compress Video in realtime using a several different codec technologies including wavelet.
The idea is that TV stations think nothing on spending $100,000 on hardware, but people streaming web video keep trying to use a cheap > $2000 PC. We tried the other approch. The more CPU power a compression algorythem uses the more efficient the compression. This is almost a rule, not counting optimization(this is equal to adding CPU power). There is a limit to this, and the code needs to be able to use this.
We were able to compress 1080i HDTV with H.263l in real time.,Looked decent at 1Mbps. I'm sure doing this with the Dirac codec with this would also work.
Lyons Gate and AMC with their proprietary Digital Theatre Distribution System (DTDS),
is directly going against DCI - Digital Cinema Initiatives that is made up from Disney, Fox, MGM, Paramount, Sony Pictures Entertainment, Universal and Warner Bros.
They are fighting for control and standards for the new Digital Cinema.
AMC's approch was very slick, they started puting low res tv add up, and deploying these digital projectors then very quickly are pushing movies out. I can't find any info on what AMC's resolution or projectors or or the Satellite system used.
DCI is using microspace or Huges for it's system and has standardized on 2K projectors 2048x1080 this is about where HDTV 1080p/24 is 1920x1080.
DCI also supports 4K 4096x2160 , but from my visit at there test bed, the USC, ETC center they were using 1024x768 video to drive everything.
I have a lot more written on this at
http://www.videotechnology.com/0904/formats.html
http://www.videotechnology.com/old0904.html
http://www.videotechnology.com/old1004.html
http://www.videotechnology.com/old0804.html
http://www.videotechnology.com/old0604.html
Hey Publishing in JAVA is practicaly Open Source since the code decompiles back to very clean source with original variable names and everything so most JAVA players are derivative works. See Mocha or JAD for decompilers.
BTW I am assembling a list of all Java streams systems
Reinvented really.
4 /0 9/10/2053245&tid=108&tid=97&tid=95&tid =1
The Livecam server we developed in 1995 and dominated the adult industry already did all this and supported more viewers with better quality.
We supported Motion JPEG or H.236 in 1999 with GSM audio, with 20Kbps to 70 Kbps streams.
I just love it when someone else come out with it all over again and everyone thinks it's new.
----Original Message-----
From: James S
Sent: Friday, September 10, 2004 5:18 PM
To: sokol@videotechnology.com
Cc: Jesse Monroy
Subject: Hey These guys just invented the Java player we created in 1999
Check this out. It's our player.
http://developers.slashdot.org/article.pl?sid=0
James
SpaceShipOne Flights are planed for September 29th, 2004, October 13th, 2004 at the Mojave Airport, Civilian Aerospace Test Center in Mojave,
California.
Press release
I just went up to Mojave this weekend to see what was around and there was almost nothing to see and all was closed.
I Don't think they are in any way prepared for the turnout they are going to get for this next flight.
Rutan's team has a very good chance success being he's already done it with one test pilot. No matter this outcome these will be a historic event.
I have taken several 250GB Maxtor hard drives and poured a thick layer of special rubber on them then been able to throw then out a second story window. Then hook them up and have then work fine.
I have also taken several power supply including an antec 480 watt and remove it from it's case and fill it with epoxy and stuff to increase it's thermal conductivity.
And even saw off the heat sink using a band saw from a $500 ATI Radeon x800 xt graphics card and pour epoxy over it. This was the only part that I was feeling queasy over.
Even unsoldering the heatsinks off of the northbridge on motherboards, and cutting off various connectors.
The end result has been a completely fanless, absolute dead silent computers running at 3.4 Ghz.
www.nisvara.com
Mine you many computers were killed in the learning process of how to do this correctly.
From School to work, the world keeps trying to put people into pigeonholes. Expecting individuals to spend 10 years to become adequate at one profession. A resume that shows expertise in Digital and analog hardware design or windows and Linux software tend to turn many prospective employers off. Just add Machine shop, mechanical design, chemistry, optics, number theory, cryptography to that and now they just don't know what to make of you.
