IIRC, those tests were done by a lab closely associated with Microsoft. i.e, MS had already fixed up those problems in IE and deliberately got someone to "discover" how it was better in this one tiny area. Just like the infamous Mindcraft tests all those years ago. I don't know if Mozilla has fixed its code yet.
That's why it's in quotes - "iPod killer". The iPod has become so ubiquitous with portable audio that anything similar coming along is inevitably labeled an "iPod killer". The Media just love to portray such situations in great adversarial language. But the phrase has been used so much now, and failed to come true, that it's become a cliche. So now it's used as a parody of the pundits and journalists that once (and still) use the term.
That sounds like a hard way to manage cookies. I just set "Cookie lifetime policy" to "ask for each cookie". Since I always have the "use this choice for cookies from this site" checkbox ticked, I don't get the dialog poping up too often. Just the first time I go to a new site, which usually isn't too often. Once you've approved your usual haunts, it isn't much of a hassle.
I wonder what would have happened if the AMD64 had not come along...
I also wonder what would have happened if Intel had created their own 64-bit extensions to x86 instead of creating ia64. Considering that the ia64 project with HP started 10 years ago, we could have seen an x86-64 architecture from Intel instead of AMD. Which AMD would surely have cloned. Or Intel might have developed the 64-bit extensions at about the same time as AMD, in which case we'd have two competing x86-64 architectures! But instead we have AMD64 and an almost-clone from Intel.
I think you might be onto something there. I wonder how much pressure there is from Sony Music, and the rest of the RIAA behemoth, on the rest of Sony to cripple anything dealing with digital music/audio. Making a portable digital music player these days that doesn't play MP3, WMA, or AAC and forcing their own crappy proprietary DRM'ed codec on people is just plain dumb.
The RIAA is really a bunch of dinosaurs and I derive great satisfaction from watching non-music-industry companies (and the Open Source community) show them the kinds of truly great things that can be done with technology.
</rant>
Ok, it was a bit of a generalization. From what I understand, most consumer/prosumer cameras are in a certain range of sensor sizes. Like you said, price is the main force "wanting" to make the sensors small. But you also have sensitivity and noise issues wanting to make the sensors large. Different models and makers have different priorities and acceptable performances. At one end there's crappy little sensors for little "toy" cameras. At the other end is large sensors for professional cameras. In the middle are a range of consumer and "prosumer" cameras that probably make up most of the market.
If you want huge, the Hasselblad H1D has a 22MP 36.7x49mm sensor. I can't imagine the yield from fabricating such a large chip to be very good. Big $$$
However, it seems like these days (especially at the lower end of the market) there is too much emphasis on megapixels - the manufacturers push out shiny 4MP cameras and everyone buys them because they think that the more pixels the better, but often the optics on those cameras are so poor that the images are not really any better than you'd get out of a 2 or 3MP camera.
Also, more pixels == smaller pixels. There goes the low-light sensitivity. And the signal coming off of the sensor has to be amplified to compensate for the smaller photon-collecting area of each pixel. So any image will have more noise. I'd rather get a 2-3MP camera and focus more on the lens and other features. You do have a point about cropping though.
I've handed down my old Olympus 460Z to my daughter (4 1/2 years old) to play with, and she's having a blast.
I can imagine. I got my first camera when I was closer to 7 or 8 (9?). My brother and I each got a Kodak disc camera for christmas. I used up most of my first two discs (15 exposures each!) just photographing my presents. I calmed down a bit after that:) They went into the closet after a few years because of the inconvenience of only 15 exposures and the trouble of getting new film. If there was anything important to photograph then we used my mums' much better 35mm compact. Now with the extreme convenience of digital photography, I can't imagine going back to such a world. Well, except for semi-serious photography.
So what sort of things does a 4.5 y/o with a digicam photograph? How does she handle transferring the photos to a PC? And/or does she use the composite video out?
For example, when I want to kill a print job that is sending garbage to my printer, where's the admin gui that lets me browse queues, select the errant job, and kill it? I rely on lpq for that today.
Um... The web interface perhaps? You know, the thing hanging out on port 631. Are you trying to tell me you're using CUPS without the web interface? That could certainly explain some of your problems. The web interface isn't perfect, but it's a lot easier to work than the CLI. And it doesn't require a telnet/ssh login.
