No. 3db = twice as much power. 10db = perceived as twice as loud.
In addition, the "twice as loud" perception is highly dependent on the type of noise. The 10db = twice as loud rule only works for pure tones at 1KHz. At other frequencies / types of sound, the perception is different.
Please note: I'm not talking about a situation where the government built an airport or some such thing near a previously quiet neighborhood. I'm talking about cases where the home-owner knew (or should have known) the conditions prior to purchase.
I've seen the same sort of short-sighted buying in the US in rural areas. I've seen places where people have built new houses a half mile from a livestock farm that has been there for 50 years, then when they finally move in, they discover that when the wind is blowing the right way, there's a smell. Then they try to get zoning changed, or they sue, or some other tactic, to try to get the farm closed. What, you didn't think pigs smelled? Or did you even check to see who your neighbors were?
Hm. I'll take a look at it. However, I like Agent, and for the same money, Agent gives me a fairly capable newsreader as well. I just wish they'd get moving and get 2.0 out, it's got most of what I've been wanting for the last 4 years.
Instead of backspace, you can try "b" for back a message. It works for me.
No, it doesn't. It just goes to the previous message in the list. it does NOT go to the most recently displayed message. I just tried it. If I "N" or spacebar through a group, hitting the new messages, say, numbers 5, 10,300, and 522, when I hit backspace, it goes back to 300, then to 10, then to 5. The B key in Thunderbird takes you to 521, then 520, then 519. Totally unhelpful.
The spam filter only works well when you mark a ton of messages as "spam" and a ton of messages as "not spam".
This is not at all clear, if true. I have thousands of messages that are marked as not spam. Do you mean I have to toggle them all to spam, then back to not spam? Is that how you mean? That's a pain. I've never seen a bayesian filter that makes you specifically mark things as nonspam. I'm using POPFile, and anything that I didn't tell it was spam, it assumes is nonspam. OK, SpamBayes asks for a set of example hams up front but after that it trains on "it's not spam so it must be ham."
I've tried 0.2 and most recently 0.3. I gave both of them about a week. Both times I went back to using Forte Agent. 0.2 was just not stable for me. 0.3 was stable, never crashed or lost stuff. My big problem is that after a few days of usage, it just started getting horribly slow. Also there are some usability problems that I start out thinking I can live with, but they eventually bug me too much:
- "thread" view is sorted wrong. They group by thread then sort ALPHABETICALLY. Sorry, group then sort BY TIME, at least optionally. Otherwise I've got recent threads at the top, ancient threads at the bottom, and thousands of emails between them. Browsing becomes nearly impossible; recent threads become needles in a haystack.
- No "backspace" when reading emails: both Agent and Thunderbird (and others) allow spacebar and "N"ext message to quickly browse through messages, Agent has a "backspace" key that remembers which messages you've read and backs you up through them. When you're in thread view of a mailing list that generates 100+ emails a day and you have 6 month's archives in the folder, once you leave a message you have almost no hope of finding it again without this feature.
- the spam filter is hopeless. I tagged well over 1000 spams, and it still was getting about 50% false negatives, and even worse, about 20% false positives. I'd pick up 50 emails, have 20 spams in there, it wouldn't ID 10 of the spams, and it would throw 5+ legitimate emails into the spam filter. POPfile got to be almost perfect far before this. Yes, I could use POPfile with Thunderbird, but I was hoping the feature would actually work.
- using it in large binaries groups is completely hopeless, especially on a good server with long retention. I went into a group that had about 300,000 messages on the server, and it just about coughed up a lung. It took it forever to do anything once that was loaded. Also it doesn't even appear to combine all parts of a multipart post into one display item; without this feature, you actually have to LOOK at all 300,000 items; this is ridiculous, other newsreaders have had this important feature for years.
There are other problems, but I've already forgotten them (I switched back to Agent two days ago).
Yeah, I could fix some of these if I wanted, and I did look in to that, but setting up the build environment is fairly involved, and I couldn't fix all of them without spending significant time learning the guts of the system.
I *want* Thunderbird to work, I just can't live with it yet. And I'm afraid some of the things that bug me about it might be fairly hard to fix.
If they're fined, they'll just increase their rates, perhaps they'll add another 10 cents to everyone's "Number portability surcharge". It will cost them nothing.
I'm seriously considering the 800 MHz fanless VIA as a replacement for my workhorse 200 MHz pentium pro. The low power consumption is a big plus, and it would still be plenty of speed for me, especially with a half GB of RAM in it.
