As I said, it was 2 summers ago. (that was a summer internship, currently working on getting my masters/phd at columbia)
AFAI Remember What I Signed, the NDA didn't last for that long. (6 months was the limit it I think). Also, everything confidential that I said (basically limited to the citizen information) has already been posted in the news (As well as/.)
and if I did release any info I shouldn't have, all I can say is sorry, I made a mistake in trying to show the world IBM's "coolness".
that startup scripts aren't all there, but you should be able to get a shell via busybox, run udhcpc ("micro" dhcp client), and then run startx which would connect you to our windows terminal server, edit the startx file (think it's in/bin) to change that.
Oh, and these shared libraries might be really "Stripped" i.e. the point oft the project was to see how small we could make the C library, by pulling out everything we dont need. (making a bootable terminal services floppy was just sort of the "cool" factor)
IBM wanted to figure out how to put linux in small places. However, what would you rather work on. "Linux on your coffee maker" or "linux on a watch", linux on a watch, has much higher "Cool" factor.
on the IBM linux watch our experiments seemed to imply that using a wheel that you could scroll around (selecting things) and also "click" was a good comprimise, however, one cant do everything that one would do with a palm w/ a device like that.
It's like a tradeoff b/w those REX pcmcia devices and a full fledged palm. Do you want something that you aren't going to do full fledged input on (but can input things into), but is mostly for info retrieval, or something that you actually want to do lots of input into.
I worked on it 2 summers ago at IBM Research (get paid well to work on cool things, who can ask more).
It was never meant to be an IBM product (at least from my understanding). It was meant to be a sort of cool "testbed" for all these mini technologies. Such as OLEDs (at the time we had the highest density OLED displays in production on our prototypes), mini linux on the software side and similiarly on the hardware side.
We were also trying to figure out what cool things we could do with it, such as we had a bluetooth module for it, so we initally had a demo of that that one could switch ones slides via the watch's thumbwheel, (so one's giving a talk, and walking around and just tap the watch to switch slides back and forth). But we were experimenting with different things as well, such as imagine sticking a GPS receiver in your backpack while you are hiking and just use the watch as a display for the GPS, it stores the data in the backpack. Since the OLED display was 640x480 (1bpp) it was fairly detailed.
doing this also involved getting things like PPP working on the watch (which since I had gcc working on our testbed's that had ethernet (nfs mount w/ nfs swap), meant we could get gcc working on the watch, though wouldn't have been so much fun as testbench was 70-80mhz, while watch ran at something like 17mhz (these numbers might be off)
We were also doing work on getting the system even smaller than it was (glibc's a hog, a simple fix was getting uClibc working on it, I was then able to take this know how to make a bootable linux floppy that boots directly into a windows terminal services server full screen (basically kernel, X, dhcp, rdesktop all on a single bootable floppy) for a school project).
That summer there was interest from citizen (They actually made some PR announcmenets) in making a product out of it, but dont know whats happened since in regards to that.
people in charge of IBM Research wanted to put funding behind the watch to give out samples to U's involved in pervasive devices (I believe CMU and Georgia Tech are big into it) to basically have them see what they can do with these pieces of tech, but dont know what's happened on that either.
Not really (at least IMHO) they figured out how to make use of the original DLLs. You will still need the DLL's from a QuickTime 5 installation (as well as wine acc. to the description). This is not reverse engineering the codec, just figuring out how spit encoded frames to the dll and understand the decoded frames it spits back. What was done with the previous sorenson codecs (of actually figuring out how to decode) was much more impressive (at least to me).
They already use the "use dlls" methods for real codecs, except in that case they have the real linux.so's to use.
The only times Galeon has hung on me is when its rasterizing fonts. a quick switch to a terminal and a run of top shows that xfs or xfstt are chewing up all my cpu. google seems to do this a lot to me.
I think wine gets around this (that big wait first time you run wine), not sure why gnome doesnt do the same thing.
so if we end up having hard drives with support for the 4 IDE devices normally supported today, one is going to need 4 connectors? or as some motherboard support 6 or 8 devices, we are going to need 8 connectors? Isn't that going to take up a decent amount of space on the motherboard, not so much for the connectors, but for all the extra wiring on the mobo that's needed.
hmm, slightly not thinking not 1 lb of mozerrella, but 16 oz (fluid oz's I believe, perhaps it equates to lb's some how). Also 8 oz of the shreded mozerella.
