Re:Any advantages over having only one connector?
on
eSATA Connectors
·
· Score: 1
Yep, the mistakes of SCSI are being repeated. When Apple put SCSI on their boxes in 1986, there WAS NO STANDARD for the external connector, and DB-25 (which encouraged malfunctional printer-cable and serial-cable connections) and 'Centronics' (actual Micro Blue Ribbon-50) with the funky springlatch were rampant for years. Neither was crimpable directly to ribbon cable (but the Centronics cables grew an expensive variant that converted the incompatible spacing, so lotsa folk didn't know of that problem).
Then a standard was issued for the external connector; it only crimped onto fine-pitch ribbon cable (everyone had to switch over), and used a funny spring latch.
But SCSI was used for complex systems that lived in wiring closets, you CAN'T tolerate loose connections, and those spring latches were terrible. With SCSI-3 came a new external connector standard, with a wider bus AND screwdown connector latches. Took three tries to get it right.
So here we see SATA with standard specs for internal connectors only. Folk need external, so they cobble a shield around the internal (but the unshielded cables still fit-- OOPS!). The standard gets an update that requires a shield and enforces it with a new plug, but uses flimsy latches. Consumers will like it for temporary plugins, but the serial-attached-SCSI types are claiming they interplug with SATA, and I just know a third connector will be issued for the high-end folk, and it will be a repeat of SCSI.
What a waste of money, brains, time (WOMBAT).
I once labeled all the different SCSI connectors I'd seen used. A, B, C.... got to about S. There's about 50 pounds of cables in my basement, with various ends, all in the boxes marked SCSI.
>Firewire? Why do enterprise desktop users need firewire? The only reason you need it is for digital video and >audio or extremely fast file transfers. Not desktop use.
That's not my experience at all! A dead or dying Macintosh can be rebooted into target disk mode and all its data transferred to another machine via that Firewire port. Or you can use it for TCP/IP at 400 mbits/second, or to attach a port-powered bootable hard disk.
None of those uses is insignificant, nor can USB be considered a suitable substitute.
Any system admin with a bunch of Macintoshes should buy a laptop-size Firewire-powered emergency disk drive. Buy one, you'll never be sorry. Won't do much good on PCs, though; gotta use the USB for them, bootability is problematic, power cord required, etc.
>They had USB... you could just plugin a USB floppy drive if you needed it.
There was also a header connection, on the logic board, and with some sheetmetal cutting it was possible to wire to it and run a Macintosh floppy drive.
The Macintosh floppy disks used 400k and 800k formats that required multispeed spindle motors, and WEREN'T playable or recordable on 1.44M drives, but the Macintosh mechanisms handled all three types. The USB plugin drives only ever handled 1.44M. They wouldn't read the original Mac 400k or 800k format.
Imation's SuperDisk drive handled both 120MB cartridges and 1.44M floppies, but that was the last gasp of compatible flexible disks for USB. Zip drives, still common, are the only modern-ish floppy. I think they're up to 750 MB?
>It doesn't matter at all because the vast majority of business applications are not available for the mac.
No person, no business, needs 'the vast majority of' available software. Macintosh computers work and are priced fairly. Some arrogant dismissal doesn't change that.
Not really; the 'Xserve' model is intended for full-service support with short turnaround, and the warranty is void if non-Apple RAM or drives are installed (but the warranty includes some nifty plus-es, is served onsite etc.).
If you need flexibility, getting a server-software-equipped desktop machine is the way to go; there's a bunch of drive drawers, and third-party addons are not discouraged by Apple. Warranty service, though, might be carry-in/days.
The original poster, however, doesn't need flexibility. He just needs to pay for the 750 MB drives and accept that he's getting more than bare minimum.
Re:video technologies confusing?
on
Plasma or LCD?
·
· Score: 1
There are lots of ways color info can be lost, like: card sends PAL and TV only receives NTSC, card has missing signal where the colorburst should be, R/G/B output buffers are shorted together.
What you probably have, though, is a card with analog RGB outputs, and you have connected ONE of the three colors to your TV. Check the info on your video card to see if it outputs COMPOSITE video. Because composite video is only capable of plain-old-analog (NTSC) performance, it might be that they didn't give your card that output.
Help is on the way; for digital cable boxes, a single type of output socket (Firewire) is required, here in the US. Use that, and there's no way to cross the wiring. Alas, the digital-rights-management crowd wants to define their own connectors and force incompatibility on the public...
>>A 44.1Khz sampling rate perfectly records a 22.05Khz signal
>NO IT DOESN'T.... Sampling without aliasing is not a perfect recording.
There's no musical signal allowed into the analog/digital converter at the 22.05 kHz frequency, because filtering must remove all/almost all of 22.06 kHz and the filters aren't abrupt.
And, if it was the case t hat you needed 22.05 kHz digitized, remember that you have (equally important) terms
A Sin(22.05 kHz *t) + B Cos(22.05 kHz *t) and a sampling at 0 and 180 degrees (i.e. two samples per cycle) will measure B and never measure any value for A.
The Nyquist limit isn't a 'it's perfect-here' point, it's a 'never-above-here' limit. You lose half the information at the Nyquist frequency in this example.
Audiophiles want to talk about the limit, but engineers just want the limit to be above the musical reproduction range. The engineers win every time.
>the eactor used in Chernobyl was designed by a fool. No sane person would use >a graphite moderated reactor today.
There is a bit of a misconception here: most reactors on the planet are graphite moderated and that WASN'T the salient feature that made Chernobyl pop. The Chernobyl design was over-moderated (which is sometimes referred to as positive-void-coefficient), so that a minor fault in the moderator would result in increased neutron flux.
The operation rules were well designed to prevent the initial event, and the operators ignored those rules for a test (and were duly prosecuted). When, after the test, shutdown was initiated, there was an isotope mix (due to the odd test conditions) that caused a minor bit of damage to the core. Which caused the explosion, because the minor damage got amplified due to the positive void coefficient into a major burst of neutrons and explosion/fire/radioactive dust was the result.
The major lesson of Chernobyl is that the operation rules are less immune to tampering that some kinds of built-in features.
The only reason graphite-core reactors aren't considered 'modern designs' is that they're all variations on the classic time-tested designs.
And the only reason 'nuclear is quite scary' is scaremongering. The safety of wood-fires through the years may be lots worse, but doesn't get the same hysterical press.
>1. If the drive is no longer recognized by the controller then the circuitry on the board has been damaged. Replace the board from an identical drive...
It may be that a board swap from a similar drive will restore the disk to usability (enough to use Norton Wipe Disk or other security-style erase program). Then you put the dead board back on the drive and get your warranty replacement.
I've done it.
But there's a good possibility that the problem is noisy read/write (either the heads or the amplifiers), and then it won't pass its little power-on-self-test. I've seen a lot of dead drives that seemed to spin, but couldn't control speed or find tracks, both of which are functions mediated by reading the magnetizations on the drive, and they always went into NOT READY state. Detailed disk status could read NO DISK, DISK NOT READY, or OFFLINE.