The only problem I've had with CUPS is with my parent's cheap Lexmark inkjet. The CUPS driver (or something) is unable to sense when the printer is off. I think the printer must still hang around on the USB bus without sending a disconnect to the PC. So if someone tries printing while the printer is off, then CUPS gets all jammed up. I have to shutdown CUPS, and then do a forcefull (SIGKILL) shutdown of the usb process (CUPS is very modular). I also have to pull the USB plug out of the printer and reinsert it just to be sure that the kernel knows what is going on. Then I can turn the printer on and restart CUPS. The only solution is to keep the printer on all the time. I've never had this problem with my old Epson STC740. When I turn it off, I can see in the kernel log that it disconnects from the USB bus.
Jeff Belluzo, or however his name is spelt. I read recently that before SGI he was the CEO of HP (or DEC? somewhere) and had pretty much done the exact same thing there as well.
Hey, lets drop our great Unix-on-RISC machines and make Wintel PCs! We'll lose control of most of the hardware and software, and be competing with every other beige-box maker! We'll still charge more though, for our brand-name and support. That is, until everyone wises up. Sounds like a great plan! Lets hear it for conformity and blandness!
You are correct though, in that it is a management problem. If I understand things correctly, the model intended for Jabber is one where ISP's and other organizations set up their own Jabber servers. Just like email. These servers would likely be much more managable because the sysadmin(s) would have more information on hand than some monkey at GlobalMegaCorp. I'm sure AOL couldn't care less about dropping <1% of their AIM accounts. Let ISP's and others run their own Jabber servers. In the case of an ISP then the jabber account would at least be tied to their modem/ADSL/email account. Mistakenly deleting an account would be very unlikely. In some cases (e.g a small org or company) the sysadmin will know most of the users.
Exactly. Lukes' posts here don't present a very good picture of his personality. Frankly, he comes off as a pompus jerk. Maybe he is as good as he claims and has made some good contributions. But just about every one of his posts here contain some annecdote claiming I was too good and they were threatened by my skill and knowledge. I know I'd find it hard to work with someone like that.
Jeremy Allison has already weighed in saying simply that working with him was difficult and didn't justify the benefits.
Well, yes we do. But don't be mistaken into thinking that "root" is equivalent to "fuck". "fuck" is a very flexible word and can be used in many, many ways. "root" simply means to have sex, like "screw" or "shag".
And thus concludes todays lesson in Australian slang by a bored aussie. Tune in tomorrow to hear me explain "koala ears" and "budgie smugglers".
AFAIK, there is no standard mapping for >2 channels in Ogg Vorbis. So yes, there is that problem. I thought that Ogg and/or Vorbis was supposed to have more meta-data (and meta-meta-data?) than other formats. Perhaps I'm mistaken.
As for file size, Vorbis does have a very flexible system for channel coupling, much like joint-stereo in MP3. In the reference encoder it's only used for q >= 6 though. So adding more channels should have a diminishing impact on bitrate requirements, assuming that similar sounds are coming out all/most of the channels.
The Boost project has Boost.Python. I haven't used it yet, but the docs make it sound very interesting. It looks quite simple to write base classes in C++ and then subclass them in Python.
Small, very concealable cameras have been available for quite a long time.
Yeah, because you can't take pictures of kids with a regular camera...
It all comes down to the purpose of the device. A camera takes photos and nothing else. If you get caught with a camera, it's pretty obvious that you'd be using it to take photos. But modern mobile/cell phones have cameras incorporated into the device. A camera-phone is more innocuous because it still looks like an ordinary phone. And they're a much more common part of everyday life than a camera. They're something that a person carries with themselves everywhere, just a part of the scenery.
Now, merely being *seen* with a camera in a sensitive area automatically raises suspicion that you just have or will shortly use the camera to take a photo. But you don't necessarily get "caught" when someone sees you with a mobile/cell phone, unless (or until) you're clearly pointing it at a subject to take a photo.
You own the phone but have no control over the software running on it.
Well of course. We know that. But most people probably don't realize the possibilities. Or that they could possibly (?) be reprogrammed remotely without the mobile/cell phone ever leaving their possession. People just naturally trust their own devices. That's what I was trying to get across. People have this natural assumption that if something is theirs that they control it.