I'm looking at building a new machine, and I'm planning on using the slowest chip I can find that saves me any money (IOW, I won't go from 2GHz to 1GHZ to save $5, but I would to save $50).
My workhorse, a headless linux box in the basement, is running on a Pentium Pro 200 with 96 megs of RAM. I'm upgrading that machine, it could really use it. But if it was a 800 MHz machine with 256 megs, I'd still be happy with it for a while. In all liklihood it'll get bumped to a 2+ GHz machine with 512M, because what the hell, it's cheap. But I don't need it.
I think that concern has been answered by the nForce series of MB chipsets.
I have only one machine to compare by, but I'd agree with this. I built my most recent machine around an ABit N7F mainboard (built-in VGA, I'm not a gamer and wouldn't even know if my 3D stuff wasn't working). It's the most solid machine I've ever built, and at $105 with built in VGA/Ethernet/USB 2.0x6/etc it's a steal.
FWIW, I've never owned an intel CPU except in laptops where I had no choice, and one old 386sx machine.
The LEAST stable machine was built around a Cyrix chip. What a piece of crap. That was the CPUs fault though.
When I say 56K is cheating, I mean it in terms of it being an analog modem. It's not really an analog modem, it's a hybrid. The modem on one end is sending pure digital down the line and is using the phone company's equipment as their D/A & A/D converters. If you use 56K modems in pure analog mode, you have a 33.6K modem.
ADSL isn't even pretending to be an analog modem, and doesn't belong in this discussion any more than an ethernet card does.
That is about right. Modern modems are running about that baud, though I don't know exactly what the number is.
But you probably mean BITS PER SECOND, which is the speed at which bits are sent over the modem. Baud is something different, and is probably one of the most misused terms around.
I remember people saying that 9600 baud is about as fast as we're ever going to get. Then 28.8K, then 33.6K. Every one of them was claimed to be "just about all we can squeeze out".
OK, well, 33.6K is actually probably about right. 56K modems are cheating by using the analog phone lines as a partially digital system.
I remember when I was working in a shop where we built clones. We got our first cases with 2.5 digits; 2 normal digits with a "1" before them. We had not yet built any machines > 66 MHz.
I was talking to the supplier and said, jokingly, "Hey, these are only good until we get to 200 MHz! Ha, ha!" I really thought it would be a long time before that happened. It was 2 years I think.
But at the time, we had been at 16 to 33 MHz for several years, and the 66 MHz 486's were these funky "speed multiplied" thing that we all though were kind of a gimmick, and it took a lot of hand tweaking to get 50 MHz 486's to run on a synchronous bus (it wasn't strictly an allowed configuration; VESA local bus was only supposed to go 33 MHz - we had to try different cards to find ones that would work).
That depends on the authoring of the DVD. The early Disney DVDs that I have, the menus can NOT be skipped. The recent ones can in any player, and in fact a screen comes up before the previews that says "Press your MENU button to skip the previews."
I've never used a PS2, so I don't know how broken it is, but every other DVD player I've used work as expected on these discs. The Apex I used to have DID let me skip "unskippable" stuff as well, and I know that's not "correct" behavior.
For instance, if you installed a retail version of PhotoShop, your warranty would be void. However, this is not illegal.
It might be. If, for instance, you can show that they market and sell the computer and imply that the computer can run software other than what came with it. If you bought the computer with the expectation that you could run this software, and they encouraged that belief, then they can't then pull your warranty when you follow through.
Otherwise, it'd be like a car dealer not honoring the warranty on the powertrain because you bolted in a fire extinguisher in the trunk.
Well, it would be more like, Juliard spends $10 million designing a new curriculum to teach music in a new way, and then the students finding out that the new coursework was developed by a half dozen guys with 2 year music degrees from a community college in Swaziland.
Likewise, I have made ripped "movie-only" copies of most of the Disney DVDs that we own. Without doing that, you wind up watching many minutes of commercials at the beginning. OK, you can skip them, but why should I have to?
Well, the "real programmer" parent is being pretty naive in regards to Microsoft.
Certainly, he's right, IN THEORY. However, the truth is that people come to RELY ON undocumented behavior in Microsoft APIs. When you do something under the hood that changes one undocumented behavior to another, you stand a chance of breaking things that a programmer wrote, intending to take advantage of that undocumented behavior.
Sure, you can blame the 3rd party programmer for trying to use an undocumented behavior.
But guess what? You can't write serious apps for the Microsoft platform without bumping into undocumented behavior, or behavior that is DIFFERENT than what is documented as "correct."
I work on a mature, very large, vertical market product that runs under Windows. Our programmers sometimes have to spend timeblack-box testing some API to find out how it REALLY works, as opposed to how Microsoft says it's supposed to work.