Lotsa disks with good motors and write heads don't let you try to erase, no matter what circuit board swaps you have available.
In the vein of fraud concerning titles being hard to undo, one must remember that the early days of the Credit Mobilier scandal included a fraudulent taking of the charter to build a transcontinental railway.
In the U.S.A., the fraudsters took the intercontinental railroad, not just a puny little single house.
I don't recall the full details, it concerned the principal being a southerner in the post-War Between the States era, and the later parts of the scandal included some congressmen with payoffs, but the courts never DID make any of the victims whole.
I've had some (physics lab) equipment operating at low temperatures, and my coworkers have done it down to VERY low temperatures. The problems are many and subtle.
First (as another poster mentioned) transistors (bipolar junction transistors) lose gain at low temperatures, while MOS adds gain. Either is sufficient to cause problems (the bipolar transistors lose drive capability and the MOSFETs oscillate at frequencies you didn't think they could reach). Second, there are little wire connections that were made at solder-melt temperatures and cooled to room temperature (with some builtin strain). The further cooling to low temperatures, with applied currents and possibly vibration, can cause mechanical stresses to go over the elastic limits. 'Storage Temperature" specifications on individual parts are usually going to tell you something of the range that's permissible, and typically those specs bottom at -65 to -85 C.
A few old chips have lots of temperature variants; Z80s were available in 0 to 70C (S grade) down to -55C to +125C (M grade), so some really simple computers can be had that won't mind the temperatures you have in mind. Old-fashioned is good enough, sometimes.
I had an odd failure in a liquid nitrogen dipstick processor (*) where the eventual resolution was that the CMOS circuit would work fine in the bore of the liquid nitrogen dewar, but its power supply kept blowing transistors. The CMOS conductance was extra high because of the low temperatures, and the clock-transient during switching was high current, enough to cause 'exploding wire' phenomenon in the pass transistor. Average currents were very low (maybe half a milliamp) but the little short transients had such a rise-time that they blew up internal wires (rated for 250 mA or more) by ultrasonic shockwave. The wires inside the CMOS chip were in an epoxy package and that kept them stationary so it was always the power transistor that failed, never the CMOS.
Temperature sensor ICs that have frequency output can be multiplexed with grid wiring and diode+sensor at each crosspoint. That'd mean only 20 wires into the cold box for 100 temperature sensors. You can run a few dozen wires into the cold box, surely? Diodes, and most temp sensor chips, are cold-tolerant.
Even if your chips will take the temperature, it's possible that a BIOS won't like it. If the CPU has a temperature sensor, is -14 degrees gonna make it panic? Unless your computers are spec'ed for the temperature, you shouldn't assume they'll even TRY to work.
Pulling a computer component out for maintenance will practically guarantee it gets condensation at first contact with warm moist air. Condensing moisture is not kind to connectors and switches, and does violate most warranties; you can work around this (put the cold component into a baggie with some dessicant, keep it shut for a few hours with dry air as it warms up) but it severely complicates the maintenance issue. And, the 'correct' procedure is unlikely to be the one the customer uses when you're not around.
(*) The long story: Carol wanted Chuck home in the evenings, but Chuck's experimental apparatus could only keep running if he refilled the cryogens at few-hours intervals on multiple-day experiment runs. So we cobbled up a capacitive sensor for nitrogen level using two coaxial stainless tubes, a 74HC14 Schmitt trigger to oscillate it, and frequency- triggered relay driver to allow compressed gas to push more liquid N2 into the reservoir. When it got to working, we retired the laboratory cot.
Yes, all that is true, but the wood can bind if it has internal flaws and splits during the cut, or if a knot comes loose, or if it just isn't completely straight. You can surface all four sides of your wood only for a few pieces before it turns the lovely woodwork hobby into a chore.
And the guard is removed for any dado cuts, and the splitter/riving knife has to be removed if you want to reverse cut (it's nearly obligatory when dealing with melamine coatings to make a reverse cut to nick the veneer, and makes for a better finish in lots of plywood situations too). There's so much reconfiguring required for safe table saw operation that most folk don't always do it. Some even get angry at the safety fittings (no, not me; I've had a couple of tutors who could hold up 9.5 fingers).
The capacitive-sensor with DSP processing sounds like a good idea, ESPECIALLY because it doesn't have to be bypassed for normal operations. Another doesn't-get-in-the-way item is to periodically degunk the saw blade with a mild lye solution. I like to follow up with a buffing wheel in a Dremel tool; you can get carbide teeth nice and clean with a little emery compound.
I was taken with the comment "scientists can be fussy"; we need a nap, I guess?
Seriously, guard bits and high-precision accumulators and addition algorithms that add the small numbers first (so the roundoff error in the big numbers doesn't unduly affect the result) are all good ways to treat roundoff errors. The old VAX/VMS math function libraries were available with single, double, H_floating, and G_floating just in case you needed many-digit exponents and really LONG mantissas.
What this article seems to suggest, though, is DECIMAL arithmetic (shades of COBOL!) as a solution. It is to laugh. Decimal arithmetic manipulation is familiar from childhood, but has no other virtue. Arrays of 'values' are containers of variable-length records, you gotta traverse a tree to find an element (or index tables, or worse, do a linear search); Time to complete an operation is dependent on the data values. There are about 70 functions on numbers in the FORTRAN math library; does anyone want to rewrite them for decimal arithmetic? What kind of precision 'improvement' do you expect in the case of functions like SIN? How would you even specify a precision goal for such things?
And few if any scientists need more precision than double (there were those guys in Hans Dehmelt's group, at U. of WA, that did a fourteen-digit measurement... but they would have got it right even if they had to use abacuses...)
A more interesting point is on the Fourier transform; there's a theorem (Parseval's theorem, if you must know) that says a bit of error (noise) added to a data set will become, after Fourier transforming that data set, the exact SAME amplitude of noise in the output data. That means that the Fourier transform doesn't reduce the signal/noise ratio of the input data. I often have wished that other mathematical and program procedures had the same property. They don't (and this has been known since Lagrange found the problem in orbit calculations). And it doesn't matter whether the roundoff is perfect, because ALL data starts with some noise. And it can never get better.
While it may be simple to say
>>The CIA is a "JOB". The president is your "BOSS". that ignores several important aspects of the current situation.
Firstly, the CIA is established by Congress, by the assembly of representatives of the people of the United States. Its charter defines the duties of the people who work there, and allows for LOTS of insulation against the whims of political appointees or other control by the President. Civil servants can't be fired for this kind of thing, by law. The oversight by the President and executive branch simply doesn't extend that deep.
That law was written after major abuses of Executive oversight of civil agencies in the 1800s. Call me a conservative, but I want that law STRICTLY enforced, it is definitely preferable to the strong-executive-fires-at-whim model.
