With the advent of PVRs and PVR software such as MythTV, it's a damn good idea to combine a computer and a TV.
I think at some point someone was working on a MythTV module called MythRecipe even...
The nice thing about Myth is that the kitchen kiosk doesn't need a coaxial drop - A cat5 or 802.11g connection will give you both net access for bringing up recipes and the ability to stream video from a central recording box that contains the tuners.
I've been willing to forgive a lot of editorial inconsistencies on the part of the/. editors (dupes, etc)... Overall they've done a good job, it's hard to manage such a large site with so much traffic.
But please, stop posting all this unsubstantiated slander and bashing in the stories. First there was the bashing of Six Apart when they were purchasing LiveJournal, without ANY evidence WHY Six Apart was bad or even why the author didn't like them. (Which directly conflicts with everything I've heard from personal friends on LJ's staff, who were all extremely happy about the buyout - Many of them who were contractors with LJ were promoted to full time when SA purchased LJ.)
Now there is a story directly bashing a person, not just a company, with no real evidence as to why that person would deserve such bashing. The mailing list looks to ME like the developer in question politely handling complaints from a rather whiny user.
Is because the type of people who download Firefox are more likely to keep the torrent running after completion, resuling in many more seeders despite the small file size.
Take a look at the Knoppix torrents - Lots of people continue seeding the main ones long after completion, as a result the Knoppix torrents are insanely fast.
Overall I've been pretty unhappy with Linux support for external drives. Througput and performance are nowhere close to the performance of the same drive under Windows, especially if the drive is FAT32. (In general, the Linux FAT32 drivers suck performance-wise. Combine them with 1394 and it gets REALLY bad.)
I played around with CVS for a bit a few years ago, and found it to be too much of a pain in the ass for personal use.
SVN is so much better, and IMO easier to use. I'm using it for all future personal projects. (Already have the LaTeX source for my resume in a personal SVN repos)
In addition, I would not be surprised if JPEG and TIFF compression were done in dedicated hardware, and PNG would be impossible to add to most digital cameras without massive sacrifices in performance.
The Canon PowerShot S100 (aka Digital Elph) and all of its Digital Elph descendants basically were THE definition of a compact camera. (At least they were if you actually wanted decent camera functionality, those micro-pen-cams, etc really suck)
Take a guess which memory format the Digital Elphs use.
b) AT&T was never one of the telcos in question, or if they were, they were one of the first ones to change their tune. AT&T has been behind VoIP for many years, they were wise enough to see that it was the wave of the future and that it was better to embrace the technology and make money off of it rather than to fight it.
Personally, I'm surprised AT&T won in the Installation and Configuration categories. From what I recall, AT&T uses MGCP (rather than SIP, which everyone else uses), and SIP is much more NAT-friendly than MGCP. The people doing the testing probably never tested in a NAT enviroment, which is probably far more prevalent than a non-NAT environment these days.
Many of the companies listed (IBM, etc.) actually DO innovate, and use the patent system the way it was intended. IBM rarely tries to sue infringers of their patents out of existence - They (and MOST other companies) are smart enough to know that it's better to negotiate a reasonable licensing agreement.
Unfortunately, it's the abusers that open up immediately with a lawsuit that give the system a bad name.
A family member of mine worked for Lucent's intellectual property division. For them, it was considered to be a last resort to take a patent assertion to court. 99% of the time there was a behind-the-scenes licensing (often cross-licensing - "I let you use my patents, you let me use yours") agreements.
Lucent has clearly gone downhill over the years. In the past, they most likely would have been on that list thanks to Bell Labs. Lucent/AT&T used to spend a LOT of money on R&D, including very forward-thinking basic research. Those expenditures brought us things like the transistor (which is generally considered to be an IP licensing success story - The transistor was licensed out VERY reasonably.)
a) Transmit amp was being overdriven and became nonlinear. Not a problem for FSK modulation, but BAD for almost any other modulation scheme.
b) Transmit amp overheated, making any nonlinearities even worse. The WRT54G has utterly shitty thermal management. I'm shocked that the unit didn't crash within 15 minutes from overheating with the transmit power cranked - My friends' WRT54G was just returned after only two days because it couldn't even last 15 minutes with moderate WLAN usage before overheating and crashing.
The most common router supported by that router (Linksys WRT54G) is a POS with horrible thermal management.
