Nice rant. I even read about the first half of it before I got bored.
So hey, listen: It's my country, too. And I'd welcome another 20 million heads, if they're immigrated properly. Then, they can enjoy all of the prosperity of the US, and pay taxes just like the rest of us.
Of course, current immigration laws can make this hard for an individual to accomplish. So I'd much rather see a greater focus on making legal immigration easier, than on making illegal immigration harder, especially since efforts toward the latter tend to make my country seem less free for me. But if immigrating were easy enough, it won't be worth the trouble and the risk to come over here illegally.
(And if you're thinking "Preposterous! You'd be opening the flood gates!", and preparing a lengthy rant about that: I know. I'm OK with that. The US was founded as a diverse nation of immigrants, and I remember where we came from.)
Lithium batteries don't suffer from "the memory effect."
They perform best (as in: most usefully, and most economically in terms of useful lifespan) if you follow a "charge early, charge often" method. They've got enough smarts to keep you from hurting them by charging them too long, or draining them too dry, or whatever. But purposefully discharging them all the way, only to purposefully charge them all the way back up again does far more damage to the battery than, you know, just using it in a convenient fashion and forgetting all of that nonsense.
Besides, they only last a handful of years anyway, no matter how you treat them -- even if you leave them on a shelf, never used, they're still slowly failing.
So, in synopsis: Charging lithium batteries shortens their lifespan. Discharging lithium batteries shortens their lifespan. Doing either of these in any extreme shortens the lifespan even more. And leaving then alone in a shoebox shortens their lifespan.
Get your head out of the 80's. Memory effect (if it ever really was a problem) only existed with NiCd batteries, and you won't find these in any phone made in at least the last decade or so.
So if you think she'll like a N900, then just buy her the damned phone. The battery will only last a few years no matter what, so when it's tired, just buy a new battery, you cheap bastard.
I had one in my locally-built 10MHz XT, along with CGA, 5.25" and 3.5" low-density floppies, a battery-backed real-time clock card, and a dedicated game port card.
I did, at one point, run Windows on that thing. It was a stupid mistake that I never bothered to repeat.
I agree -- simple, easy to understand units are just dumb. This is why, when expressing fuel economy, I prefer to use Hogshead/League. I think that speed is best expressed in terms of rods per Káshthá.
I also prefer to express time in fractional fortnights. And I like to measure energy consumption in terms of calories/hour instead of Watts, though I find it much easier to review my past usage on my electric bill if I convert from kilowatt hours to horsepower hours.
Hmm. Go back and read what I wrote. If you still find something disagreeable about it, read it again. And then, if you still can't figure out how not to disagree, just go away.
Once, during the original Pentium counterfeiting spree (where they'd grind down the ceramic top and re-screen the labels), I was at the Dayton Hamvention.
There was a vendor there that I knew, and trusted, and had bought RAM and CPUs and other commodity silicon from in the past, all of which worked fine.
He was selling counterfeit Pentiums. But they were marked as such. They were cheaper than the genuine product, and he was happy to explain to people what had happened to them. Stuff like "These were all probably Pentium 100's, but they've all checked out at 133, and now they're just marked as such."
And: "We've got real P133's here for you to buy, too."
And he'd explain the difference, and how to tell (the ceramic wasn't as smooth as the real article, for instance).
I stood and watched people at that booth for some time: Some folks, knowingly, bought the counterfeits. Some folks bought the genuine ones. Nobody really seemed to be scared away by the concept or the vendor, though, probably because of his openness, honesty, and willingness to teach.
In retrospect, it looked a whole lot more like the retail overclocking scene does today, than it did of someone trying to cheat someone else.
You're so stupid, I feel ashamed to have a UID close to yours. But: mine's still lower.
So listen, kid:
The GGP used a bad layout, and people complained. That's all there is to say.
Meanwhile, screen resolutions have moved from 640x480 to 800x600 to 1024x768 to, nowadays, 1920x1200. Except for portable devices like the iPhone, at 320x480. Or the various Android devices, at their various resolutions. Or any of the other competently-useful Web-browsing devices folks have in their pockets these days.
So, plainly: Pixels mean nothing. That GGP couldn't figure it out to the satisfaction of his customers (who, if they were complaining, probably don't even know what a pixel IS) means that he's incompetent at the task, not that the idea is unsound.