But, the person involved was a contractor, not a civil servant. Thus, cancelling the contract WAS a legal option. It is not an option that has a benign odor, however. It stinks.
By law, the practice of torture is illegal; our Congress passed recent legislation to that effect, when the issue seemed unclear. I see no problem with civil servants discussing this, even taking positions on it. They're required to stay within the law,. and discussion can clear up confusion.
The specter of discussion causing loss of income is... disturbing.
From the root comment: "Except I only heard the commentary from Scott Cleland. It was chock full of misinformation and outright lies."
Verbum sap.
So, what DOES the violation of net neutrality do for those big access providers? I think they're trying to make a viable channel for video-on-demand. That means porn, and they'd rather call it 'let the market decide' because it sounds respectable (to Republicans, anyhow).
The servers we all use on the internet (Google, Ask, Ralph's-pretty-good-grocery.com...) are already paying according to (1) pipe size and (2) aggregate flow and the proposal is to add (3) continuous-service-no-busy-signals. In the water analogy, it's about your utility delivering nothing for a few seconds/days. And my water utility DOES interrupt my service from time to time when high-flow uses intrude (occasional street cleaning using the water mains, stirs up sediment and loses water pressure).
This probably means everyone on the current high-interest services can do just fine with the lowest of the 'priorities' that would exist under the proposed scheme. They hate the idea of taking a well-understood rate structure and making unspecified changes (and the doubletalk from the tiered-internet proponents isn't gonna allay those fears, nor should it- yes, there WILL be gouging tried if net neutrality goes by the wayside).
Probably they CAN'T do any digital operations. This professor's other work is in phased-array radar, so this 'chip' is a processor only in the sense that my Cuisinart is a processor. It isn't a digital computer.
At most, it has some digital enable/control inputs.
But, that doesn't mean it isn't a real and important accomplishment. Vacuum tubes worked for decades in radio before their qualities as digital building blocks were exploited! Umm... maybe that's a bad example.
For retroreflection inside the camera (which is the feature of CCD cameras they are detecting, and which defines the difference between cameras and eyeballs to the sensors in the article), there's a known technical solution.
Apply to the front of the lens a linear polarizer (reduces light by a factor of two; adjust your exposure accordingly) and a suitably oriented quarter wave plate. The pair is available as a 'circular polarizer', has been used to increase contrast of CRTs and other displays (does anyone remember Nixie tubes? this technology IS that old).
The circular polarized light, when reflected back, changes handedness (from right hand to left hand) and the back passage through the quarter wave plate turns it to plane polarized, at right angles to the linear polarizer. Nullification of the "cat's-eye" reflection effect is the result.
Some other posters have suggested SLR or other shutter-usually-closed cameras, but that just raises the possibility that your shutter opens between blinks of the film projector (and takes a pure black shot). Remember, the movies of film-on-screen variety present 24 frames per second, blinking each frame twice (because the 24 Hz flicker was annoying); there are 48 dark intervals each second between the light-on-screen instants. With a fast enough shutter speed, you'd definitely capture an occasional dark field, and a lot of half-bright ones.
To clarify things, here's several scenarios for erasure: "delete file" erasure: tell the OS that that part of a file system doesn't have any current ownership,
and that the filename doesn't exist, i. e. doesn't point to any data. "overwrite sectors" erasure: direct the hard disk drive to put new, noninformative, data into the
spaces formerly occupied by a file's data (and maybe metadata, like the file's icon and such) "multiple remagnetize" erasure: direct the hard disk drive to put all (in binary terms, both) physical
magnetizitions onto the data area, so that data's remnant traces are not informative "whole-disk multiple" erasure: ensure that all areas on the hard disk and all other data-holding parts (flash ROM)
are multiply rewritten. This would make the bad-block list disappear, might even make the
original format (how many tracks and sectors) unknowable to an investigator.
After "delete file", unerase software can bring much data to light by scanning the drive through the normal hardware. Because EVERYONE KNOWS THIS, there are 'secure erase' options in many disk tools (Norton "Wipe File", Mac OS X "Secure Empty Trash" etc.)
Those secure erase tools do multiple "write-over-sector", but there are some regulations that require "multiple remagnetize" erasure, and even 'dd/dev/random' isn't guaranteed there; you gotta pay money for a tool certified for that use. Here's why:
What everyone DOESN'T know, is that "write-over-sector" leaves behind some small regions (magnetic domains) in places the read/write heads cannot access, which can be sensed by exotic techniques (optical rotation, neutron scattering, electron beam microprobing). The erase-35-times and DOD (military) multiple-erase requirements are aimed at this kind of exotic stuff. Nothing you can do in software would get data back from "write-over-sector" erasure.
The modern disk drive compacts the data into a serial bit stream of known bandwidth and containing parity/error correcting code information, and DOES NOT put ones down on the disk when ones are in the data (MFM, RLL, and suchlike encoding schemes are in use on ALL media I'm aware of). This embedded-clock-and-data stream is hard to predict (what does Hitachi use on sATA drives this week? I don't know. Does anyone?), but WITH KNOWLEDGE of the encoding scheme, there are different recommended patterns for ensuring erasure to the standard of 'put ones on every spot, then zeros on every spot' . The use of software with ones in the DATA INPUT is not going to cause ones in the MAGNETIZED PATTERN, but you can come up with a set of data inputs that DOES effectively hit every bit of the surface. The famous paper on erasure has thirty-five scenarios for the encoding on the disk, and attempts to give a full remagnetize (with 'dd/dev/pattern01' through 'dd/dev/pattern35' kinds of operations).
So, that's a third kind of erase, intended to remagnetize all portions of the disk surface. The formal requirement to remagnetize the surface is ridiculously strict, becaue the exotic techniques DON'T KNOW HISTORY. Those random little domains can be left over from the manufacturer's bad-block scan, or from last December's diagnostic reformat, or from the camera run from last week, or from this week's most sensitive information, or can be a combination of all of those.
Or, it could be a bit of cosmic ray induced damage. The exotic reconstruction technique doesn't have any noise margin, it doesn't ignore the insignificant; noise is guaranteed.
The engineering isn't as direct an approach to show infeasibility as the physics. Let's consider the safety first...
Capacitors have energy stored in the electric field, and when that field gets too high, they spontaneously short out. One familiar example of this behavior is in a Geiger counter; when you hear a click, it means a cosmic ray caused an avalanche breakdown and the whole high voltage power supply just got dumped (shorted to ground, basically). You have to honor safety limits on the energy stored, or that 'click' will scale up to the whole-energy-to-run-the-car-a-hundred-miles kinda !C!L!I!C!K! that is best appreciated in bunker far from ground zero.
You can improve capacitors by minimizing the waste space (make the metal thinner, because the field in the insulator holds the energy). You can improve capacitors by maximizing the field in the insulator (raising the voltage rating for a given geometry). And you can improve capacitors by storing more energy in the field (using an insulator with a high dielectric constant).
