Depends on the pixel art in question. ZSNES comes with a variety of scalers it can use on old Super Nintendo games. The pixel-duplication scaler will always look good, but sometimes one of the others (usually the HQ2x filter) will give results that look better than the original art.
Which brand/model of monitor are you using? I've got a Viewsonic VA2026w next to me, and at any resolutions besides the native 1680x1050 and certain "common" resolutions (eg. 1024x768), it varies between "ugly" and "unusably blurred".
Maybe you should teach them this and test it on Drivers License exam instead of usual round the block with automatic transmission then, eh?
Do you know how many multilane roads there are with all-way stop signs? I know of one in Grant Village in Yellowstone, but other than that, I don't think there are any within 500 miles of me. Does it really make sense to teach people how to handle extremely rare events?
You work your way up, just like you do in the current model. You might release your first works for free, to get an audience. Then, as your audience grows, you increase the cost accordingly.
For expensive things like movies, you'd need a sponsor for your initial works. Imagine 20th Century Fox giving George Lucas $10 million to make A New Hope and release it for free, on the understanding that if it's successful, he'll make two sequels. Half the design price of the sequels would go to to 20th Century Fox, with the other half being split between the cost of making the sequels and GL's paycheck.
The bulk of the swap will only be used during a few compiles (glibc and gcc come to mind). For normal operation, you won't be using more than a megabyte or two of it.
You can't install Gentoo unassisted on that machine: you can't fit both the Portage tree and the compile environment into 720MB, and 28MB of RAM requires the use of a great deal of swap. I'd do the following:
1) Partition the hard drive into a 250MB swap partition, a 20MB bootstrap partition, and a 450MB system partition 2) Install a floppy-era Linux on the bootstrap partition. 3) Using the bootstrap Linux to give you network access, mount network drives for/usr/portage,/var/tmp, and/tmp. 4) Install Gentoo using gcc-3.4 and an appropriately old glibc. 5) Install a lightweight DE and apps.
Steps 4 and 5 will take you about a week on the hardware mentioned. After that, routine upgrades will take no more than an hour or so.
The pace of the books just kept slowing down. My impression was that he could have wrapped things up in book 1 with another hundred pages or so. Book 2 could have been the second book in a trilogy. I gave up after book 5, when it became clear that he wasn't going to manage a septology.
The mandatory car analogy: I have the right to give you a ride from the bank to your house. But if I do so with the knowlege that you've just robbed said bank, I can be charged with "accessory after the fact to bank robbery".
Laws are as much about context and intention as they are about actions.
Just as an aside, the Digium SwitchVox platform, which is our commercial re-packaging of Asterisk, has as an element of it's GUI a tool that indicates the relative strength of passwords. We'd encourage any other re-packagers or users of Asterisk to implement a similar UI hint that forces good password behavior by users and local admins.
There's no good algorithm for telling the strength of a password (password strength is related to the Kolmogorov complexity of the password, which is incomputable), and every password-strength indicator I've seen uses heuristics that either accept weak passwords ("Password1" is strong because it's nine characters, a mix of upper- and lower-case, and has a non-letter) or rejects strong passwords ("this password is very very strong even though it only contains lowercase letters because it is a long password and plain english has between one and two bits of randomness per letter").
I've run Linux servers on variously an Intel Atom (70 watts, including 5 hard drives and a RAID card), a Celeron 433 (40 watts, including two quad-port Ethernet cards), a Pentium MMX 233 (35 watts), and a NSLU2 (8 watts). If you need CPU power, the Atom is probably your best bet; if you're just shoving static files around, the NSLU2 will work just fine.
That's at the high end. But what if you only need 4GB? A solid-state drive will run you about $3.35/GB. Rotating media? Can't do that -- the smallest you can get is an 80GB drive for $35, which works out to $8.75 per gigabyte of used storage.
(Before you object to my choice of solid-state drives, please note that the CompactFlash interface is parallel ATA with a smaller pin spacing, and adapters cost next to nothing.)
Your point #4 is wrong: Win 3.1 can access a FAT32 filesystem if the underlying version of DOS can. You won't get long filenames, but you can access the files themselves.
3. Peer-to-peer networks have gotten more efficient over the years. Early implementations of Gnutilla, for example, generated something like 75% of the Internet's traffic just holding itself together, about another 5% moving search requests and results around, and 1% transferring data.
I prefer the US solution: if you want to pay your taxes in pennies, you can -- but you are the one who needs to count it and prove it's the right amount.
Yes, something doesn't add up about Google's numbers: the error rates the study reports work out to 1600-5000 single-bit errors per gigabyte per year. I ran a small university computer lab for two years. The computers had parity (not ECC) memory, and the OS would panic on a single-bit error. Based on the crash rate, I figured the error rate was around 15 single-bit errors per gigabyte per year.
Likewise, it doesn't match the results people are getting from Memtest86. According to Google's numbers, you'd expect between 4 and 13 errors per gigabyte of memory in a single 24-hour run, but people almost always report error-free runs.
I worked out the numbers for my Honda Civic a year or two ago: assuming a solar array 14 feet long by 5 feet wide, on a tiltable (but not steerable) frame, I could drive between 15 (winter) and 30 (summer) miles on a day's sunlight.
If antivirus protectors could collect data from machines and users
This idea stopped being a good one here.
Think about a corporate environment where this level of information is readily available: if your automated system can spot a virus working its way through the PHBs, the system could block it before it gets to Accounting and starts interfering with people who actually do work.
They produce ads that are so embarrassing that the message is really clear: "oh, did you see the lame ad, these guys should stick to doing what they do best, which is programming. Win 7 is actually quite good you know."
No, the message they're sending is "not only does our software blow dead goats, we can't hide it with fancy marketing".