Slashdot Mirror
2.60? Shit, Linus really should jump the numbering to 3.0! :)
Merry christmas from Australia.
IIRC, those tests were done by a lab closely associated with Microsoft. i.e, MS had already fixed up those problems in IE and deliberately got someone to "discover" how it was better in this one tiny area. Just like the infamous Mindcraft tests all those years ago. I don't know if Mozilla has fixed its code yet.
That's why it's in quotes - "iPod killer". The iPod has become so ubiquitous with portable audio that anything similar coming along is inevitably labeled an "iPod killer". The Media just love to portray such situations in great adversarial language. But the phrase has been used so much now, and failed to come true, that it's become a cliche. So now it's used as a parody of the pundits and journalists that once (and still) use the term.
That sounds like a hard way to manage cookies. I just set "Cookie lifetime policy" to "ask for each cookie". Since I always have the "use this choice for cookies from this site" checkbox ticked, I don't get the dialog poping up too often. Just the first time I go to a new site, which usually isn't too often. Once you've approved your usual haunts, it isn't much of a hassle.
I also wonder what would have happened if Intel had created their own 64-bit extensions to x86 instead of creating ia64. Considering that the ia64 project with HP started 10 years ago, we could have seen an x86-64 architecture from Intel instead of AMD. Which AMD would surely have cloned. Or Intel might have developed the 64-bit extensions at about the same time as AMD, in which case we'd have two competing x86-64 architectures! But instead we have AMD64 and an almost-clone from Intel.
No, still Mhz. IIRC, gigE is made of four 250Mhz channels. That's all four pairs of a cat5 cable, thus it's only half-duplex.
I think you might be onto something there. I wonder how much pressure there is from Sony Music, and the rest of the RIAA behemoth, on the rest of Sony to cripple anything dealing with digital music/audio. Making a portable digital music player these days that doesn't play MP3, WMA, or AAC and forcing their own crappy proprietary DRM'ed codec on people is just plain dumb.
The RIAA is really a bunch of dinosaurs and I derive great satisfaction from watching non-music-industry companies (and the Open Source community) show them the kinds of truly great things that can be done with technology.
</rant>
Ok, it was a bit of a generalization. From what I understand, most consumer/prosumer cameras are in a certain range of sensor sizes. Like you said, price is the main force "wanting" to make the sensors small. But you also have sensitivity and noise issues wanting to make the sensors large. Different models and makers have different priorities and acceptable performances. At one end there's crappy little sensors for little "toy" cameras. At the other end is large sensors for professional cameras. In the middle are a range of consumer and "prosumer" cameras that probably make up most of the market.
If you want huge, the Hasselblad H1D has a 22MP 36.7x49mm sensor. I can't imagine the yield from fabricating such a large chip to be very good. Big $$$
Also, more pixels == smaller pixels. There goes the low-light sensitivity. And the signal coming off of the sensor has to be amplified to compensate for the smaller photon-collecting area of each pixel. So any image will have more noise. I'd rather get a 2-3MP camera and focus more on the lens and other features. You do have a point about cropping though.
Thanks for the correction. I was sure that I'd mispelled his surname, not realising that I'd totally mis-remembered his first name! :P
An AC gave this link to Rick Belluzzo's bio at Quantum.
I can imagine. I got my first camera when I was closer to 7 or 8 (9?). My brother and I each got a Kodak disc camera for christmas. I used up most of my first two discs (15 exposures each!) just photographing my presents. I calmed down a bit after that :) They went into the closet after a few years because of the inconvenience of only 15 exposures and the trouble of getting new film. If there was anything important to photograph then we used my mums' much better 35mm compact. Now with the extreme convenience of digital photography, I can't imagine going back to such a world. Well, except for semi-serious photography.
So what sort of things does a 4.5 y/o with a digicam photograph? How does she handle transferring the photos to a PC? And/or does she use the composite video out?
Um... The web interface perhaps? You know, the thing hanging out on port 631. Are you trying to tell me you're using CUPS without the web interface? That could certainly explain some of your problems. The web interface isn't perfect, but it's a lot easier to work than the CLI. And it doesn't require a telnet/ssh login.