This nanotube dingus mainly works by reducing the metal volume (small amount of metal foil with lots of area due to hairy bits). The internal field has lots of curvature, because of the pointy/highly curved nanotube surface, so field inhomogeneity is going to be high, and the maximum field limit (at the pointy parts) means the average fields are going to be low-ish.
Batteries, on the other hand, suck up the charge into atoms, one electron per atom, and make the field due to that charge vanish; you only get one voltage option (set by the chemistry; for alkaline cells it's 1.5V), but the amount of charge is HUGE compared to what a capacitor can suck up before it flashes and bangs.
My calculations: if you could safely sustain 10,000V per millimeter in the capacitor, with modest dielectric constant, and assuming the full field in 5 cm**3 of volume (about the same as an AA cell), your energy storage would come to 0.44 Watt-hour. A NiMH AA cell with 2000 mAh capacity has about 2.4 Watt-hour of stored energy.
The area of the electrode, so carefully maximized, only means that the charge is high and the voltage relatively low, it doesn't change AT ALL the total-stored-energy number. You have to diddle the other variables somehow, if you want to beat batteries.
What I've done (for 5000 plus titles) is to put together a Filemaker database, and forget about barcodes, just type in author and name... it helps that I'm an adequate typist. Ideally, there would be full card-catalog info, but I've only entered one other bit of info, the size (my paperback shelves and my hardback/trade paperback shelves are distinct, 'cuz you get better packing that way). For future consideration, it might be good to catalog short-story entities within the volumes... A friend does something similar and logs the prices (in case of insurance claim).
It's working fine, and won't break until my Macintosh hardware will no longer run the old version 2.1 of Filemaker.
To bookhunt, I've exported a tab-delimited file of the author and name fields, and use split (yeah, the old BSD utility) to make small files from it, which go onto an iPod in the Extras/Notes subdirectory. iPod notes browsing works with short files only. It's good enough to slow my rate of duplicate acquisition, but something on a PDA where you could jump to the middle (by text entry on author or title) would save time. I'm considering that, though I usually have my iPod and only occasionally carry the PDA.
For reference, consider that author/title info for 5000 books still is under 1 MB of data, so the big-data problems are unlikely to occur. Even an old low-capacity PDA will work, and any iPod except the shuffle.
I've seen pro photographers using this kind of drive, to good effect.
A full day's photoshoot would exceed the capacity of the digital camera's media card (remember, it's about $60/gigabyte for compactFlash, having dozens of cards is not a great option). But even the raw mode (uncompressed) will fit easily onto a suitable disk.
You need a good laptop screen to preview the shots anyhow (before the model changes her dress, you need to KNOW that there's a good image), so you plug in the outboard drive (get three, they're small), preview and when it looks good, send the decorative woman back to the dressing room while you do your download and sort files into labeled folders.
And the entire assembly is portable so you can go on the road for backdrops or to visit the clothing-prototype collection. If there were good eSATA laptops, it'd be a contender. But, there aren't. An Apple powerbook with firewire drives can get a LOT of use before the hardware gets obsolete. The use of a stack of spare batteries means you can be far from AC power during the whole event (summer fashions on the beach, you know).
While it isn't usually done, the Mac mini actually DOES have a second fast network port (for a total of four: modem, Airport/802.11g, Ethernet 100baset are the others). On the new Core minis, it's 1000baseT Ethernet and no modem...
Firewire is a capable interface, and IP traffic is one of its capabilities. It works with a Firewire cable, peer-to-peer, between two Macs and at 400 Mbit/sec it's sometimes an improvement on crossover-cable-Ethernet.
There is actually a reason for the Mac Mini to be thought of as a good media computer, which hasn't been touched on yet. The future holds a threshold date after which all analog TV transmissions become null and void, when digital receivers (set-top boxes) must be used to get any off-the-air broadcasts.
And all those set-top boxes are going to have Firewire ports. Most PCs are unsuited to the entertainment center because they lack the basic amenities (silence, remote control, low power consumption, firewire). The Mac Mini doesn't suck. It IS suited to this location, and it's available now. Heck, it was available last year, and the year before that...
Remember, too, that adding a set-top-box means you have to have multiple boxes to do familiar tasks (tape the news on channel N while watching the movie on channel M?) and that the task of tuning in a channel is no longer something your VCR can do, 'cuz it takes the signal from that set-top-box...
You'll want a sane interface, using a single remote control, reading a single menu from a screen, and having it all JUST WORK without any of the little gotchas...
Remember that VCR that was hard to tell whether it was AM or PM? Remember that VCR that wouldn't do its timed record unless it was in
OFF/standby mode? Remember the recording that hit the end of the tape (or DVD) and lost
the final scenes? Remember the power glitched, and nothing kept its settings (a computer
with filesystem and backup battery would have solved that problem)? Remember the cute accessory outlet on the cable box that turned off the TV,
and how the before-sleep ritual of turning the TV off meant the VCR was
taping from a turned-off cable box and it taped a lot of nothing? Remember how the universal remote got bumped to TV and you tuned the TV instead
of the cable box and nothing showed up right? Remember how the Beta and VHS recorders were daisy-chained and somehow the
signal was getting noisy until you unplugged some of them?
All those glitches are soluble but the solutions are in the form of integration of functions and good software modeling and display/control functions. Tivo does most of this, but not all (multiple-channels-at-once? Gonna cost ya!) The need for information integration and a control terminal (keys on the remote, menu and status on the big screen) with responsibility for the whole media center is here, NOW.
I think a lot of folk will have a computer of some sort next to the TV in the next few years. The happiest of those folk might have a Mac Mini.
"When I buy a piece of software,,,
If I want to put extra ones and zeros in it, while forfeiting any warranty, that's MY damn business. "
Alas, your point has a kind of validity, in that Apple would be hardly able to show any 'damage' done to them in a civil suit.
Note, however, the Apple communication referred to DMCA, which is a federal criminal statute and which DOESN'T require Apple to be there in court at all. Some prosecutor, who collects a salary at public expense, would be your opponent.
DMCA (Digital Millennium Copyright Act) is a piece of relatively obscure, not easily tested, law, which Apple's lawyers have LOTS of expertise in dealing with. It might be they are giving a friendly warning here, and not hinting at impending lawsuit at all.
Most open-software principles are violated by the sort of obscurity that DMCA is intended to protect. Most folk reading this site are probably saddened at this development, and maybe Apple is, too.
Slashdot Mirror
Yep, the mistakes of SCSI are being repeated. When Apple put SCSI on their boxes
in 1986, there WAS NO STANDARD for the external connector, and DB-25 (which
encouraged malfunctional printer-cable and serial-cable connections) and
'Centronics' (actual Micro Blue Ribbon-50) with the funky springlatch were
rampant for years. Neither was crimpable directly to ribbon cable (but the
Centronics cables grew an expensive variant that converted the incompatible
spacing, so lotsa folk didn't know of that problem).