For myself and many people who I have met like me that can rise above average levels
, It is a challenge. There are many people who are threatened and territorial, and others who try to box you into one thing or control you. Or, if you are not Buzzword compliant will pass you bye.
I have never had the time/patients to deal with college and with no formal education find that I have constantly made a much higher income then most of my college graduate peers.
Income is about social interactions, and relating with other people, be it business owner or employee. So if your really smart in all areas and income is what your after this is the area for study. It took me a while to realize that better mousetraps don't bring money, selling them does.
With diverse interests and intelligence, you just need to focus yourself on profitable projects and avoid the interesting money pits. Things like making a telescope, robotics, model railroads, hot rods, and prime number theory unless you plan these for profit (in a realistic way) from the start, are bad. Focus only things that you can see potential profit in the near future, are good.
Another problem is what ever you pick and start doing for other people can get boring really fast. Work like fixing things, programming, or after creating some brilliant invention they expect you to be the full time engineer/scientist thereafter and forego all other interests are a real hazard.
I find being an all around expert in everything from Video Compression, Physics, Thermodynamics, mathematics, mechanical design, electronics, material science, keeps me constantly stimulated.
So I have many small companies that I work for or have founded, (little profit centers) that I just come in and do the interesting-creative-brilliant parts, pass on my knowledge to others and move on.
They key for anything is a long-term consistent effort. The odds of instant success are very low. Usually I find the technical part to be the most straightforward part but success is anything but instant. Even when creating lots of diverse things, I can't just go around starting projects then losing interest once I know the answers. I may start many things, but also make sure that I have others that I can then pass it over to when the interesting parts are worked out. And I must still make myself available so that they can stay on track and follow through.
I install (win2k + patches + drivers)
Virus/firewall (Trend Micro)
cygwin
Hummingbird exceed
putty
lemmy
Visual Studio
Mozilla
winzip
vnc
divx
quicktime player
Windows media player
trillian
I had comeup with a way to do live streaming audio over a P2P network (without file sharing) and tried to raise funding, but P2P had such a bad name at the time I was forced to drop it,
those iterested can see more at www.livecamserver.com and www.ecip.com
There site is www.decash.com , but there is very little information on the site right now.
If Microsoft posted the code then in a few years they can sue the Linux companies like SCO is doing claming that someone contaiminated the Linux source tree with it.
The responsiblity should be one the mail package writers. Not the novices who attempt to run a mailserver off there dsl and the experts who don't really have this problem.
If all mail packages reported a sucessful relay for any connection but then would either just toss the relayed message into
But it's even harder then this. Many spammers own there relays and operate them from off shore in asia/russia/south america and other places where there is little authority or desire on the part of the local authorities to shut these people down. The burden is then on us or the incomming fiber/backbone providers/ and or ISP's to block the IP of spammers. Attempting to get individual Mail server operators to track and block hundreads of thousands of spam servers from off shore just isn't practical...
I don't think I actualy credited Feynman as it's inventor, just as someone who lectured on it....
And now has a book printed in 1996 that seems to do a good job explaining it.
This book by Richard Feynman is based on a series of lectures given at CalTech in the mid 1980s.
In it he discusses Reversible Computation and the Thermodynamics of Computing and quantum computing.
As usual, Feynman was way ahead of his time.
I highly recomend this book.
The basic idea is heat is only generated when information is destroyed. So don't destroy information when performing computations.
How this relates to something actualy practical is hard to say, but it didn't strike me as something that would apply to silicon very easily.
John
One word, Linux wouldn't be what it is, or even the entire internet for that matter, if it were not for BSD. BSD was until reciently been considered a research OS and the TCP/IP stack was prototyped there.
Much of early linux borrowed from Net2 and 386BSD and later replaced this code.
Linux also had similar growing pains but lack the copyright problems that held back BSD allowing Linux to capture the lions share of the market before BSD was able to be officialy free of AT&T.