The only problem I've had with CUPS is with my parent's cheap Lexmark inkjet. The CUPS driver (or something) is unable to sense when the printer is off. I think the printer must still hang around on the USB bus without sending a disconnect to the PC. So if someone tries printing while the printer is off, then CUPS gets all jammed up. I have to shutdown CUPS, and then do a forcefull (SIGKILL) shutdown of the usb process (CUPS is very modular). I also have to pull the USB plug out of the printer and reinsert it just to be sure that the kernel knows what is going on. Then I can turn the printer on and restart CUPS. The only solution is to keep the printer on all the time. I've never had this problem with my old Epson STC740. When I turn it off, I can see in the kernel log that it disconnects from the USB bus.
Jeff Belluzo, or however his name is spelt. I read recently that before SGI he was the CEO of HP (or DEC? somewhere) and had pretty much done the exact same thing there as well.
Me, cynical? Nah...
Yeesh, what is it with you and fanboys ?
You are correct though, in that it is a management problem. If I understand things correctly, the model intended for Jabber is one where ISP's and other organizations set up their own Jabber servers. Just like email. These servers would likely be much more managable because the sysadmin(s) would have more information on hand than some monkey at GlobalMegaCorp. I'm sure AOL couldn't care less about dropping <1% of their AIM accounts. Let ISP's and others run their own Jabber servers. In the case of an ISP then the jabber account would at least be tied to their modem/ADSL/email account. Mistakenly deleting an account would be very unlikely. In some cases (e.g a small org or company) the sysadmin will know most of the users.
Back.... Up? What is this "backup" of which you speak?
Exactly. Lukes' posts here don't present a very good picture of his personality. Frankly, he comes off as a pompus jerk. Maybe he is as good as he claims and has made some good contributions. But just about every one of his posts here contain some annecdote claiming I was too good and they were threatened by my skill and knowledge. I know I'd find it hard to work with someone like that.
Jeremy Allison has already weighed in saying simply that working with him was difficult and didn't justify the benefits.
A 3D screensaver on a server? Now that's the sort of innovative shit that Linux needs to succeed in the marketplace ;)
(Yes, I know about Xscreensaver)
Well, yes we do. But don't be mistaken into thinking that "root" is equivalent to "fuck". "fuck" is a very flexible word and can be used in many, many ways. "root" simply means to have sex, like "screw" or "shag".
And thus concludes todays lesson in Australian slang by a bored aussie. Tune in tomorrow to hear me explain "koala ears" and "budgie smugglers".
AFAIK, there is no standard mapping for >2 channels in Ogg Vorbis. So yes, there is that problem. I thought that Ogg and/or Vorbis was supposed to have more meta-data (and meta-meta-data?) than other formats. Perhaps I'm mistaken.
As for file size, Vorbis does have a very flexible system for channel coupling, much like joint-stereo in MP3. In the reference encoder it's only used for q >= 6 though. So adding more channels should have a diminishing impact on bitrate requirements, assuming that similar sounds are coming out all/most of the channels.
The Boost project has Boost.Python. I haven't used it yet, but the docs make it sound very interesting. It looks quite simple to write base classes in C++ and then subclass them in Python.
...or one that says "This one isn't for critics!" :)
I thought the Tux icon was also animated. Don't his eyes move also?
MacHall suggests a good book for attracting smart women :)
It all comes down to the purpose of the device. A camera takes photos and nothing else. If you get caught with a camera, it's pretty obvious that you'd be using it to take photos. But modern mobile/cell phones have cameras incorporated into the device. A camera-phone is more innocuous because it still looks like an ordinary phone. And they're a much more common part of everyday life than a camera. They're something that a person carries with themselves everywhere, just a part of the scenery.
Now, merely being *seen* with a camera in a sensitive area automatically raises suspicion that you just have or will shortly use the camera to take a photo. But you don't necessarily get "caught" when someone sees you with a mobile/cell phone, unless (or until) you're clearly pointing it at a subject to take a photo.
Well of course. We know that. But most people probably don't realize the possibilities. Or that they could possibly (?) be reprogrammed remotely without the mobile/cell phone ever leaving their possession. People just naturally trust their own devices. That's what I was trying to get across. People have this natural assumption that if something is theirs that they control it.
Will it advise you not to ring your ex when you've had too many drinks?