I was just at a small LAN party at a friends' house. For simplicity, we decided to try and go wireless.
Big mistake. Router overheated and crashed every 10-15 minutes when the WLAN transceiver was being used heavily, and every 2-3 hours when only wired connections were being used.
So if your errors are +- 50, then you'd better work on how you control your bloodsugars. (Assuming you mean mg/dl, where 80-120 is the "normal" range.)
I'm a serious geek, but downloading my numbers to my computer (even with my new IR transfer capability) is just more trouble than it's worth. I usually don't bother except for one dump just before visiting the endocrinologist, and that's usually only good for averages since I can't be arsed to keep the clock accurate. (My Glucometer's clock drifted BADLY, the Accu-Chek's battery compartment is too loose and it completely loses its time frequently.)
Implement wireless Palm HotSync-style syncing. (Hit button on the meter with zero interaction on the meter itself) and people might use it. Bluetooth might be a good bet here. Make the fucking protocol open by the way.
Someone mentioned a "pseudo-meter" in a CF device. Bad idea. The hottest feature in meters right now is multi-strip cartridges/drums. Can't fit those in a CF form factor. After going multi-strip (first with a Glucometer Dex and now an Accu-Chek Compact) I will NEVER go back to a single-strip device.
Go to the Zaurus User Group forums (I think www.zaurususergroup.net), in the Off-Topic section there's a thread with data dumps of the Compact's protocol and a Perl script I wrote to dump the data. Right now it just prints time/date/reading to stdout, but it would be easy to modify to dump into a MySQL DB.
Not 100% accurate is an understatement - it was apparently severly inaccurate if you were sweating.
It also caused significant skin irritation (People preferred the occasional pinprick) and was insanely expensive. I was looking forward to it greatly but when it was released, the reviews were so horrible I didn't even bother.
I believe the company is out of business now. The company didn't even come close to selling enough units to pay for all the R&D, the watch bombed so badly.
The One Touch II (a classic, and for a long time THE meter used by diabetics.) Big, clunky, single-strip design.
Glucometer Dex (There is a Dex II now, I can't see what the difference is.)
Accu-Check Compact
I would never use the One Touch again - Lifescan still has no cartridge-based meters, and being able to load the meter with multiple strips is REALLY nice.
The Glucometer was a damn nice meter. 10 strips per cartridge (they were the first company to release a cartridge-based meter), you could order the interface cable for free and their software (WinGlucofacts) was pretty nice, and it was free (free as in beer, that is) too.
The Accu-Chek Compact is also really nice. Pros: IR data transfer 17 strips per cartridge Formulary with my insurance company. That's $30 per strip refill. (You'll see a common theme that the insurance company dictates what type of insulin/meter you use to some degree. Although I'm using a Novo Nordisk insulin pen with Novo cartridges because there are no pens for Lilly insulins that dispense half-unit increments anymore.) AA batteries - the button cells in the Dex were annoying. Form factor and carrying case make it great to put in a jacket pocket for a night out. Adjustable puncture depth lancet device. (The other meters had this too but you had to change tips to do so.)
Cons: Battery holder is loose. I often find myself having to push the batteries back in. Meter loses its time frequently because of this and has to reset the drum. The "find next unused drum slot" feature is nice in cases where you accidentally take the batteries out for too long or accidentally hit the cartridge chamber release. Cartridge chamber is too easy to unlatch, and comes unlatched often just by removing the meter from its case. No open protocol documentation, software is $30 and may be rigged to only work with their external serial-to-IR adapter instead of a normal serial IRDA port (they do not mention IRDA ports at all, they ALWAYS try to sell their $30 adapter even though I know the meter can speak to a normal IRDA port.) There are protocol dumps linked to from the Zaurus User Group forums, I think the "off-topic" section. I wrote a Perl script from those dumps, it's posted to those forums. Form factor of the meter/case is good for a night on the town but not as good for extended trips. The Dex case had lots of extra pockets for pen needles, extra lancets, etc. and was very flat. The Compact case has no extra pockets, just an elastic band to hold a spare test strip drum and one to hold the lancet device. Lancet device is not compatible with half the lancets on the market, including my personal favorite, the B-D Ultra Fine. Most lancet devices have a round holder that can also hold "flat" lancets via a friction fit. The Accu-Chek lancet device can only hold "flat" type lancets, not round ones like the B-Ds.