And guess what? Next service pack, it might just break our code. What is our recourse? Why, to fix OUR BUG, of course. Obviously it's OUR BUG because it'd be silly to claim that MICROSOFT was at fault.
The truth is, Microsoft does regression testing against THEIR *CURRENT* software. You can tell because when their service packs break 3rd party software, it never breaks MS Office. This is what leads customers to think that obviously it's OUR problem not Microsoft's.
Yes, but it's not clear that's what really happened; it appears that the actual info may have been muddled by the reporter. I suggest looking for Krastof in news.google.com and read some other articles.
Here's an excerpt from another article on this matter:
The suspect led the police right to his door when he decided to go online. Gascoyne alerted the police that someone had used his account since the burglary. America Online helped investigators link the dial-up computer connection to a phone number, which SBC then linked to a phone jack at Krastof's home.
This is TOTALLY un-scary. The Wells-Fargo guy apparently has his password cached on the machine. This guy just clicks "login" and logs in AS THE GUY WHOS COMPUTER WAS STOLEN. At this point it's a trivial bit of work to go catch the guy.
Hmm. I wrote a hell of a lot of stuff in it, for all it's being "not a programming language." One was a very full-featured (for its day) BBS for the TRS-80 model I, with a linked-list messaging filesystem including garbage collection, etc, XModem file downloading, and way more features than the leading BBS of the day, which was written for the Apple ][.
My boss when I was in high school wrote his own complete accounting suite and ran his multiple businesses off of it. But if it's not a programming language, I guess that never happened.
we desing products (like the hubble), to be not easily servicable.
The Hubble is INCREDIBLY servicable. The only reason it's hard is that IT'S IN SPACE. This makes it extremely expensive to get to. I don't think that would be alleviated that much by using a robot instead of a person to work on it.
It would probably cost way more to develop a robotics system that could work on something that complex, than to pay the little bit more every flight that it costs to shoot up people. Besides, people can improvise when things go wrong (and they do, almost every flight). With a robot, even a remote-controlled one, you may just not be able to rescue a mission. One lost mission can kill a hell of a lot of savings.
Slashdot Mirror
No. 3db = twice as much power. 10db = perceived as twice as loud.
In addition, the "twice as loud" perception is highly dependent on the type of noise. The 10db = twice as loud rule only works for pure tones at 1KHz. At other frequencies / types of sound, the perception is different.
So 1db is really quite close.
Please note: I'm not talking about a situation where the government built an airport or some such thing near a previously quiet neighborhood. I'm talking about cases where the home-owner knew (or should have known) the conditions prior to purchase.
I've seen the same sort of short-sighted buying in the US in rural areas. I've seen places where people have built new houses a half mile from a livestock farm that has been there for 50 years, then when they finally move in, they discover that when the wind is blowing the right way, there's a smell. Then they try to get zoning changed, or they sue, or some other tactic, to try to get the farm closed. What, you didn't think pigs smelled? Or did you even check to see who your neighbors were?
Hm. I'll take a look at it. However, I like Agent, and for the same money, Agent gives me a fairly capable newsreader as well. I just wish they'd get moving and get 2.0 out, it's got most of what I've been wanting for the last 4 years.
The thread view is sorted by thread and then time, at least in 0.4, and I think it always has been....
Hey, you're right. My bad. I could have sworn it used to be the other way. Sorry for throwing stones at the wrong target on this one.
Instead of backspace, you can try "b" for back a message. It works for me.
No, it doesn't. It just goes to the previous message in the list. it does NOT go to the most recently displayed message. I just tried it. If I "N" or spacebar through a group, hitting the new messages, say, numbers 5, 10,300, and 522, when I hit backspace, it goes back to 300, then to 10, then to 5. The B key in Thunderbird takes you to 521, then 520, then 519. Totally unhelpful.
The spam filter only works well when you mark a ton of messages as "spam" and a ton of messages as "not spam".
This is not at all clear, if true. I have thousands of messages that are marked as not spam. Do you mean I have to toggle them all to spam, then back to not spam? Is that how you mean? That's a pain. I've never seen a bayesian filter that makes you specifically mark things as nonspam. I'm using POPFile, and anything that I didn't tell it was spam, it assumes is nonspam. OK, SpamBayes asks for a set of example hams up front but after that it trains on "it's not spam so it must be ham."