Slashdot Mirror
Depends on the pixel art in question. ZSNES comes with a variety of scalers it can use on old Super Nintendo games. The pixel-duplication scaler will always look good, but sometimes one of the others (usually the HQ2x filter) will give results that look better than the original art.
Which brand/model of monitor are you using? I've got a Viewsonic VA2026w next to me, and at any resolutions besides the native 1680x1050 and certain "common" resolutions (eg. 1024x768), it varies between "ugly" and "unusably blurred".
You want a laugh? Ask for the IBM share price in zimbabwean dollars -- apparently the low price was -Z$47.51, about four months ago.
Do you know how many multilane roads there are with all-way stop signs? I know of one in Grant Village in Yellowstone, but other than that, I don't think there are any within 500 miles of me. Does it really make sense to teach people how to handle extremely rare events?
You work your way up, just like you do in the current model. You might release your first works for free, to get an audience. Then, as your audience grows, you increase the cost accordingly.
For expensive things like movies, you'd need a sponsor for your initial works. Imagine 20th Century Fox giving George Lucas $10 million to make A New Hope and release it for free, on the understanding that if it's successful, he'll make two sequels. Half the design price of the sequels would go to to 20th Century Fox, with the other half being split between the cost of making the sequels and GL's paycheck.
The bulk of the swap will only be used during a few compiles (glibc and gcc come to mind). For normal operation, you won't be using more than a megabyte or two of it.
Might it have something to do with the perpetual budget crisis brough on by Proposition 13?
You can't install Gentoo unassisted on that machine: you can't fit both the Portage tree and the compile environment into 720MB, and 28MB of RAM requires the use of a great deal of swap. I'd do the following:
1) Partition the hard drive into a 250MB swap partition, a 20MB bootstrap partition, and a 450MB system partition /usr/portage, /var/tmp, and /tmp.
2) Install a floppy-era Linux on the bootstrap partition.
3) Using the bootstrap Linux to give you network access, mount network drives for
4) Install Gentoo using gcc-3.4 and an appropriately old glibc.
5) Install a lightweight DE and apps.
Steps 4 and 5 will take you about a week on the hardware mentioned. After that, routine upgrades will take no more than an hour or so.
The pace of the books just kept slowing down. My impression was that he could have wrapped things up in book 1 with another hundred pages or so. Book 2 could have been the second book in a trilogy. I gave up after book 5, when it became clear that he wasn't going to manage a septology.
The mandatory car analogy: I have the right to give you a ride from the bank to your house. But if I do so with the knowlege that you've just robbed said bank, I can be charged with "accessory after the fact to bank robbery".
Laws are as much about context and intention as they are about actions.
Esperanto? No, Latin is the One True Lingua Franca!
There's no good algorithm for telling the strength of a password (password strength is related to the Kolmogorov complexity of the password, which is incomputable), and every password-strength indicator I've seen uses heuristics that either accept weak passwords ("Password1" is strong because it's nine characters, a mix of upper- and lower-case, and has a non-letter) or rejects strong passwords ("this password is very very strong even though it only contains lowercase letters because it is a long password and plain english has between one and two bits of randomness per letter").
I've run Linux servers on variously an Intel Atom (70 watts, including 5 hard drives and a RAID card), a Celeron 433 (40 watts, including two quad-port Ethernet cards), a Pentium MMX 233 (35 watts), and a NSLU2 (8 watts). If you need CPU power, the Atom is probably your best bet; if you're just shoving static files around, the NSLU2 will work just fine.
That's at the high end. But what if you only need 4GB? A solid-state drive will run you about $3.35/GB. Rotating media? Can't do that -- the smallest you can get is an 80GB drive for $35, which works out to $8.75 per gigabyte of used storage.
(Before you object to my choice of solid-state drives, please note that the CompactFlash interface is parallel ATA with a smaller pin spacing, and adapters cost next to nothing.)
Your point #4 is wrong: Win 3.1 can access a FAT32 filesystem if the underlying version of DOS can. You won't get long filenames, but you can access the files themselves.
3. Peer-to-peer networks have gotten more efficient over the years. Early implementations of Gnutilla, for example, generated something like 75% of the Internet's traffic just holding itself together, about another 5% moving search requests and results around, and 1% transferring data.
Why not do both?
The article sounds like a mashup of the many-worlds interpretation of quantum mechanics, the transactional interpretation, and quantum immortality, as filtered through pop-science journalism.
I prefer the US solution: if you want to pay your taxes in pennies, you can -- but you are the one who needs to count it and prove it's the right amount.
That's why you were having trouble: ECC memory is almost invariably buffered.
Yes, something doesn't add up about Google's numbers: the error rates the study reports work out to 1600-5000 single-bit errors per gigabyte per year. I ran a small university computer lab for two years. The computers had parity (not ECC) memory, and the OS would panic on a single-bit error. Based on the crash rate, I figured the error rate was around 15 single-bit errors per gigabyte per year.
Likewise, it doesn't match the results people are getting from Memtest86. According to Google's numbers, you'd expect between 4 and 13 errors per gigabyte of memory in a single 24-hour run, but people almost always report error-free runs.
I worked out the numbers for my Honda Civic a year or two ago: assuming a solar array 14 feet long by 5 feet wide, on a tiltable (but not steerable) frame, I could drive between 15 (winter) and 30 (summer) miles on a day's sunlight.
Think about a corporate environment where this level of information is readily available: if your automated system can spot a virus working its way through the PHBs, the system could block it before it gets to Accounting and starts interfering with people who actually do work.
The plastic lining in aluminum cans is because the acid in most drinks will slowly dissolve the aluminum, giving the drink a metallic flavor.
No, the message they're sending is "not only does our software blow dead goats, we can't hide it with fancy marketing".