The Compact also allows "alternative site" testing (i.e. not the finger) but has so many restrictions on when you can do it (see parent posts' mention of lag in the reading) that I never bother.
Don't try to go too fancy. People say they will want it, but so far every attempt at a meter that "does everything" (onboard statistics/data collection) has bombed. Non-invasive testing is a VERY desirable feature, but so far no one has succeeded in making an accurate NI meter that wasn't more trouble than it was worth. (The Glucowatch was expensive, known for causing skin irritation, and not very accurate. Due to the skin irritation, people preferred the occasional pinprick, which with a good lancet and properly adjusted lancet device depth, you barely feel.)
My advice: Try and convince Lilly and Novo Nordisk to bring back their 1.5 mL pen cartridges. Pens for 3 mL carts such as my new NovoPen Junior are fucking huge.:(
ZSNES has had hardware scaling via OpenGL for quite a while. Although my favorite so far has been 4x software scaling with HQ4X (which is a scaling algorithm that is designed specifically for video games, which typically have lots of line art. HQnX attempt to perform pattern detection on the input graphics in order to guess what it's supposed to look like, and for most video games, the HQnX algorithms and their predecessor 2xSAI work quite well.) Of course, nowadays I follow HQ4X with further hardware scaling on my laptop's 1600x1200 screen.:)
Not sure how hard it would be to get the SDL Linux code to work under OSX... Might not be hard at all.
Slashdot Mirror
With the advent of PVRs and PVR software such as MythTV, it's a damn good idea to combine a computer and a TV.
I think at some point someone was working on a MythTV module called MythRecipe even...
The nice thing about Myth is that the kitchen kiosk doesn't need a coaxial drop - A cat5 or 802.11g connection will give you both net access for bringing up recipes and the ability to stream video from a central recording box that contains the tuners.
IIRC, BroadVoice is about $25/month, and includes unlimited calling to the US, Canada, and a few other countries.
I've been willing to forgive a lot of editorial inconsistencies on the part of the /. editors (dupes, etc)... Overall they've done a good job, it's hard to manage such a large site with so much traffic.
But please, stop posting all this unsubstantiated slander and bashing in the stories. First there was the bashing of Six Apart when they were purchasing LiveJournal, without ANY evidence WHY Six Apart was bad or even why the author didn't like them. (Which directly conflicts with everything I've heard from personal friends on LJ's staff, who were all extremely happy about the buyout - Many of them who were contractors with LJ were promoted to full time when SA purchased LJ.)
Now there is a story directly bashing a person, not just a company, with no real evidence as to why that person would deserve such bashing. The mailing list looks to ME like the developer in question politely handling complaints from a rather whiny user.
Really, it's getting out of hand...
Adds version control for the config file replacement system, and lets you see what changes are going to be made.
It's much nicer than etc-update
Is because the type of people who download Firefox are more likely to keep the torrent running after completion, resuling in many more seeders despite the small file size.
Take a look at the Knoppix torrents - Lots of people continue seeding the main ones long after completion, as a result the Knoppix torrents are insanely fast.
I think in some cases the throughput can be lower than a hard drive.
Plus USB2 is a horrendous bottleneck.
FAT32 + Firewire = HORRIBLE performance
Overall I've been pretty unhappy with Linux support for external drives. Througput and performance are nowhere close to the performance of the same drive under Windows, especially if the drive is FAT32. (In general, the Linux FAT32 drivers suck performance-wise. Combine them with 1394 and it gets REALLY bad.)
I played around with CVS for a bit a few years ago, and found it to be too much of a pain in the ass for personal use.
SVN is so much better, and IMO easier to use. I'm using it for all future personal projects. (Already have the LaTeX source for my resume in a personal SVN repos)
In addition, I would not be surprised if JPEG and TIFF compression were done in dedicated hardware, and PNG would be impossible to add to most digital cameras without massive sacrifices in performance.
The Canon PowerShot S100 (aka Digital Elph) and all of its Digital Elph descendants basically were THE definition of a compact camera. (At least they were if you actually wanted decent camera functionality, those micro-pen-cams, etc really suck)
Take a guess which memory format the Digital Elphs use.
Yeah, Urban Terror is lots of fun.