I've tried 0.2 and most recently 0.3. I gave both of them about a week. Both times I went back to using Forte Agent. 0.2 was just not stable for me. 0.3 was stable, never crashed or lost stuff. My big problem is that after a few days of usage, it just started getting horribly slow. Also there are some usability problems that I start out thinking I can live with, but they eventually bug me too much:
- "thread" view is sorted wrong. They group by thread then sort ALPHABETICALLY. Sorry, group then sort BY TIME, at least optionally. Otherwise I've got recent threads at the top, ancient threads at the bottom, and thousands of emails between them. Browsing becomes nearly impossible; recent threads become needles in a haystack.
- No "backspace" when reading emails: both Agent and Thunderbird (and others) allow spacebar and "N"ext message to quickly browse through messages, Agent has a "backspace" key that remembers which messages you've read and backs you up through them. When you're in thread view of a mailing list that generates 100+ emails a day and you have 6 month's archives in the folder, once you leave a message you have almost no hope of finding it again without this feature.
- the spam filter is hopeless. I tagged well over 1000 spams, and it still was getting about 50% false negatives, and even worse, about 20% false positives. I'd pick up 50 emails, have 20 spams in there, it wouldn't ID 10 of the spams, and it would throw 5+ legitimate emails into the spam filter. POPfile got to be almost perfect far before this. Yes, I could use POPfile with Thunderbird, but I was hoping the feature would actually work.
- using it in large binaries groups is completely hopeless, especially on a good server with long retention. I went into a group that had about 300,000 messages on the server, and it just about coughed up a lung. It took it forever to do anything once that was loaded. Also it doesn't even appear to combine all parts of a multipart post into one display item; without this feature, you actually have to LOOK at all 300,000 items; this is ridiculous, other newsreaders have had this important feature for years.
There are other problems, but I've already forgotten them (I switched back to Agent two days ago).
Yeah, I could fix some of these if I wanted, and I did look in to that, but setting up the build environment is fairly involved, and I couldn't fix all of them without spending significant time learning the guts of the system.
I *want* Thunderbird to work, I just can't live with it yet. And I'm afraid some of the things that bug me about it might be fairly hard to fix.
If they're fined, they'll just increase their rates, perhaps they'll add another 10 cents to everyone's "Number portability surcharge". It will cost them nothing.
I'm seriously considering the 800 MHz fanless VIA as a replacement for my workhorse 200 MHz pentium pro. The low power consumption is a big plus, and it would still be plenty of speed for me, especially with a half GB of RAM in it.
I'm looking at building a new machine, and I'm planning on using the slowest chip I can find that saves me any money (IOW, I won't go from 2GHz to 1GHZ to save $5, but I would to save $50).
My workhorse, a headless linux box in the basement, is running on a Pentium Pro 200 with 96 megs of RAM. I'm upgrading that machine, it could really use it. But if it was a 800 MHz machine with 256 megs, I'd still be happy with it for a while. In all liklihood it'll get bumped to a 2+ GHz machine with 512M, because what the hell, it's cheap. But I don't need it.
I think that concern has been answered by the nForce series of MB chipsets.
I have only one machine to compare by, but I'd agree with this. I built my most recent machine around an ABit N7F mainboard (built-in VGA, I'm not a gamer and wouldn't even know if my 3D stuff wasn't working). It's the most solid machine I've ever built, and at $105 with built in VGA/Ethernet/USB 2.0x6/etc it's a steal.
FWIW, I've never owned an intel CPU except in laptops where I had no choice, and one old 386sx machine.
The LEAST stable machine was built around a Cyrix chip. What a piece of crap. That was the CPUs fault though.
When I say 56K is cheating, I mean it in terms of it being an analog modem. It's not really an analog modem, it's a hybrid. The modem on one end is sending pure digital down the line and is using the phone company's equipment as their D/A & A/D converters. If you use 56K modems in pure analog mode, you have a 33.6K modem.
ADSL isn't even pretending to be an analog modem, and doesn't belong in this discussion any more than an ethernet card does.
That is about right. Modern modems are running about that baud, though I don't know exactly what the number is.
But you probably mean BITS PER SECOND, which is the speed at which bits are sent over the modem. Baud is something different, and is probably one of the most misused terms around.
I remember people saying that 9600 baud is about as fast as we're ever going to get. Then 28.8K, then 33.6K. Every one of them was claimed to be "just about all we can squeeze out".
OK, well, 33.6K is actually probably about right. 56K modems are cheating by using the analog phone lines as a partially digital system.
I remember when I was working in a shop where we built clones. We got our first cases with 2.5 digits; 2 normal digits with a "1" before them. We had not yet built any machines > 66 MHz.