Then a standard was issued for the external connector; it only crimped onto
fine-pitch ribbon cable (everyone had to switch over), and used a funny spring
latch.
But SCSI was used for complex systems that lived in wiring closets, you CAN'T
tolerate loose connections, and those spring latches were terrible. With SCSI-3
came a new external connector standard, with a wider bus AND screwdown
connector latches. Took three tries to get it right.
So here we see SATA with standard specs for internal connectors only. Folk
need external, so they cobble a shield around the internal (but the unshielded
cables still fit-- OOPS!). The standard gets an update that requires a shield
and enforces it with a new plug, but uses flimsy latches. Consumers
will like it for temporary plugins, but the serial-attached-SCSI types
are claiming they interplug with SATA, and I just know a third connector
will be issued for the high-end folk, and it will be a repeat of SCSI.
What a waste of money, brains, time (WOMBAT).
I once labeled all the different SCSI connectors I'd seen used. A, B, C.... got to
about S. There's about 50 pounds of cables in my basement, with various
ends, all in the boxes marked SCSI.
We're gonna see the same with SATA.
>Firewire? Why do enterprise desktop users need firewire? The only reason you need it is for digital video and >audio or extremely fast file transfers. Not desktop use.
That's not my experience at all! A dead or dying Macintosh can be rebooted into target disk
mode and all its data transferred to another machine via that Firewire port. Or you can
use it for TCP/IP at 400 mbits/second, or to attach a port-powered bootable hard disk.
None of those uses is insignificant, nor can USB be considered a suitable substitute.
Any system admin with a bunch of Macintoshes should buy a laptop-size Firewire-powered
emergency disk drive. Buy one, you'll never be sorry. Won't do much good on PCs, though;
gotta use the USB for them, bootability is problematic, power cord required, etc.
Anonymous Coward wrote:
>They had USB... you could just plugin a USB floppy drive if you needed it.
There was also a header connection, on the logic board, and with some sheetmetal
cutting it was possible to wire to it and run a Macintosh floppy drive.
The Macintosh floppy disks used 400k and 800k formats that required multispeed
spindle motors, and WEREN'T playable or recordable on 1.44M drives, but the
Macintosh mechanisms handled all three types. The USB plugin drives only
ever handled 1.44M. They wouldn't read the original Mac 400k or 800k format.
Imation's SuperDisk drive handled both 120MB cartridges and 1.44M floppies, but
that was the last gasp of compatible flexible disks for USB. Zip drives, still common,
are the only modern-ish floppy. I think they're up to 750 MB?
>It doesn't matter at all because the vast majority of business applications are not available for the mac.
No person, no business, needs 'the vast majority of' available software.
Macintosh computers work and are priced fairly. Some arrogant
dismissal doesn't change that.
> ... can buy drives elsewhere and put them in
Not really; the 'Xserve' model is intended for full-service support with
short turnaround, and the warranty is void if non-Apple RAM or drives
are installed (but the warranty includes some nifty plus-es, is served onsite etc.).
If you need flexibility, getting a server-software-equipped desktop machine
is the way to go; there's a bunch of drive drawers, and third-party addons
are not discouraged by Apple. Warranty service, though, might be carry-in/days.
The original poster, however, doesn't need flexibility. He just needs to
pay for the 750 MB drives and accept that he's getting more than bare
minimum.
There are lots of ways color info can be lost, like: card sends PAL and
TV only receives NTSC, card has missing signal where the colorburst should
be, R/G/B output buffers are shorted together.
What you probably have, though, is a card with analog RGB outputs, and
you have connected ONE of the three colors to your TV. Check the info
on your video card to see if it outputs COMPOSITE video. Because composite
video is only capable of plain-old-analog (NTSC) performance, it might
be that they didn't give your card that output.
Help is on the way; for digital cable boxes, a single type of output socket (Firewire)
is required, here in the US. Use that, and there's no way to cross the wiring.
Alas, the digital-rights-management crowd wants to define their own
connectors and force incompatibility on the public...
>>A 44.1Khz sampling rate perfectly records a 22.05Khz signal
... Sampling without aliasing is not a perfect recording.
>NO IT DOESN'T.
There's no musical signal allowed into the analog/digital converter at the
22.05 kHz frequency, because filtering must remove all/almost all of 22.06
kHz and the filters aren't abrupt.
And, if it was the case t hat you needed 22.05 kHz digitized, remember
that you have (equally important) terms
A Sin(22.05 kHz *t) + B Cos(22.05 kHz *t)
and a sampling at 0 and 180 degrees (i.e. two samples per cycle) will
measure B and never measure any value for A.
The Nyquist limit isn't a 'it's perfect-here' point, it's a 'never-above-here' limit.
You lose half the information at the Nyquist frequency in this example.
Audiophiles want to talk about the limit, but engineers just want the limit
to be above the musical reproduction range. The engineers win every time.
I'm a mathematician, I just cringe. A lot.
>the eactor used in Chernobyl was designed by a fool. No sane person would use
>a graphite moderated reactor today.
There is a bit of a misconception here: most reactors on the planet are graphite moderated
and that WASN'T the salient feature that made Chernobyl pop. The Chernobyl design
was over-moderated (which is sometimes referred to as positive-void-coefficient),
so that a minor fault in the moderator would result in increased neutron flux.
The operation rules were well designed to prevent the initial event, and the operators
ignored those rules for a test (and were duly prosecuted). When, after the test, shutdown
was initiated, there was an isotope mix (due to the odd test conditions) that caused
a minor bit of damage to the core. Which caused the explosion, because the
minor damage got amplified due to the positive void coefficient into a major
burst of neutrons and explosion/fire/radioactive dust was the result.
The major lesson of Chernobyl is that the operation rules are less immune to
tampering that some kinds of built-in features.
The only reason graphite-core reactors aren't considered 'modern designs' is
that they're all variations on the classic time-tested designs.
And the only reason 'nuclear is quite scary' is scaremongering. The safety of
wood-fires through the years may be lots worse, but doesn't get the same
hysterical press.
>1. If the drive is no longer recognized by the controller then the circuitry on the board has been damaged. Replace the board from an identical drive ...
It may be that a board swap from a similar drive will restore the disk to usability (enough
to use Norton Wipe Disk or other security-style erase program). Then you
put the dead board back on the drive and get your warranty replacement.
I've done it.
But there's a good possibility that the problem is noisy read/write (either the heads
or the amplifiers), and then it won't pass its little power-on-self-test. I've seen a lot
of dead drives that seemed to spin, but couldn't control speed or find tracks, both of
which are functions mediated by reading the magnetizations on the drive,
and they always went into NOT READY state. Detailed disk status could
read NO DISK, DISK NOT READY, or OFFLINE.
Lotsa disks with good motors and write heads don't let you try to erase, no matter
what circuit board swaps you have available.