1 9*12 roaster/baker aluminum pan 1 lb of ziti noodle 1 jar of marinara sauce 1 lb of Ricotta Cheese 8 oz of Mozerella
boil water, put noodles in water, cook for the 10 minutes (or as per directions).
drain water, dump noodles into pan, dump sauce into pan, dump cheese into pan. mix well.
layer on top with mozerella.
bake in oven at 350-400 for 30 minutes (or until done). Depending on how you can cover your ziti pan with aluminum foil. Stores and reheats easily. Can make 2, 3... pounds of nododles at once. (I normally do 2, lasts me an entire week)
I also think they tinkered with it during the match, which annoyed kasparov, as the program that started the match wasnt the same as the one at the end.
Not exactly sure why prices in CA would be for the double gulp, but here in NYC one can easily fine 2 liter bottles of soda for a buck a bottle. that seems to be the standard "sale" price, and many times it cheaper, like this week at my local super market one can get a 2 liter bottle of pepsi for 85 cents.
The large majority of CDEX's usage is legal. As you have the physical CD, and are ripping it. Assuming you own it (as one would assume most rips are) then it's 100% legal. The "illegal" part would be either distributing the mp3s (cdex not involved) or ripping a CD you don't own.
I wonder if people having no use of their floppy drives can be in any way corollated with AOL not inunandating us with free writable media.
AOL floppies were so usefull in the day of multi floppy debian installs (to get base up and running, then network to get the rest, even ppp was sufficiant). now one just has to jidgo down a cd image and burn it.
One could release their code under the GPL and BSD licenses. That would solve that problem. Under problem would be if you wanted to use an "Open Source" license that is more restrictive than the GPL.
Slashdot Mirror
The problem I have with the idea is basically you would be designing your home in the likeness of a sewer..
there you go encouraging the teenage mutant ninja turtle demographic.
Not really formatted well, just copied everything off the hot deals forum at Anandtech (as they havent taken down the ads yet)
http://www.columbia.edu/~sp2015/blackfriday.txt
the file probably has somewhere in the 400-500 prices. I'd be willing to put more there, so just email me.
As I said, it was 2 summers ago. (that was a summer internship, currently working on getting my masters/phd at columbia)
/.)
AFAI Remember What I Signed, the NDA didn't last for that long. (6 months was the limit it I think). Also, everything confidential that I said (basically limited to the citizen information) has already been posted in the news (As well as
and if I did release any info I shouldn't have, all I can say is sorry, I made a mistake in trying to show the world IBM's "coolness".
as I told the other person, believe what you want. you could confirm easily enough if its on my resume or not.
here's what I was able to scrounge up (kernel only has 3c905 driver installed)
/bin) to change that.
http://www.cs.columbia.edu/~spotter/floppy.bin
that startup scripts aren't all there, but you should be able to get a shell via busybox, run udhcpc ("micro" dhcp client), and then run startx which would connect you to our windows terminal server, edit the startx file (think it's in
Oh, and these shared libraries might be really "Stripped" i.e. the point oft the project was to see how small we could make the C library, by pulling out everything we dont need. (making a bootable terminal services floppy was just sort of the "cool" factor)
think what you want.
you are sort of right. :)
IBM wanted to figure out how to put linux in small places. However, what would you rather work on. "Linux on your coffee maker" or "linux on a watch", linux on a watch, has much higher "Cool" factor.
on the IBM linux watch our experiments seemed to imply that using a wheel that you could scroll around (selecting things) and also "click" was a good comprimise, however, one cant do everything that one would do with a palm w/ a device like that.
It's like a tradeoff b/w those REX pcmcia devices and a full fledged palm. Do you want something that you aren't going to do full fledged input on (but can input things into), but is mostly for info retrieval, or something that you actually want to do lots of input into.
I worked on it 2 summers ago at IBM Research (get paid well to work on cool things, who can ask more).
It was never meant to be an IBM product (at least from my understanding). It was meant to be a sort of cool "testbed" for all these mini technologies. Such as OLEDs (at the time we had the highest density OLED displays in production on our prototypes), mini linux on the software side and similiarly on the hardware side.
We were also trying to figure out what cool things we could do with it, such as we had a bluetooth module for it, so we initally had a demo of that that one could switch ones slides via the watch's thumbwheel, (so one's giving a talk, and walking around and just tap the watch to switch slides back and forth). But we were experimenting with different things as well, such as imagine sticking a GPS receiver in your backpack while you are hiking and just use the watch as a display for the GPS, it stores the data in the backpack. Since the OLED display was 640x480 (1bpp) it was fairly detailed.
doing this also involved getting things like PPP working on the watch (which since I had gcc working on our testbed's that had ethernet (nfs mount w/ nfs swap), meant we could get gcc working on the watch, though wouldn't have been so much fun as testbench was 70-80mhz, while watch ran at something like 17mhz (these numbers might be off)
We were also doing work on getting the system even smaller than it was (glibc's a hog, a simple fix was getting uClibc working on it, I was then able to take this know how to make a bootable linux floppy that boots directly into a windows terminal services server full screen (basically kernel, X, dhcp, rdesktop all on a single bootable floppy) for a school project).