Systems shouldn't crash, anyway. If they do crash, they need fixing.
I ran software RAID on a busy Linux box for five years -- it was finally decommissioned yesterday. It never crashed. It got rebooted a few times, once due to a tornado, another for more RAM, another for a rare remote kernel exploit. It was never shut down improperly.
And, yes, this "RAID controller" was battery-backed.
He's only got physical room for three access points (or, more likely: three clusters of access points).
There are three non-overlapping channels. So, 3 channels * 3 clusters = 9 APs, maximum. But 9 APs * 50 users each 500.
It's important to remember that 50 is just a useful number, and needn't be a hard limit: It's not spelled out in 802.11 that "there shall not be more than 50 users per node."
And besides, it's not like folks are going to physically locate themselves for optimum WiFi distribution. They'll be wherever they are. There's always going to be more than, or less than, 50 users per access point eventually, no matter what. Nevermind the fact that even running in a somewhat degraded state, these 9 independent access points are NOT going to be the bottleneck -- the backbone to Teh Intarwebs will be.
So:
1. Use three clusters of three WRT54GL (the L suffix is important, in case the asker doesn't know -- Google it), all running OpenWRT or Tomato or somesuch freewheeling firmware.
2. Turn the power down. Lots. More power means that the contention issues of 802.11 get worse, not better. A bit of low-tech experimentation with a couple of devices (any old laptop, a jailbroke iPod Touch, an Android phone, etc) will help find the sweet spot for power in that environment.
Plucker on such a device doesn't do any better than it does on my old Handspring Visor Deluxe.
Sure, the Visor is monochrome. But so are books. And the resolution of PalmOS devices sucks, just some less than others.
Not that Plucker isn't very good at what it does. It very much is.
Lately, when I read ebooks away from the computer, I've been doing it on a first-gen iPod Touch. Works good, and was free for me, thanks to a silly rebate program from Netgear. Its display also gets dim enough to be useful in a dark room, unlike my Droid, or the Zire 71 I used to have (which died during battery replacement surgery).
There may be, for example, 64KB pages, containing many packets. None of the packets can be decoded until the entire 64KB page is received and its CRC checked. This may sound small, but for 32-64kbit stream, that's 10 seconds of latency right there. Alternatively, you can have 1 page per packet, but on 32-64kbit streams you end up with about 5-10% overhead from the container. It is a REAL problem.
So picture it:
You have a widget with knobs on it. Stuff goes into it, and other stuff comes out depending on how the knobs are set.
You look at this widget and say to yourself, "Self, how can I make this widget do something bad -- something against the end-result that I'm attempting to achieve?"
You notice that one of the knobs is labeled "Ouch" and another is labeled "Masochism." You turn the Ouch knob all the way up to "Fuck you, buddy!" and the "Masochism" knob to the "I like the hot poker in the eye!" setting.
And, then, you have the gall to be surprised when the widget produces the badness that you asked it to?
Pick reasonable tradeoffs, and you'll have a reasonable result.
This was even several months before Netscape released the source code for their browser, which was the event that made Firefox possible in the first place.
Even Microsoft had the beginning of useful PNG support in Internet Explorer starting with version 4.0, released in September of 1997.
Now, sure: As I recall, Firefox was way before IE in fully supporting all of the features of PNG. But to say that it supported PNG before IE, when IE supported PNG before Firefox could even have begun to exist, is a little anachronistic.
Slashdot Mirror
We're Americans, here. We only get one holiday, and it only happens at one predetermined point of the year. You insensitive clod.
Nice rant. I even read about the first half of it before I got bored.
So hey, listen: It's my country, too. And I'd welcome another 20 million heads, if they're immigrated properly. Then, they can enjoy all of the prosperity of the US, and pay taxes just like the rest of us.
Of course, current immigration laws can make this hard for an individual to accomplish. So I'd much rather see a greater focus on making legal immigration easier, than on making illegal immigration harder, especially since efforts toward the latter tend to make my country seem less free for me. But if immigrating were easy enough, it won't be worth the trouble and the risk to come over here illegally.
(And if you're thinking "Preposterous! You'd be opening the flood gates!", and preparing a lengthy rant about that: I know. I'm OK with that. The US was founded as a diverse nation of immigrants, and I remember where we came from.)