In the vein of fraud concerning titles being hard to undo, one must remember
that the early days of the Credit Mobilier scandal included a fraudulent
taking of the charter to build a transcontinental railway.
In the U.S.A., the fraudsters took the intercontinental railroad, not just
a puny little single house.
I don't recall the full details, it concerned the principal being a southerner
in the post-War Between the States era, and the later parts of the scandal
included some congressmen with payoffs, but the courts never DID
make any of the victims whole.
I've had some (physics lab) equipment operating at low temperatures, and
my coworkers have done it down to VERY low temperatures. The
problems are many and subtle.
First (as another poster mentioned) transistors (bipolar junction transistors) lose gain
at low temperatures, while MOS adds gain. Either is sufficient to cause problems
(the bipolar transistors lose drive capability and the MOSFETs oscillate at frequencies
you didn't think they could reach). Second, there are little wire connections that were
made at solder-melt temperatures and cooled to room temperature (with some
builtin strain). The further cooling to low temperatures, with applied currents and
possibly vibration, can cause mechanical stresses to go over the elastic limits.
'Storage Temperature" specifications on individual parts are usually going to tell you
something of the range that's permissible, and typically those specs bottom at -65 to -85 C.
A few old chips have lots of temperature variants; Z80s were available in 0 to 70C (S grade)
down to -55C to +125C (M grade), so some really simple computers can be had
that won't mind the temperatures you have in mind. Old-fashioned is good enough,
sometimes.
I had an odd failure in a liquid nitrogen dipstick processor (*) where the eventual resolution
was that the CMOS circuit would work fine in the bore of the liquid nitrogen dewar, but
its power supply kept blowing transistors. The CMOS conductance was extra high
because of the low temperatures, and the clock-transient during switching was high current,
enough to cause 'exploding wire' phenomenon in the pass transistor. Average currents
were very low (maybe half a milliamp) but the little short transients had such a rise-time
that they blew up internal wires (rated for 250 mA or more) by ultrasonic shockwave.
The wires inside the CMOS chip were in an epoxy package and that kept them stationary
so it was always the power transistor that failed, never the CMOS.
Temperature sensor ICs that have frequency output can be multiplexed with
grid wiring and diode+sensor at each crosspoint. That'd mean only 20 wires into the
cold box for 100 temperature sensors. You can run a few dozen wires into the cold
box, surely? Diodes, and most temp sensor chips, are cold-tolerant.
Even if your chips will take the temperature, it's possible that a BIOS won't like it. If the CPU has a temperature sensor, is -14 degrees gonna make it panic? Unless your computers are spec'ed for the temperature, you shouldn't assume they'll even TRY to work.
Pulling a computer component out for maintenance will practically guarantee
it gets condensation at first contact with warm moist air. Condensing moisture is not
kind to connectors and switches, and does violate most warranties; you can work around
this (put the cold component into a baggie with some dessicant, keep it shut for
a few hours with dry air as it warms up) but it severely complicates the maintenance
issue. And, the 'correct' procedure is unlikely to be the one the customer uses when
you're not around.
(*) The long story: Carol wanted Chuck home in the evenings, but Chuck's
experimental apparatus could only keep running if he refilled the cryogens at few-hours intervals on
multiple-day experiment runs. So we cobbled up a capacitive sensor for nitrogen level
using two coaxial stainless tubes, a 74HC14 Schmitt trigger to oscillate it, and frequency-
triggered relay driver to allow compressed gas to push more liquid N2 into the reservoir.
When it got to working, we retired the laboratory cot.
Yes, all that is true, but the wood can bind if it has internal flaws and splits during the
cut, or if a knot comes loose, or if it just isn't completely straight. You can surface all four sides
of your wood only for a few pieces before it turns the lovely woodwork hobby into a chore.
And the guard is removed for any dado cuts, and the splitter/riving knife has to be
removed if you want to reverse cut (it's nearly obligatory when dealing with melamine
coatings to make a reverse cut to nick the veneer, and makes for a better finish in lots of
plywood situations too). There's so much reconfiguring required for safe table saw operation
that most folk don't always do it. Some even get angry at the safety fittings (no, not me; I've had
a couple of tutors who could hold up 9.5 fingers).
The capacitive-sensor with DSP processing sounds like a good idea, ESPECIALLY because it
doesn't have to be bypassed for normal operations. Another doesn't-get-in-the-way item is
to periodically degunk the saw blade with a mild lye solution. I like to follow up with a buffing wheel
in a Dremel tool; you can get carbide teeth nice and clean with a little emery compound.
I was taken with the comment "scientists can be fussy"; we need a
nap, I guess?
Seriously, guard bits and high-precision accumulators and addition
algorithms that add the small numbers first (so the roundoff error in the
big numbers doesn't unduly affect the result) are all good ways to treat
roundoff errors. The old VAX/VMS math function libraries were available
with single, double, H_floating, and G_floating just in case you needed
many-digit exponents and really LONG mantissas.
What this article seems to suggest, though, is DECIMAL arithmetic (shades of
COBOL!) as a solution. It is to laugh. Decimal arithmetic manipulation is
familiar from childhood, but has no other virtue. Arrays of 'values' are
containers of variable-length records, you gotta traverse a tree to
find an element (or index tables, or worse, do a linear search); Time
to complete an operation is dependent on the data values. There are
about 70 functions on numbers in the FORTRAN math library; does anyone
want to rewrite them for decimal arithmetic? What kind of precision
'improvement' do you expect in the case of functions like SIN? How
would you even specify a precision goal for such things?
And few if any scientists need more precision than double (there were
those guys in Hans Dehmelt's group, at U. of WA, that did a fourteen-digit
measurement... but they would have got it right even if they had to
use abacuses...)
A more interesting point is on the Fourier transform; there's a theorem
(Parseval's theorem, if you must know) that says a bit of error (noise) added
to a data set will become, after Fourier transforming that data set, the exact
SAME amplitude of noise in the output data. That means that the Fourier
transform doesn't reduce the signal/noise ratio of the input data. I often
have wished that other mathematical and program procedures had the
same property. They don't (and this has been known since Lagrange found
the problem in orbit calculations). And it doesn't matter whether the roundoff
is perfect, because ALL data starts with some noise. And it can never
get better.
While it may be simple to say
>>The CIA is a "JOB". The president is your "BOSS".
that ignores several important aspects of the current situation.
Firstly, the CIA is established by Congress, by the assembly of
representatives of the people of the United States. Its charter
defines the duties of the people who work there, and allows
for LOTS of insulation against the whims of political appointees
or other control by the President. Civil servants can't be fired
for this kind of thing, by law. The oversight by the President
and executive branch simply doesn't extend that deep.
That law was written after major abuses of Executive oversight
of civil agencies in the 1800s. Call me a conservative, but
I want that law STRICTLY enforced, it is definitely preferable
to the strong-executive-fires-at-whim model.
But, the person involved was a contractor, not a civil servant.
