Matrox are still the only people (of which I am aware) shipping a quad-head DVI video card. And it's passively cooled, to boot. It's the video card Bloomberg sends you with a new terminal. ATI and Nvidia would kill to have a reliable price-insensitve market like that, but they just can't make the quality goods like Matrox can. The other two are stuck in the l33t overclocker eXtreme universe.
Actually the mobile police data system barely works at all. Hence the need for something to replace it. As far as the fire department, I'm with you. The trucks should keep their radios.
Ah, so what you are saying is the Free Market (TM)(R)(C 2005) that exists in wireless phone service has failed to provide you with good service? I'm shocked. Surely the slashbot libertarian parrots will be unable to appreciate the nuance of your situation.
Well, let's see: the police department, the fire department, the DPT, the DPW, MUNI, the libraries, and the schools. All these can cancel their existing expensive proprietary licensed radio other telecommunications service in favor of cheap commericial off-the-shelf technology.
RAID 6 can survive multiple disk failures (two, exactly), as can RAID 1+0, RAID 1+5, and even RAID 1. All of these are "sane" and in fact widely used. RAID5 is the only level that can't survive multiple failures.
Here's why I think POD is great. Besides lots of generated formats like HTML, POD works just like man: it shows *everything* about the package on one page, where I can read it. And since Perl programmers actually use and enjoy POD, I know there will be sufficient documentation to go from zilch to productivity.
Take, as an example, the DBI perldoc. It starts with succinct method documentation in the form of examples, in the order they might be used in an actual program. Then it moves on to useful newbie information, and a long, complete-sentence description of the way DBI works. And, since the page is in a linear format, I can find the specific documentation I need with one tap of my little finger on the / key.
Now let's look at a Javadoc from the Blackberry J2ME software. The front page has all the class names on the left, with terse descriptions in tables on the main page. There is no main description which tells me what the software is, nor how to use it. The packages are described like "The UI API provides a set of features for implementation of user interfaces for MIDP applications", however there is no definition of MIDP here, nor indeed is there any such definition anywhere in the Javadoc tree. Worse, since the javadoc is in a huge tree of separate HTML pages, I can't just quicksearch for the bit I'm looking for. I have to switch to the terminal and use grep -r, or try my luck with Google.
Here's the description for Class MessageFormat:
Formats and parses messages in a locale-sensitive manner.
Locale-specific formatting is passed on to other formatter objects that inserted into this formatter's pattern (this class supports the string, date, and time element formats).
That's swell. WTF is a formatter object? A pattern? Maybe the superclass will tell us:
Defines fundamental functionality for classes that format locale-sensitive information.
Examples of information that might be locale-sensitive are dates, system messages, and numbers.
Huh, no help there either. I don't know what it is about POD or Javadoc, but most of the time 'perldoc Whatever::Package' gives you some seriously useful documentation, and in my experience nearly any given Javadoc is totally useless to the beginner.
The following is an example of useless documentation:
frobWoggleFfloofMoing
public void frobWoggleFfloofMoing(String, String, String)
Frobs the Woggle
...
You see, running Doxygen over your header files may produce some output in HTML format, but it doesn't produce what I like to call "documentation." For instance, documentation would explain what is a Woggle, and when should it be Frobbed?
You weep for slashdot? Take your order-5 userid and cram it up your rectum. You cannot construct an argument by cutting and pasting talking points from freerepublic.com.
Yes, and if you weren't an ignorant twat you would realize that modern refrigerators have coefficients of performance in excess of unity. Well in excess, normally.
What are the differences between Windows, Unix, and mainframe programmers? What do we all need to know to get along in each other's worlds?
The difference is one programs Windows, one Unix, and one mainframes. As a fifth-year geek, you should take the rantings of Joel, ESR, and any other pointless windbag and send them to the bit bucket.
I can't seem to scrape together a lot of pity for Cendant here, either. Cendant's patent is for "System and Method for Providing Recommendation of Goods or Services Based on Recorded Purchasing History." Doesn't anyone remember FireFly circa 1996 and subsequently bought and murdered by Microsoft? The entire point of FireFly was to recommend stuff based on how well you liked or disliked records. In fact, it's possible that Cendant's technology shares a common ancestor with FireFly's: Chris Bergh, a technologist at FireFly and NetMarket.
As far as I am concerned, Cendant drew its sword and now they cannot avoid battle. Tough shit for them.
Yeah, nice CPU, difficult for software authors. I read a paper recently wherein the authors managed to reduce L4 microkernel message passing (up to 8 bytes) to 36 clock cycles, which is far faster than any other platform. But this was done by hand, and the compiler blurted out a routine that required 508 cycles. The gulf between what you can really do with an Itanium, and what normal software writers can do with it, remains huge.
I don't even give a damn about the multichannel stuff. I'd like to see more releases of high-resolution stereo recordings from analog tape masters. Of course, we'll never see anything Sony owns on DVD-A.
I think a real problem for DVD-A is the previous two decades of direct-to-digital production. If a recording was made to 16/44.1 stereo masters, there's absolutely no point in releasing it on high definition multichannel media. The best you can hope for is to sound exactly like the CD, and what's more likely to happen is a corny surround remaster where the drum kit is inexplicably behind the listener. So there's this huge catalog of (mostly pop) music from the 80s and 90s made to low-definition digital formats, and they can never be improved by DVD-A nor SACD.