Lithium batteries don't suffer from "the memory effect."
They perform best (as in: most usefully, and most economically in terms of useful lifespan) if you follow a "charge early, charge often" method. They've got enough smarts to keep you from hurting them by charging them too long, or draining them too dry, or whatever. But purposefully discharging them all the way, only to purposefully charge them all the way back up again does far more damage to the battery than, you know, just using it in a convenient fashion and forgetting all of that nonsense.
Besides, they only last a handful of years anyway, no matter how you treat them -- even if you leave them on a shelf, never used, they're still slowly failing.
So, in synopsis: Charging lithium batteries shortens their lifespan. Discharging lithium batteries shortens their lifespan. Doing either of these in any extreme shortens the lifespan even more. And leaving then alone in a shoebox shortens their lifespan.
Get your head out of the 80's. Memory effect (if it ever really was a problem) only existed with NiCd batteries, and you won't find these in any phone made in at least the last decade or so.
So if you think she'll like a N900, then just buy her the damned phone. The battery will only last a few years no matter what, so when it's tired, just buy a new battery, you cheap bastard.
Hmm. So much for my reasonably-quick, inexpensive ground shipping for Newegg orders from Edison, NJ to Ohio.
*sigh*
As a back-end guy, I don't want it.
Happy?
That was an ST-225.
I had one in my locally-built 10MHz XT, along with CGA, 5.25" and 3.5" low-density floppies, a battery-backed real-time clock card, and a dedicated game port card.
I did, at one point, run Windows on that thing. It was a stupid mistake that I never bothered to repeat.
Spoken like someone with true sanity about them.
Unfortunately, not everyone is so blessed.
I agree -- simple, easy to understand units are just dumb. This is why, when expressing fuel economy, I prefer to use Hogshead/League. I think that speed is best expressed in terms of rods per Káshthá.
I also prefer to express time in fractional fortnights. And I like to measure energy consumption in terms of calories/hour instead of Watts, though I find it much easier to review my past usage on my electric bill if I convert from kilowatt hours to horsepower hours.
Simple, meaningful units are for the birds.
They have a Linux version, right?
Yes.
Plainly, as with any other multitasking system, such problems depend on the apps.
You've just chosen better-behaved applications than GP has.
Hmm. Go back and read what I wrote. If you still find something disagreeable about it, read it again. And then, if you still can't figure out how not to disagree, just go away.
Thank you.
Once, during the original Pentium counterfeiting spree (where they'd grind down the ceramic top and re-screen the labels), I was at the Dayton Hamvention.
There was a vendor there that I knew, and trusted, and had bought RAM and CPUs and other commodity silicon from in the past, all of which worked fine.
He was selling counterfeit Pentiums. But they were marked as such. They were cheaper than the genuine product, and he was happy to explain to people what had happened to them. Stuff like "These were all probably Pentium 100's, but they've all checked out at 133, and now they're just marked as such."
And: "We've got real P133's here for you to buy, too."
And he'd explain the difference, and how to tell (the ceramic wasn't as smooth as the real article, for instance).
I stood and watched people at that booth for some time: Some folks, knowingly, bought the counterfeits. Some folks bought the genuine ones. Nobody really seemed to be scared away by the concept or the vendor, though, probably because of his openness, honesty, and willingness to teach.
In retrospect, it looked a whole lot more like the retail overclocking scene does today, than it did of someone trying to cheat someone else.
guarantee
There's lots of guarantees out there. Craftsman hand tools come immediately to mind -- if it breaks, I take it back and get a new one.
Of course, they're shoddy tools, and break all the time, but at least I can get a new one when it happens.
A guarantee isn't a certainty; it's a fallback.
You're so stupid, I feel ashamed to have a UID close to yours. But: mine's still lower.
So listen, kid:
The GGP used a bad layout, and people complained. That's all there is to say.
Meanwhile, screen resolutions have moved from 640x480 to 800x600 to 1024x768 to, nowadays, 1920x1200. Except for portable devices like the iPhone, at 320x480. Or the various Android devices, at their various resolutions. Or any of the other competently-useful Web-browsing devices folks have in their pockets these days.
So, plainly: Pixels mean nothing. That GGP couldn't figure it out to the satisfaction of his customers (who, if they were complaining, probably don't even know what a pixel IS) means that he's incompetent at the task, not that the idea is unsound.
