Perhaps Disney will start seriously wooing John McCain, now that he's replaced Fritz Hollings as the Senate's Commerce Committee chair?

This is great stuff, and I found a more technically detailed article article on physicsweb.org.

It is curious why this press release is out now. The first paper on the subject came out a year ago.

I am unclear how this advance (which is really cool, don't get me wrong) leads to tremendous storage increases--the toughest long term issue in magnetic storage is 1) making disks, and 2) profit. Until those are solved, I am betting that flash memory variants are the way of the future.

Maybe I'm just out of touch, but Im missing what the tough long term issue is about making disks.

The physicsweb.org article claims that the storage capacity in bits per square inch is larger with ballistic magnetoresistance read heads than with current heads using giant magnetoresistance, which would be the "tremendous storage increase" in question.

As to flash memory, if it's possible to manufacture a flash memory storage cell that's smaller than the magnetic "spots" on the hypothetical disk platters, then flash could have greater storage capacity per square inch than BMR disks. It doesn't seem likely, on the face of it, that flash memory would be cheaper to manufacture (and thus easier to sell at a profit) because each bit of flash memory corresponds to an individual electronic device (capacitor and transistor in the "permanent" part of the flash memory Silicon, at a minimum), while a hard disk supports a fairly large ratio of storage location to active devices.

So, what am I missing here?

(P.S. Am I slipping even worse than that would suggest, or is remembering personal information working now?)

How I would use great heaps of capacity:

My home PC, a rather modest beast, has about 5 gigs of Stuff, ranging from binaries to photos to five year TurboTax files to ancient, ancient snippets of BASIC code.

I would love to be able to snapshot this information every few weeks. Save and timestamp the whole shabang.

OTOH, as impressive as great heaps of storage are, what we really need are better batteries.

If "remember personal information" is working for some users, great; chalk it up to the upgrade to MT 2.6. It still doesn't work for me.

Not working for me, either, as it turns out, just an artifact of commenting via the page retrieved using the permalink.

Hi Bob,

First, apologies in advance if I sound a bit didactic and lame-teacher like. I have no idea of your background, so I may be being insultingly simple. Apologies in advance and all.

Second, take what I say with a grain of salt--I worked on simple read heads, and know a few disk and head engineers--but I am NOT an absolute expert--what follows is just what I think is the future. I would be happy to be wrong on this one.

Sufficiently disclaimed--I think so!

The sensitivity of the head is not necessarily the determining factor in increasing storage density. You can read smaller bits with a BMR sensor, but we are getting awefully close to the absolute physical limit on how small a single magnetic bit can be made(the superparamagnetic limit). Even if we can sense super small bits, we can't really make them dense enough.

There are tricks to this--perpendicular recording, antiferromagnetically coupled media ("pixie dust") and patterned media being the best options I know of--but these are very hard. And they lead to all sorts of trickier problems in the long run with writing the data. In a more subtle way, the creation of patterned media requires the same processing as silicon--making the whole cost-equation really tricky. It is this last issue which makes me think that universal storage will be Flash (and with Flash, add in MRAM, FeRAM, and the other Non-Volatile, semiconductor processing based variants).

So my personal opinion is that some variant on flash memory will win the day in the longer run--10 years. For the middle term, BMR may be the trick, especially if it gets sufficient temperature stability and yield. It may be that this will really help integration, and lead to much, much cheaper heads. Right now, head fabrication is an extraordinarily difficult process, requiring hundreds of steps. So even if we don't get vast amounts of extended storage out of BMR (because making dense disks ends up being harder than making good sensors), we get lots more hard drives that are very cheap. Which is almost, but not quite, the same thing.

And hell, maybe someone will finally make a really top notch, super fast holographic system. That would be nice, too.

Indirectly related, for those interested, is this piece by Charlie White, touting the coming revolution from 32-bit to 64-bit. This is more about speed than data space, but we digital video freaks get really excited about it.

Here's hoping his prediction that Apple will be ready with its new IBM-based 2GHZ processor models by the end of this summer is on the money.

Brennan, that was great! I haven't really been keeping up with development -- I think I stopped tracking around the time IBM (?) came out with "pixie dust," and wasn't aware of a lot of the current platter-coating technologies you mention. The stability and fabrication problems for the little Ni clusters in the BMR heads did occur to me, but I figured that it would have been covered in one of the articles if it were an obvious obstacle to development.

Some quick-and-ignorant googling yields results that suggest various teams are working to break the superparamagnetic barrier to areal density, so rotating magnetic media are perhaps still in the race, but I take your point about the manufacturing steps required to produce platter surfaces being comparable with those for active devices in Si. And I'm guessing that fabricating write heads for some of the proposed high-density technologies is going to be quite challenging, too.

Anyway, thanks for helping to answer my question about what I was missing.

Another approach to maxi-storage in mini-form, using micro-machines, sort of Hollerith cards you would need an electron microscope to see:

IBM's 'Millipede' Project Demonstrates Trillion-Bit Data Storage Density
Store the equivalent of 25 DVDs on a surface the size of a postage stamp

and another story. This stuff seems seriously wack, yet: spinning disks with bearings and read heads, all seems so Edisonian. The future must be solid-state -- unless it's these little teeny punchcards! Niven had a song, "Little Teeny Eyes."

OT: Congrats on the Skylark!

>Niven had a song, "Little Teeny Eyes."

Nope. Tom Digby.

"...And you need little teeny eyes for reading little teeny print like you need little teeny hooks for microfiche..."

I dunno which technology will win out (people have been predicting the demise of disks in favor of some kind of solid state memory since, oh, about 2 hours after the invention of the integrated circuit) but I do think the consequences will be interesting. When this general topic arose a while back on RASFF, I suggested the possibility of keeping a permanent record of one's entire life -- video camera "glasses" to record what you see, and stereo mic's to record what you hear. It's both increasingly feasible, and, frighteningly, may be increasingly necessary as the only way to defend oneself against arbitrary accusations. But then, what about subpoenas of your "diary"?

We have met Big Brother, and he is us?

I've been meaning to read David Brin's "Transparent Society" book for a while, a thought recalled by Jordin's mention of ubiquitous personal recording. One point in favour of the proposal for recording one's surroundings at all times is that if other agencies are doing it anyway, the additional loss of privacy incurred for personal recording is small and the lack of a recording of one's own view could be rather like being the only party to a legal action not represented in court.

Atomic Holographic Nanotechnology Sets the Pace for the Future

> 40,000 terabits / cu. cm. or

>40,000,000 gigabits/cu.cm. !!!!