Don't blame it on the people, blame it on the publishers. Your collection must consist of almost every DVD-A ever released. The DVD-A catalog just isn't that appealing to most people, and then there's the expensive and/or crippled players that lack full-resolution digital outputs. There's also consumer confusion regarding SACD vs. DVD-A. A format war, in other words.
If the situation were sane, people would buy more DVD-Audio.
I think you are on crack. There's no p-layer in a Schottky barrier and therefore no (or little) stored charge when conducting. Schottky rectifiers typically have reverse conduction of about 100% forward current during recovery, and settle to DC reverse current within 50ns and in a single damped excursion. I've never even heard of anyone using snubbers with Schottky diodes in applications up to 1GHz.
Two things. First, most low-power audio equipment are still using linear supplies, not switchers. Secondly, I wasn't implying that the regulation needs to work into the high frequencies, I was stating that the power supply as a whole needs to offer low impedance across a large band. Which, as you point out, means the bypass needs to be up to the task.
Most equipment uses both bad regulation and bad bypass, so the point is really moot. Replacing the capacitors in or modifying the design of the bypass can yield a substantial improvement, especially for DACs. Replacing bad linear regulators with better designs can give you good regulation at 10kHz where previously the case was hopeless.
Probably nobody uses DVD-Audio. The effort spent cracking the copy protection will be proportional to the popularity of the medium. When the DVD Audio section at your local record store is as large as the CD section, we may see some more time being spent on such things.
I think you are right. Avaya offers VoIP equipment. It has the same unbelievably high price and astonishing lack of quality as Avaya's other equipment (attention Intuity engineers: I'm looking at you). Meanwhile Cisco offers the same or better quality at low prices and without the bad service and clueless support.
It's inevitable that the netheads will bury the bellheads. The only question is: when? 2007? 2020?
Sure. I hit the "Forward" button on my Lucent digital set, punch in my phone number, and any calls to my desk ring my cell or home phones. No flaky VPN nor softphone required. The PSTN is many things, but featureless is not one of them.
By the way, "PBX" and "VoIP" are orthogonal, not exclusive. You can have a VoIP-connected PBX just the same as you can have a T1 or POTS or even a cellular-connected PBX. The PBX is just a telephony device that doesn't care what kind of line card you use to make and receive calls.
The problem with Vonage is they don't control the quality of service on the downstream. The Vonage adapter can prioritize packets on the upstream, but your DSL provider, who has no agreement with Vonage and doesn't want one, ignores any QoS request on the packets returning to your phone.
With business DSL services you often have an entire dedicated line (such as a T1) to your VoIP provider's head end, which allows prioritization of packets in both directions. The branch office where I work has VoIP on a dedicated T1 to Speakeasy and, after some initial problems with the connection, we are now getting better voice quality and better service for about a twentieth what we used to pay Avaya and WorldCom for the same number of lines.
Yes! A high-quality clock can be recovered by buffering the input and synching the clock on the DAC to the very-long-term average clock rate recovered from the S/PDIF. The only problem here is multirate, where your PLL might not pick up the change quickly enough, and as you mentioned the video sync problem.
Another tactic is to employ a sample rate converter, which can isolate two clock domains but unfortunately introduces some distortion. The distortion can be pretty significant for ratios very close to 1. Sample rate conversion is nearly universal in consumer equipment with a digital input. Unfortunately the clocks on both sides of the SRC are usually crap. There's plenty of opportunity to improve consumer equipment by modifying or replacing the clock or the clock's power supply.
Slashdot Mirror
Matrox are still the only people (of which I am aware) shipping a quad-head DVI video card. And it's passively cooled, to boot. It's the video card Bloomberg sends you with a new terminal. ATI and Nvidia would kill to have a reliable price-insensitve market like that, but they just can't make the quality goods like Matrox can. The other two are stuck in the l33t overclocker eXtreme universe.
At first I laughed, but that's damn useful. It's passively cooled and uses a slot that is on my motherboard but currently useless. Good stuff.
Actually the mobile police data system barely works at all. Hence the need for something to replace it. As far as the fire department, I'm with you. The trucks should keep their radios.
Ah, so what you are saying is the Free Market (TM)(R)(C 2005) that exists in wireless phone service has failed to provide you with good service? I'm shocked. Surely the slashbot libertarian parrots will be unable to appreciate the nuance of your situation.
Well, let's see: the police department, the fire department, the DPT, the DPW, MUNI, the libraries, and the schools. All these can cancel their existing expensive proprietary licensed radio other telecommunications service in favor of cheap commericial off-the-shelf technology.
RAID 6 can survive multiple disk failures (two, exactly), as can RAID 1+0, RAID 1+5, and even RAID 1. All of these are "sane" and in fact widely used. RAID5 is the only level that can't survive multiple failures.