Thus, cancelling the contract WAS a legal option. It is
not an option that has a benign odor, however. It stinks.
By law, the practice of torture is illegal; our Congress passed
recent legislation to that effect, when the issue seemed unclear.
I see no problem with civil servants discussing this, even
taking positions on it. They're required to stay within the law,.
and discussion can clear up confusion.
The specter of discussion causing loss of income is... disturbing.
From the root comment:
"Except I only heard the commentary from Scott Cleland. It was chock full of misinformation and outright lies."
Verbum sap.
So, what DOES the violation of net neutrality do for those big access providers?
I think they're trying to make a viable channel for video-on-demand. That means porn,
and they'd rather call it 'let the market decide' because it sounds respectable
(to Republicans, anyhow).
The servers we all use on the internet (Google, Ask, Ralph's-pretty-good-grocery.com...) are
already paying according to (1) pipe size and (2) aggregate flow and the proposal is to
add (3) continuous-service-no-busy-signals. In the water analogy, it's about your
utility delivering nothing for a few seconds/days. And my water utility DOES interrupt my
service from time to time when high-flow uses intrude (occasional street cleaning using
the water mains, stirs up sediment and loses water pressure).
This probably means everyone on the current high-interest services can do just fine with
the lowest of the 'priorities' that would exist under the proposed scheme. They hate the idea
of taking a well-understood rate structure and making unspecified changes (and the
doubletalk from the tiered-internet proponents isn't gonna allay those fears, nor
should it- yes, there WILL be gouging tried if net neutrality goes by the wayside).
Probably they CAN'T do any digital operations. This professor's other work is in phased-array
radar, so this 'chip' is a processor only in the sense that my Cuisinart is a processor. It isn't
a digital computer.
At most, it has some digital enable/control inputs.
But, that doesn't mean it isn't a real and important accomplishment. Vacuum tubes worked for
decades in radio before their qualities as digital building blocks were exploited!
Umm... maybe that's a bad example.
For retroreflection inside the camera (which is the feature of CCD cameras they are
detecting, and which defines the difference between cameras and eyeballs to the sensors
in the article), there's a known technical solution.
Apply to the front of the lens a linear polarizer (reduces light by a factor of two; adjust
your exposure accordingly) and a suitably oriented quarter wave plate. The pair is available
as a 'circular polarizer', has been used to increase contrast of CRTs and other displays
(does anyone remember Nixie tubes? this technology IS that old).
The circular polarized light, when reflected back, changes handedness (from right
hand to left hand) and the back passage through the quarter wave plate turns it to plane
polarized, at right angles to the linear polarizer. Nullification of the "cat's-eye" reflection
effect is the result.
Some other posters have suggested SLR or other shutter-usually-closed cameras, but that
just raises the possibility that your shutter opens between blinks of the film projector (and takes
a pure black shot). Remember, the movies of film-on-screen variety present 24 frames per second,
blinking each frame twice (because the 24 Hz flicker was annoying); there are 48 dark intervals
each second between the light-on-screen instants. With a fast enough shutter speed, you'd
definitely capture an occasional dark field, and a lot of half-bright ones.
To clarify things, here's several scenarios for erasure:
/dev/random' isn't
/dev/pattern01' through 'dd /dev/pattern35'
"delete file" erasure: tell the OS that that part of a file system doesn't have any current ownership,
and that the filename doesn't exist, i. e. doesn't point to any data.
"overwrite sectors" erasure: direct the hard disk drive to put new, noninformative, data into the
spaces formerly occupied by a file's data (and maybe metadata, like the file's icon and such)
"multiple remagnetize" erasure: direct the hard disk drive to put all (in binary terms, both) physical
magnetizitions onto the data area, so that data's remnant traces are not informative
"whole-disk multiple" erasure: ensure that all areas on the hard disk and all other data-holding parts (flash ROM)
are multiply rewritten. This would make the bad-block list disappear, might even make the
original format (how many tracks and sectors) unknowable to an investigator.
After "delete file", unerase software can bring much data to light
by scanning the drive through the normal hardware. Because EVERYONE KNOWS THIS, there
are 'secure erase' options in many disk tools (Norton "Wipe File", Mac OS X "Secure Empty Trash" etc.)
Those secure erase tools do multiple "write-over-sector", but there are some
regulations that require "multiple remagnetize" erasure, and even 'dd
guaranteed there; you gotta pay money for a tool certified for that use. Here's why:
What everyone DOESN'T know, is that "write-over-sector" leaves behind some small regions
(magnetic domains) in places the read/write heads cannot access, which can be sensed by
exotic techniques (optical rotation, neutron scattering, electron beam microprobing). The
erase-35-times and DOD (military) multiple-erase requirements are aimed at this kind of
exotic stuff. Nothing you can do in software would get data back from "write-over-sector"
erasure.
The modern disk drive compacts the data into a serial bit stream of known bandwidth and
containing parity/error correcting code information, and DOES NOT put ones down on the
disk when ones are in the data (MFM, RLL, and suchlike encoding schemes are in use on ALL
media I'm aware of). This embedded-clock-and-data stream is hard to predict (what does
Hitachi use on sATA drives this week? I don't know. Does anyone?), but WITH KNOWLEDGE
of the encoding scheme, there are different recommended patterns for ensuring
erasure to the standard of 'put ones on every spot, then zeros on every spot' . The use of
software with ones in the DATA INPUT is not going to cause ones in the MAGNETIZED PATTERN,
but you can come up with a set of data inputs that DOES effectively hit every bit of the surface.
The famous paper on erasure has thirty-five scenarios for the encoding on the disk,
and attempts to give a full remagnetize (with 'dd
kinds of operations).
So, that's a third kind of erase, intended to remagnetize all portions of the disk surface.
The formal requirement to remagnetize the surface is ridiculously strict, becaue the exotic techniques
DON'T KNOW HISTORY. Those random little domains can be left over from the manufacturer's
bad-block scan, or from last December's diagnostic reformat, or from the camera run from last
week, or from this week's most sensitive information, or can be a combination of all of those.
Or, it could be a bit of cosmic ray induced damage. The exotic reconstruction technique
doesn't have any noise margin, it doesn't ignore the insignificant; noise is guaranteed.
As I recall, This Island Earth by Raymond F. Jones was the first
mention of a batacitor. That was in 1952, when Farmer was a promising young author...
The engineering isn't as direct an approach to show infeasibility
as the physics. Let's consider the safety first...
Capacitors have energy stored in the electric field, and when that
field gets too high, they spontaneously short out. One familiar example
of this behavior is in a Geiger counter; when you hear a click, it means
a cosmic ray caused an avalanche breakdown and the whole high
voltage power supply just got dumped (shorted to ground, basically).
You have to honor safety limits on the energy stored, or that 'click'
will scale up to the whole-energy-to-run-the-car-a-hundred-miles
kinda !C!L!I!C!K! that is best appreciated in bunker far from ground zero.