:(
Although I preferred Q3F, unfortunately Q3F is basically dead, no one plays it anymore.
My netcode rankings:
Classic QW is #1, period
Q3 is second, behind by quite a bit
UTxxx is waaaaay behind both of them.
a) That article is 8 years old.
b) AT&T was never one of the telcos in question, or if they were, they were one of the first ones to change their tune. AT&T has been behind VoIP for many years, they were wise enough to see that it was the wave of the future and that it was better to embrace the technology and make money off of it rather than to fight it.
Personally, I'm surprised AT&T won in the Installation and Configuration categories. From what I recall, AT&T uses MGCP (rather than SIP, which everyone else uses), and SIP is much more NAT-friendly than MGCP. The people doing the testing probably never tested in a NAT enviroment, which is probably far more prevalent than a non-NAT environment these days.
Many of the companies listed (IBM, etc.) actually DO innovate, and use the patent system the way it was intended. IBM rarely tries to sue infringers of their patents out of existence - They (and MOST other companies) are smart enough to know that it's better to negotiate a reasonable licensing agreement.
Unfortunately, it's the abusers that open up immediately with a lawsuit that give the system a bad name.
A family member of mine worked for Lucent's intellectual property division. For them, it was considered to be a last resort to take a patent assertion to court. 99% of the time there was a behind-the-scenes licensing (often cross-licensing - "I let you use my patents, you let me use yours") agreements.
Lucent has clearly gone downhill over the years. In the past, they most likely would have been on that list thanks to Bell Labs. Lucent/AT&T used to spend a LOT of money on R&D, including very forward-thinking basic research. Those expenditures brought us things like the transistor (which is generally considered to be an IP licensing success story - The transistor was licensed out VERY reasonably.)
Most likely two things:
a) Transmit amp was being overdriven and became nonlinear. Not a problem for FSK modulation, but BAD for almost any other modulation scheme.
b) Transmit amp overheated, making any nonlinearities even worse. The WRT54G has utterly shitty thermal management. I'm shocked that the unit didn't crash within 15 minutes from overheating with the transmit power cranked - My friends' WRT54G was just returned after only two days because it couldn't even last 15 minutes with moderate WLAN usage before overheating and crashing.
The most common router supported by that router (Linksys WRT54G) is a POS with horrible thermal management.
I was just at a small LAN party at a friends' house. For simplicity, we decided to try and go wireless.
Big mistake. Router overheated and crashed every 10-15 minutes when the WLAN transceiver was being used heavily, and every 2-3 hours when only wired connections were being used.
Needless to say, he returned it the next day.
2wire tends to sell to more technically inclined users.
There's a reason you don't see their products in Worst Buy, etc.
The number is approx 10-20%.
So if your errors are +- 50, then you'd better work on how you control your bloodsugars. (Assuming you mean mg/dl, where 80-120 is the "normal" range.)
I agree.
I'm a serious geek, but downloading my numbers to my computer (even with my new IR transfer capability) is just more trouble than it's worth. I usually don't bother except for one dump just before visiting the endocrinologist, and that's usually only good for averages since I can't be arsed to keep the clock accurate. (My Glucometer's clock drifted BADLY, the Accu-Chek's battery compartment is too loose and it completely loses its time frequently.)
Implement wireless Palm HotSync-style syncing. (Hit button on the meter with zero interaction on the meter itself) and people might use it. Bluetooth might be a good bet here. Make the fucking protocol open by the way.
Someone mentioned a "pseudo-meter" in a CF device. Bad idea. The hottest feature in meters right now is multi-strip cartridges/drums. Can't fit those in a CF form factor. After going multi-strip (first with a Glucometer Dex and now an Accu-Chek Compact) I will NEVER go back to a single-strip device.
Go to the Zaurus User Group forums (I think www.zaurususergroup.net), in the Off-Topic section there's a thread with data dumps of the Compact's protocol and a Perl script I wrote to dump the data. Right now it just prints time/date/reading to stdout, but it would be easy to modify to dump into a MySQL DB.
Not 100% accurate is an understatement - it was apparently severly inaccurate if you were sweating.
It also caused significant skin irritation (People preferred the occasional pinprick) and was insanely expensive. I was looking forward to it greatly but when it was released, the reviews were so horrible I didn't even bother.