That summer there was interest from citizen (They actually made some PR announcmenets) in making a product out of it, but dont know whats happened since in regards to that.
people in charge of IBM Research wanted to put funding behind the watch to give out samples to U's involved in pervasive devices (I believe CMU and Georgia Tech are big into it) to basically have them see what they can do with these pieces of tech, but dont know what's happened on that either.
Not really (at least IMHO) they figured out how to make use of the original DLLs. You will still need the DLL's from a QuickTime 5 installation (as well as wine acc. to the description). This is not reverse engineering the codec, just figuring out how spit encoded frames to the dll and understand the decoded frames it spits back. What was done with the previous sorenson codecs (of actually figuring out how to decode) was much more impressive (at least to me).
They already use the "use dlls" methods for real codecs, except in that case they have the real linux .so's to use.
The only times Galeon has hung on me is when its rasterizing fonts. a quick switch to a terminal and a run of top shows that xfs or xfstt are chewing up all my cpu. google seems to do this a lot to me.
I think wine gets around this (that big wait first time you run wine), not sure why gnome doesnt do the same thing.
well, here's Bill Gates denying it.
http://www.nybooks.com/articles/15180#fn*
that would be like scsi. and as there's a "serial scsi" protocol in the works, it doesnt make sense. Then again, when do things have to make sense.
so if we end up having hard drives with support for the 4 IDE devices normally supported today, one is going to need 4 connectors? or as some motherboard support 6 or 8 devices, we are going to need 8 connectors? Isn't that going to take up a decent amount of space on the motherboard, not so much for the connectors, but for all the extra wiring on the mobo that's needed.
hmm, slightly not thinking not 1 lb of mozerrella, but 16 oz (fluid oz's I believe, perhaps it equates to lb's some how). Also 8 oz of the shreded mozerella.
Baked Ziti.
1 9*12 roaster/baker aluminum pan
1 lb of ziti noodle
1 jar of marinara sauce
1 lb of Ricotta Cheese
8 oz of Mozerella
boil water, put noodles in water, cook for the 10 minutes (or as per directions).
drain water, dump noodles into pan, dump sauce into pan, dump cheese into pan. mix well.
layer on top with mozerella.
bake in oven at 350-400 for 30 minutes (or until done). Depending on how you can cover your ziti pan with aluminum foil. Stores and reheats easily. Can make 2, 3... pounds of nododles at once. (I normally do 2, lasts me an entire week)
Currently, a Copyright Royalty Arbitration Panel (CARP) meets once every two years to decide on royalty fees for web radio broadcasters.
Does one expect non CRAP'y decisions from a panel that's named like this. They even changed the acronym so we wouln't realize this.
I also think they tinkered with it during the match, which annoyed kasparov, as the program that started the match wasnt the same as the one at the end.
Not exactly sure why prices in CA would be for the double gulp, but here in NYC one can easily fine 2 liter bottles of soda for a buck a bottle. that seems to be the standard "sale" price, and many times it cheaper, like this week at my local super market one can get a 2 liter bottle of pepsi for 85 cents.
The large majority of CDEX's usage is legal. As you have the physical CD, and are ripping it. Assuming you own it (as one would assume most rips are) then it's 100% legal. The "illegal" part would be either distributing the mp3s (cdex not involved) or ripping a CD you don't own.
I wonder if people having no use of their floppy drives can be in any way corollated with AOL not inunandating us with free writable media.
AOL floppies were so usefull in the day of multi floppy debian installs (to get base up and running, then network to get the rest, even ppp was sufficiant). now one just has to jidgo down a cd image and burn it.
This hand powered webserver was featured on slashdot awhile ago. Personally I find this cooler, even if it is less usefull.
3 integer variables a, b, c
(a & b) | (b & c) | (c & a)
so the "and's" are any 2 of the variables, if the bit is true, it will be one, and the "or's" will accumulate the multiple "votes"
anyways, thats just what I thought of off the cuff, perhaps thats the slow way you mentioned.
reading above comment, it seems that perhaps I'm wrong, if they link to a gpl'd library, must be gpld, can't dual license it.
One could release their code under the GPL and BSD licenses. That would solve that problem. Under problem would be if you wanted to use an "Open Source" license that is more restrictive than the GPL.