You can improve capacitors by minimizing the waste space (make
the metal thinner, because the field in the insulator holds the energy).
You can improve capacitors by maximizing the field in the insulator (raising
the voltage rating for a given geometry). And you can improve capacitors
by storing more energy in the field (using an insulator with a high
dielectric constant).
This nanotube dingus mainly works by reducing the metal volume
(small amount of metal foil with lots of area due to hairy bits). The internal
field has lots of curvature, because of the pointy/highly curved nanotube surface, so field
inhomogeneity is going to be high, and the maximum field limit (at the pointy parts)
means the average fields are going to be low-ish.
Batteries, on the other hand, suck up the charge into atoms, one
electron per atom, and make the field due to that charge vanish; you only get
one voltage option (set by the chemistry; for alkaline cells it's 1.5V), but
the amount of charge is HUGE compared to what a capacitor can suck up
before it flashes and bangs.
My calculations: if you could safely sustain 10,000V per millimeter
in the capacitor, with modest dielectric constant, and assuming the
full field in 5 cm**3 of volume (about the same as an AA cell), your energy
storage would come to 0.44 Watt-hour. A NiMH AA cell with 2000 mAh
capacity has about 2.4 Watt-hour of stored energy.
The area of the electrode, so carefully maximized, only means that the
charge is high and the voltage relatively low, it doesn't change AT ALL the
total-stored-energy number. You have to diddle the other variables
somehow, if you want to beat batteries.
What I've done (for 5000 plus titles) is to put together a Filemaker database,
and forget about barcodes, just type in author and name... it helps that I'm an
adequate typist. Ideally, there would be full card-catalog info, but I've only entered one
other bit of info, the size (my paperback shelves and my hardback/trade paperback
shelves are distinct, 'cuz you get better packing that way). For future consideration,
it might be good to catalog short-story entities within the volumes...
A friend does something similar and logs the prices (in case of insurance claim).
It's working fine, and won't break until my Macintosh hardware will no longer
run the old version 2.1 of Filemaker.
To bookhunt, I've exported a tab-delimited file of the author and name fields, and
use split (yeah, the old BSD utility) to make small files from it, which go onto
an iPod in the Extras/Notes subdirectory. iPod notes browsing works with short
files only. It's good enough to slow my rate of duplicate acquisition, but something
on a PDA where you could jump to the middle (by text entry on author or title) would
save time. I'm considering that, though I usually have my iPod and only occasionally
carry the PDA.
For reference, consider that author/title info for 5000 books still is under 1 MB of data,
so the big-data problems are unlikely to occur. Even an old low-capacity PDA will work,
and any iPod except the shuffle.
I've seen pro photographers using this kind of drive, to good effect.
A full day's photoshoot would exceed the capacity of the digital camera's media
card (remember, it's about $60/gigabyte for compactFlash, having dozens of
cards is not a great option). But even the raw mode (uncompressed) will fit easily
onto a suitable disk.
You need a good laptop screen to preview the shots anyhow (before the model
changes her dress, you need to KNOW that there's a good image), so you
plug in the outboard drive (get three, they're small), preview and when
it looks good, send the decorative woman back to the dressing room while you
do your download and sort files into labeled folders.
And the entire assembly is portable so you can go on the road for backdrops or
to visit the clothing-prototype collection. If there were good eSATA laptops,
it'd be a contender. But, there aren't. An Apple powerbook with firewire
drives can get a LOT of use before the hardware gets obsolete. The use of
a stack of spare batteries means you can be far from AC power during the
whole event (summer fashions on the beach, you know).
While it isn't usually done, the Mac mini actually DOES have a second fast network
port (for a total of four: modem, Airport/802.11g, Ethernet 100baset are the others).
On the new Core minis, it's 1000baseT Ethernet and no modem...
Firewire is a capable interface, and IP traffic is one of its capabilities.
It works with a Firewire cable, peer-to-peer, between two Macs and
at 400 Mbit/sec it's sometimes an improvement on crossover-cable-Ethernet.
There is actually a reason for the Mac Mini to be thought of as a good
media computer, which hasn't been touched on yet. The future holds
a threshold date after which all analog TV transmissions become null
and void, when digital receivers (set-top boxes) must be used to get
any off-the-air broadcasts.
And all those set-top boxes are going to have Firewire ports.
Most PCs are unsuited to the entertainment center because they lack the
basic amenities (silence, remote control, low power consumption, firewire).
The Mac Mini doesn't suck. It IS suited to this location, and it's available
now. Heck, it was available last year, and the year before that...
Remember, too, that adding a set-top-box means you have to have multiple boxes to
do familiar tasks (tape the news on channel N while watching the movie on channel M?)
and that the task of tuning in a channel is no longer something your VCR can do,
'cuz it takes the signal from that set-top-box...
You'll want a sane interface, using a single remote control,
reading a single menu from a screen,
and having it all JUST WORK without any of the little gotchas...
Remember that VCR that was hard to tell whether it was AM or PM?
Remember that VCR that wouldn't do its timed record unless it was in
OFF/standby mode?
Remember the recording that hit the end of the tape (or DVD) and lost
the final scenes?
Remember the power glitched, and nothing kept its settings (a computer
with filesystem and backup battery would have solved that problem)?
Remember the cute accessory outlet on the cable box that turned off the TV,
and how the before-sleep ritual of turning the TV off meant the VCR was
taping from a turned-off cable box and it taped a lot of nothing?
Remember how the universal remote got bumped to TV and you tuned the TV instead
of the cable box and nothing showed up right?
Remember how the Beta and VHS recorders were daisy-chained and somehow the
signal was getting noisy until you unplugged some of them?
All those glitches are soluble but the solutions are in the form of integration
of functions and good software modeling and display/control functions.
Tivo does most of this, but not all (multiple-channels-at-once? Gonna cost ya!)
The need for information integration and a control terminal (keys on the remote,
menu and status on the big screen) with responsibility for the
whole media center is here, NOW.
I think a lot of folk will have a computer of some sort next to the TV in the next few years.
The happiest of those folk might have a Mac Mini.
"When I buy a piece of software ,,,
If I want to put extra ones and zeros in it, while forfeiting any warranty, that's MY damn business. "
Alas, your point has a kind of validity, in that Apple would be hardly able to show
any 'damage' done to them in a civil suit.
Note, however, the Apple communication referred to DMCA, which is a federal criminal
statute and which DOESN'T require Apple to be there in court at all. Some prosecutor,
who collects a salary at public expense, would be your opponent.
DMCA (Digital Millennium Copyright Act) is a piece of relatively obscure, not easily
tested, law, which Apple's lawyers have LOTS of expertise in dealing with.
It might be they are giving a friendly warning here, and not hinting at impending
lawsuit at all.
Most open-software principles are violated by the sort of obscurity that DMCA is
intended to protect. Most folk reading this site are probably saddened at this
development, and maybe Apple is, too.