IBM Says Polymer Memory Could Be Ready By 2005 145
prostoalex writes "Polymer memory is hardly anything new, and we already had HP and Princeton announcing their prototype. In a Forbes magazine article IBM promises polymer memory that's five times cheaper than current flash memory, and expects the first devices with polymer data storage systems to be delivered possibly by 2005. IBM's Zurich Lab published this article last year with description of Millipede."
5X Cheaper? (Score:3, Insightful)
Re:5X Cheaper? (Score:3, Insightful)
Re:5X Cheaper? (Score:5, Insightful)
The interesting point about the semiconductor market that makes it different from almost anything else is that it pay's to drive down the price as your cost decreases.
The reason being that the market thereby grows at a much faster rate, more than compensating the price drop. Remember Profit in $ is Unit profit * Unit.
Just look at Cell Phones as a good recent example. Industries has repeatedly learned that artificially holding the price high kills you. If interested in more info Google Clayton Christenson + Disruptive technologies.
Comment removed (Score:5, Insightful)
Re:Hmmm. Cheap long term storage? (Score:1, Insightful)
this is 1 - 2 years' time we're talking about, not 10. granted, technology moves forward quickly, but not THAT quickly. this new tech might not be the greatest thing since sliced bread, and it probably wont deliver on all the hype (like the price), but if it is better than even one of the three other competing technologies you mentioned (and no, it probably wont be obsolete by then), that's still a good thing.
Re:Hmmm. Cheap long term storage? (Score:5, Insightful)
To capture the market, this stuff has to either be:
1. Cheaper than flash or HDs.
2. More durable than flash or HDs (or even CD/DVDs)
3. Be faster than flash/HDs/optical media.
Nope. Read The innovator's dilemma [amazon.com]. All it has to do is:
- Have room for improvement
- Serve a niche market that the others can't
- Improve over time into something they aren't
Micro-computers (to use one of his examples) weren't cheaper (for the power), more durable, or faster than big iron. But they came in smaller increments and could serve markets that the big players couldn't...-- MarkusQ
This is why video compression will soon not matter (Score:3, Insightful)
Re:This is why video compression will soon not mat (Score:3, Insightful)
Re:This is why video compression will soon not mat (Score:2, Insightful)
1) Bandwidth. Even if storage increases by the factor of 20 you envision, that doesn't mean bandwidth will.
2) Would you rather not compress your video, or have better quality and more of it? Uncompressed video is RIDICULOUSLY large. 640 x 480 x 24bpp x 30fps = 221 megabits a second (27 megabytes a second) That means you can store about an hour of that in 100 GB. DVD quality is about 30:1 compression.
I don't think this would cause people to keep uncompressed AUDIO around (where audio is only 150K/s, compressed maybe 10:1 or or so); basically the cost of compression is pretty small. If your other resources are finite, it makes a lot of sense.
Re:This is why video compression will soon not mat (Score:3, Insightful)
Re:This is why video compression will soon not mat (Score:4, Insightful)
When ten movies fit on one disk/whatever without compression, giving crystal-clear video, no one will think "yes, let's compress that!".
No, when that happens, everyone will think "Hey! Let's increase the framerate, increase the pixels and increase the color depth, then compress it all so we can fit 100 of these better movies on the same device!". Currently, uncompressed NTSC video is about 18MBps, or 144Mbps, which is 18 times more data than a DVD video stream (which includes audio, subtitles and control data as well), and DVDs use the old MPEG2 compression algorithm. If you see occasional compression artifacts in your DVDs, you can be sure that if they compressed to the same data rate using MPEG4, the result would be perfect.
Looking into the future, assume we double the frame rate, increase resolution to 1080 lines, increase the color depth to 4 bytes per pixel, and store full-raster data, then the video data rate increases to about 300MBps. That would make an uncompressed two-hour movie over 2TB in size. Assuming storage sizes continue to double every 18 months, 2TB disks (or whatever) should be commonly available in 15 years. To get 10 uncompressed movies you'd need 20TB, so add another 3 years or so.
OTOH, if we can get 50:1 compression, that 2TB movie becomes a 43GB movie, and your 10-movie storage device is only a year or two away (since 200GB drives are pretty cheap now).
Further, it just doesn't make sense not to compress video. There is so much redundancy that can be discarded. I mean, even stills can be compressed dramatically without degradation, and think about how much similarity there is between each video frame and the next. Good codecs like MPEG4 can achieve 100:1 compression ratios with some degradation, or 50:1 with no perceptible degradation at all, and we can probably expect that to improve.
Video will be compressed. It's just dumb not to do it.