I believe the company is out of business now. The company didn't even come close to selling enough units to pay for all the R&D, the watch bombed so badly.
Type 1 here, no pump.
:(
I've used three meters in my life:
The One Touch II (a classic, and for a long time THE meter used by diabetics.) Big, clunky, single-strip design.
Glucometer Dex (There is a Dex II now, I can't see what the difference is.)
Accu-Check Compact
I would never use the One Touch again - Lifescan still has no cartridge-based meters, and being able to load the meter with multiple strips is REALLY nice.
The Glucometer was a damn nice meter. 10 strips per cartridge (they were the first company to release a cartridge-based meter), you could order the interface cable for free and their software (WinGlucofacts) was pretty nice, and it was free (free as in beer, that is) too.
The Accu-Chek Compact is also really nice.
Pros: IR data transfer
17 strips per cartridge
Formulary with my insurance company. That's $30 per strip refill. (You'll see a common theme that the insurance company dictates what type of insulin/meter you use to some degree. Although I'm using a Novo Nordisk insulin pen with Novo cartridges because there are no pens for Lilly insulins that dispense half-unit increments anymore.)
AA batteries - the button cells in the Dex were annoying.
Form factor and carrying case make it great to put in a jacket pocket for a night out.
Adjustable puncture depth lancet device. (The other meters had this too but you had to change tips to do so.)
Cons:
Battery holder is loose. I often find myself having to push the batteries back in. Meter loses its time frequently because of this and has to reset the drum. The "find next unused drum slot" feature is nice in cases where you accidentally take the batteries out for too long or accidentally hit the cartridge chamber release.
Cartridge chamber is too easy to unlatch, and comes unlatched often just by removing the meter from its case.
No open protocol documentation, software is $30 and may be rigged to only work with their external serial-to-IR adapter instead of a normal serial IRDA port (they do not mention IRDA ports at all, they ALWAYS try to sell their $30 adapter even though I know the meter can speak to a normal IRDA port.) There are protocol dumps linked to from the Zaurus User Group forums, I think the "off-topic" section. I wrote a Perl script from those dumps, it's posted to those forums.
Form factor of the meter/case is good for a night on the town but not as good for extended trips. The Dex case had lots of extra pockets for pen needles, extra lancets, etc. and was very flat. The Compact case has no extra pockets, just an elastic band to hold a spare test strip drum and one to hold the lancet device.
Lancet device is not compatible with half the lancets on the market, including my personal favorite, the B-D Ultra Fine. Most lancet devices have a round holder that can also hold "flat" lancets via a friction fit. The Accu-Chek lancet device can only hold "flat" type lancets, not round ones like the B-Ds.
The Compact also allows "alternative site" testing (i.e. not the finger) but has so many restrictions on when you can do it (see parent posts' mention of lag in the reading) that I never bother.
Don't try to go too fancy. People say they will want it, but so far every attempt at a meter that "does everything" (onboard statistics/data collection) has bombed. Non-invasive testing is a VERY desirable feature, but so far no one has succeeded in making an accurate NI meter that wasn't more trouble than it was worth. (The Glucowatch was expensive, known for causing skin irritation, and not very accurate. Due to the skin irritation, people preferred the occasional pinprick, which with a good lancet and properly adjusted lancet device depth, you barely feel.)
My advice: Try and convince Lilly and Novo Nordisk to bring back their 1.5 mL pen cartridges. Pens for 3 mL carts such as my new NovoPen Junior are fucking huge.
Gdesklets is nice for having such info on your desktop on a Linux box.
For "screensaver" style stuff, I'm not sure. One could probably do some gdesklets hacking to get similar info.
As one other poster said, samplerate conversion for sound is one reason. The other reason would be for graphics scaling.
ZSNES has had hardware scaling via OpenGL for quite a while. Although my favorite so far has been 4x software scaling with HQ4X (which is a scaling algorithm that is designed specifically for video games, which typically have lots of line art. HQnX attempt to perform pattern detection on the input graphics in order to guess what it's supposed to look like, and for most video games, the HQnX algorithms and their predecessor 2xSAI work quite well.) Of course, nowadays I follow HQ4X with further hardware scaling on my laptop's 1600x1200 screen. :)
Not sure how hard it would be to get the SDL Linux code to work under OSX... Might not be hard at all.