Linux has also has similar infighting but is was more obscure at that time, and so most people aren't aware of that. While the BSD was very visible while its fight were occuring.
And by most accounts BSD has alway been more stable then Linux. Why Else whould Apple choose this for the base of there new OS?
It didn't strike me till this third movie that this was just a modernized version of Tron.
A lot was altered but overall had much of the same concepts.
Did Neo really die in this one? it wasn't clear, and how did the machine know what was really going on?
With agent smith essentilay taking over all man kind how would this effect the power output for the machine city? I would think that power output would be it's only concern...
Sun had there Burritotool and Pizzatool back in 19912 01.shtm l?tid=154
There is some discussion here.
http://slashdot.org/articles/02/08/22/1732
From most of the Sun office, Sun employiess could run this tool on your workstation and it would fax an order to the appropriate restaurant that would deliver it right to your office. Very good for keeping programmers from being distracted by things like lunch, dinner, or going home even.
Anyhow wouldn't this be location specific services and information?
Or for that matter wouldn't yp.yahoo.com also predate the MS patent?
Look like that site is down.
4 _N1.jpg
g
I have seen a Unix Blow dryers in Korea, I really wanted to bring one home but didn't have the time to.
Here is the Unix brand UN-1421 blow dryer. Goto
http://www.lgeshop.com and search for UN-1421 and click on the image of the blow dryer that come up.
Because of there Java Server pages a direct link doesn't work.
Or
A copy of the image on my server.
http://www.unixprogram.com/images/109292
You really need to see the box this thing is shipped in.
I have also a copy of the Linux brand paper towels one line.
http://www.unixprogram.com/images/linux.jp
The grid being a distributed system of power generators. Each power plant is independent and is governed by a simple set of rulls, but collectivly the overall system gives rise to complex and unexpected behavior. Stephan Wolfram goes into this sort of thing in a New kind of science, but I have also seen this in Peer to Peer network or even the early studies of TCP trafic flows causing ocilations. Also highway trafic simulations exibit this collective behavior that is now well understood.
The problem here stems from power plants haveing inadequate monitoring and control to allow then map the power loads and to trigger small local rolling blackouts and predict internal and external power loads before they get out of hand.
Currently the whole power generator kicks off line! causing them to take hours to restart. This make sense if there is a crowbar short in some of the lines directly out of the plant, but not if it's an external surge triggered from another grid.
If it's just memory like the CMOS or something like the little flash on the ethernet cards that hold the mac then is still would be accessable.
Only a locked fpga or microcontroller or antifuse proms like atmel sells would be difficult but it's based on a secret, that would be what? A DES key or something, it only took 2 year to crack the des key for DVD so I don't think it would really work in practice.
There will alway be the option to write a bogus sound driver the dumps raw samples to disk or a video over lay driver or direct show filter that does the same.
I mean you just can't lock someone out of a PC with an endless varity of reverse engineering tools available.
The only way is for the media(video audio) to pass through the computer without decryption or decodeing straight to a closed output device that did all the decoding internaly and not on the host processor and never passing data across the host processor.
I have done this several times in the past and what happens is you then have an application trapped on a non-standard system. Any system that it runs that application will need your kernal patch with the new calls.
This gets even worse when the OS keeps upgrading and you are forced to migrate your changes up to be able to use current hardware. (this is need when all of the supported hardware is no longer available new)
In practice it turn into a major undertaking everytime a new OS release came out.
It's is a cool thing to have fun with. But think twice before you base a product or application on a kernel change. (unless you can get the main development tree to adopt it.) Or have the resources to maintain your own OS development team.
Adding new syscontrols and sockopts are also great fun.
At one point we had a versions of FreeBSD that could run DES encrypted Binaries, access the hard driver serial numbers and Mount a CD from HTTP or FTP connection, transmit Datagrams masquerading as TCP connections and be able to process Router Alert packets.
John