So, battery-back your system...obviously.
Systems shouldn't crash, anyway. If they do crash, they need fixing.
I ran software RAID on a busy Linux box for five years -- it was finally decommissioned yesterday. It never crashed. It got rebooted a few times, once due to a tornado, another for more RAM, another for a rare remote kernel exploit. It was never shut down improperly.
And, yes, this "RAID controller" was battery-backed.
FWIW.
Increasingly, it's not good enough that you said what you did say, and chose not to say what you clearly haven't said.
And, hey, guess what? No one said otherwise.
Well, sure I have. Just like any other self-respecting geek.
But we're a pretty small cross-section, and this task seems to involve a more generalized crowd.
:)
He's only got physical room for three access points (or, more likely: three clusters of access points).
There are three non-overlapping channels. So, 3 channels * 3 clusters = 9 APs, maximum. But 9 APs * 50 users each 500.
It's important to remember that 50 is just a useful number, and needn't be a hard limit: It's not spelled out in 802.11 that "there shall not be more than 50 users per node."
And besides, it's not like folks are going to physically locate themselves for optimum WiFi distribution. They'll be wherever they are. There's always going to be more than, or less than, 50 users per access point eventually, no matter what. Nevermind the fact that even running in a somewhat degraded state, these 9 independent access points are NOT going to be the bottleneck -- the backbone to Teh Intarwebs will be.
So:
1. Use three clusters of three WRT54GL (the L suffix is important, in case the asker doesn't know -- Google it), all running OpenWRT or Tomato or somesuch freewheeling firmware.
2. Turn the power down. Lots. More power means that the contention issues of 802.11 get worse, not better. A bit of low-tech experimentation with a couple of devices (any old laptop, a jailbroke iPod Touch, an Android phone, etc) will help find the sweet spot for power in that environment.
3. ???
4. Profit!
So, yes, I do what you do and much more.
More? Feh. I've got similar stories. Anyone does, if they're doing a decent job of fixing things.
I was just being concise when I said that I explain things and address concerns. I didn't realize it was story hour.
There's no problem using an RLL drive on an MFM controller, if that's the combination that happens to be available.
Plucker on such a device doesn't do any better than it does on my old Handspring Visor Deluxe.
Sure, the Visor is monochrome. But so are books. And the resolution of PalmOS devices sucks, just some less than others.
Not that Plucker isn't very good at what it does. It very much is.
Lately, when I read ebooks away from the computer, I've been doing it on a first-gen iPod Touch. Works good, and was free for me, thanks to a silly rebate program from Netgear. Its display also gets dim enough to be useful in a dark room, unlike my Droid, or the Zire 71 I used to have (which died during battery replacement surgery).
I wasn't arguing the GP's point for him. I was arguing against yours.
There may be, for example, 64KB pages, containing many packets. None of the packets can be decoded until the entire 64KB page is received and its CRC checked. This may sound small, but for 32-64kbit stream, that's 10 seconds of latency right there. Alternatively, you can have 1 page per packet, but on 32-64kbit streams you end up with about 5-10% overhead from the container. It is a REAL problem.
So picture it:
You have a widget with knobs on it. Stuff goes into it, and other stuff comes out depending on how the knobs are set.
You look at this widget and say to yourself, "Self, how can I make this widget do something bad -- something against the end-result that I'm attempting to achieve?"
You notice that one of the knobs is labeled "Ouch" and another is labeled "Masochism." You turn the Ouch knob all the way up to "Fuck you, buddy!" and the "Masochism" knob to the "I like the hot poker in the eye!" setting.
And, then, you have the gall to be surprised when the widget produces the badness that you asked it to?
Pick reasonable tradeoffs, and you'll have a reasonable result.
Netscape Navigator had useful support for PNG starting in November, 1997 with version 4.04. This beat even the existence of Firefox 1.0 by seven entire years.
This was even several months before Netscape released the source code for their browser, which was the event that made Firefox possible in the first place.
Even Microsoft had the beginning of useful PNG support in Internet Explorer starting with version 4.0, released in September of 1997.
Now, sure: As I recall, Firefox was way before IE in fully supporting all of the features of PNG. But to say that it supported PNG before IE, when IE supported PNG before Firefox could even have begun to exist, is a little anachronistic.