Take, as an example, the DBI perldoc. It starts with succinct method documentation in the form of examples, in the order they might be used in an actual program. Then it moves on to useful newbie information, and a long, complete-sentence description of the way DBI works. And, since the page is in a linear format, I can find the specific documentation I need with one tap of my little finger on the / key.
Now let's look at a Javadoc from the Blackberry J2ME software. The front page has all the class names on the left, with terse descriptions in tables on the main page. There is no main description which tells me what the software is, nor how to use it. The packages are described like "The UI API provides a set of features for implementation of user interfaces for MIDP applications", however there is no definition of MIDP here, nor indeed is there any such definition anywhere in the Javadoc tree. Worse, since the javadoc is in a huge tree of separate HTML pages, I can't just quicksearch for the bit I'm looking for. I have to switch to the terminal and use grep -r, or try my luck with Google.
Here's the description for Class MessageFormat:
That's swell. WTF is a formatter object? A pattern? Maybe the superclass will tell us:Huh, no help there either. I don't know what it is about POD or Javadoc, but most of the time 'perldoc Whatever::Package' gives you some seriously useful documentation, and in my experience nearly any given Javadoc is totally useless to the beginner.
I happen to think that POD is far better for in-line comments than either Javadoc or Doxygen.
The following is an example of useless documentation:
frobWoggleFfloofMoing
public void frobWoggleFfloofMoing(String, String, String)
...
You see, running Doxygen over your header files may produce some output in HTML format, but it doesn't produce what I like to call "documentation." For instance, documentation would explain what is a Woggle, and when should it be Frobbed?
Thank you, and have a nice life.
You weep for slashdot? Take your order-5 userid and cram it up your rectum. You cannot construct an argument by cutting and pasting talking points from freerepublic.com.
I don't know about parkinson's, but the same device is used to treat severe epilepsy.
Yes, and if you weren't an ignorant twat you would realize that modern refrigerators have coefficients of performance in excess of unity. Well in excess, normally.
The difference is one programs Windows, one Unix, and one mainframes. As a fifth-year geek, you should take the rantings of Joel, ESR, and any other pointless windbag and send them to the bit bucket.
As far as I am concerned, Cendant drew its sword and now they cannot avoid battle. Tough shit for them.
Yeah, nice CPU, difficult for software authors. I read a paper recently wherein the authors managed to reduce L4 microkernel message passing (up to 8 bytes) to 36 clock cycles, which is far faster than any other platform. But this was done by hand, and the compiler blurted out a routine that required 508 cycles. The gulf between what you can really do with an Itanium, and what normal software writers can do with it, remains huge.
So ... you don't have a cell phone? No Wi-Fi?
I think a real problem for DVD-A is the previous two decades of direct-to-digital production. If a recording was made to 16/44.1 stereo masters, there's absolutely no point in releasing it on high definition multichannel media. The best you can hope for is to sound exactly like the CD, and what's more likely to happen is a corny surround remaster where the drum kit is inexplicably behind the listener. So there's this huge catalog of (mostly pop) music from the 80s and 90s made to low-definition digital formats, and they can never be improved by DVD-A nor SACD.
If the situation were sane, people would buy more DVD-Audio.
I think you are on crack. There's no p-layer in a Schottky barrier and therefore no (or little) stored charge when conducting. Schottky rectifiers typically have reverse conduction of about 100% forward current during recovery, and settle to DC reverse current within 50ns and in a single damped excursion. I've never even heard of anyone using snubbers with Schottky diodes in applications up to 1GHz.
Most equipment uses both bad regulation and bad bypass, so the point is really moot. Replacing the capacitors in or modifying the design of the bypass can yield a substantial improvement, especially for DACs. Replacing bad linear regulators with better designs can give you good regulation at 10kHz where previously the case was hopeless.
Probably nobody uses DVD-Audio. The effort spent cracking the copy protection will be proportional to the popularity of the medium. When the DVD Audio section at your local record store is as large as the CD section, we may see some more time being spent on such things.
It's inevitable that the netheads will bury the bellheads. The only question is: when? 2007? 2020?
Sure. I hit the "Forward" button on my Lucent digital set, punch in my phone number, and any calls to my desk ring my cell or home phones. No flaky VPN nor softphone required. The PSTN is many things, but featureless is not one of them.
By the way, "PBX" and "VoIP" are orthogonal, not exclusive. You can have a VoIP-connected PBX just the same as you can have a T1 or POTS or even a cellular-connected PBX. The PBX is just a telephony device that doesn't care what kind of line card you use to make and receive calls.
With business DSL services you often have an entire dedicated line (such as a T1) to your VoIP provider's head end, which allows prioritization of packets in both directions. The branch office where I work has VoIP on a dedicated T1 to Speakeasy and, after some initial problems with the connection, we are now getting better voice quality and better service for about a twentieth what we used to pay Avaya and WorldCom for the same number of lines.
Another tactic is to employ a sample rate converter, which can isolate two clock domains but unfortunately introduces some distortion. The distortion can be pretty significant for ratios very close to 1. Sample rate conversion is nearly universal in consumer equipment with a digital input. Unfortunately the clocks on both sides of the SRC are usually crap. There's plenty of opportunity to improve consumer equipment by modifying or replacing the clock or the clock's power supply.