I was talking to the supplier and said, jokingly, "Hey, these are only good until we get to 200 MHz! Ha, ha!" I really thought it would be a long time before that happened. It was 2 years I think.
But at the time, we had been at 16 to 33 MHz for several years, and the 66 MHz 486's were these funky "speed multiplied" thing that we all though were kind of a gimmick, and it took a lot of hand tweaking to get 50 MHz 486's to run on a synchronous bus (it wasn't strictly an allowed configuration; VESA local bus was only supposed to go 33 MHz - we had to try different cards to find ones that would work).
That depends on the authoring of the DVD. The early Disney DVDs that I have, the menus can NOT be skipped. The recent ones can in any player, and in fact a screen comes up before the previews that says "Press your MENU button to skip the previews."
I've never used a PS2, so I don't know how broken it is, but every other DVD player I've used work as expected on these discs. The Apex I used to have DID let me skip "unskippable" stuff as well, and I know that's not "correct" behavior.
For instance, if you installed a retail version of PhotoShop, your warranty would be void. However, this is not illegal.
It might be. If, for instance, you can show that they market and sell the computer and imply that the computer can run software other than what came with it. If you bought the computer with the expectation that you could run this software, and they encouraged that belief, then they can't then pull your warranty when you follow through.
Otherwise, it'd be like a car dealer not honoring the warranty on the powertrain because you bolted in a fire extinguisher in the trunk.
Well, it would be more like, Juliard spends $10 million designing a new curriculum to teach music in a new way, and then the students finding out that the new coursework was developed by a half dozen guys with 2 year music degrees from a community college in Swaziland.
I hope you don't mean it will come up in the trial. It's irrelevant.
Likewise, I have made ripped "movie-only" copies of most of the Disney DVDs that we own. Without doing that, you wind up watching many minutes of commercials at the beginning. OK, you can skip them, but why should I have to?
Well, the "real programmer" parent is being pretty naive in regards to Microsoft.
Certainly, he's right, IN THEORY. However, the truth is that people come to RELY ON undocumented behavior in Microsoft APIs. When you do something under the hood that changes one undocumented behavior to another, you stand a chance of breaking things that a programmer wrote, intending to take advantage of that undocumented behavior.
Sure, you can blame the 3rd party programmer for trying to use an undocumented behavior.
But guess what? You can't write serious apps for the Microsoft platform without bumping into undocumented behavior, or behavior that is DIFFERENT than what is documented as "correct."
I work on a mature, very large, vertical market product that runs under Windows. Our programmers sometimes have to spend timeblack-box testing some API to find out how it REALLY works, as opposed to how Microsoft says it's supposed to work.
And guess what? Next service pack, it might just break our code. What is our recourse? Why, to fix OUR BUG, of course. Obviously it's OUR BUG because it'd be silly to claim that MICROSOFT was at fault.
The truth is, Microsoft does regression testing against THEIR *CURRENT* software. You can tell because when their service packs break 3rd party software, it never breaks MS Office. This is what leads customers to think that obviously it's OUR problem not Microsoft's.
Here's an excerpt from another article on this matter:
This is TOTALLY un-scary. The Wells-Fargo guy apparently has his password cached on the machine. This guy just clicks "login" and logs in AS THE GUY WHOS COMPUTER WAS STOLEN. At this point it's a trivial bit of work to go catch the guy.
According to Bill Gates, as reported in USA Today, Microsoft was never in talks with Google about an acquisition.
And there are no US troops in Baghdad.
Hmm. I wrote a hell of a lot of stuff in it, for all it's being "not a programming language." One was a very full-featured (for its day) BBS for the TRS-80 model I, with a linked-list messaging filesystem including garbage collection, etc, XModem file downloading, and way more features than the leading BBS of the day, which was written for the Apple ][.
My boss when I was in high school wrote his own complete accounting suite and ran his multiple businesses off of it. But if it's not a programming language, I guess that never happened.
How is pointing out the actual facts flamebait? Oh well.
we desing products (like the hubble), to be not easily servicable.
The Hubble is INCREDIBLY servicable. The only reason it's hard is that IT'S IN SPACE. This makes it extremely expensive to get to. I don't think that would be alleviated that much by using a robot instead of a person to work on it.
It would probably cost way more to develop a robotics system that could work on something that complex, than to pay the little bit more every flight that it costs to shoot up people. Besides, people can improvise when things go wrong (and they do, almost every flight). With a robot, even a remote-controlled one, you may just not be able to rescue a mission. One lost mission can kill a hell of a lot of savings.
We REALLY should put scopes on the moon.
Right, I mean after that. I'm assuming that at least the next servicing mission happens. I'd like more.