I looked into writing a JPEG2000 pixbuf loader for GTK2 once. The only open-source-ish implementation I could find was JasPer, which had too many licensing requirements to be useful. I tried to find the specifications, but not even the baseline specs are free. So I said "Screw this" and abandoned the idea of ever supporting JPEG2000 in anything, and I suspect I'm not alone.
"As for the P4/1.8GHz story try this for a test : Install MySQL on your linux PC and create a database with a table of about 5-6GB. Run alter table on it. Wait for it CRUMBLE TO DUST as it hits past 2GBs. Then get a Sun."
No way would I ever use a toy like MySQL if I had the option of Oracle, DB2, or even PostgreSQL. News flash for you: real software uses the 64-bit I/O calls.
Pretty sure it was The Trees, but it could've just as well been Freewill or A Passage to Bangkok. Sometime in my very last days of high school (1999) over a 14.4 modem (1996). Indeed, it took my parents 3 years to get around to getting online.
I don't have it anymore, as my system wasn't purchased until a year and a half later, and the 1GB drive in the old box got overcrowded in the meantime.
Those instances of Slashdotter are not necessarily the same person. Furthermore, Microsoft would be totally braindead to make a business plan from Slashdot comments. BillG probably isn't that dumb, despite what we wish for;-)
Yep, not even the Color-I Plus. No power LED for us!
My dad's had to repair it a couple times, but we still run it for video gaming. Man, that thing's had more stuff connected to it....
2 C= 64s*
C= Plus 4
2 C= Amiga 500s** (running OS 1.2 and 1.3...)
NES
Super NES
Sony PlayStation
Currently, a PS2
And the occasional VCR being repaired for friends
* I believe the Ohio Scientific with a huge 8K RAM used a different monitor, and the C= 64 was the original reason this one was purchased. But I'm too young to remember anything before the Amigas very well.
** To run the Amigas, my dad built a custom cable and added a plug to the monitor to hook the Amiga RGB output up more-or-less directly to the electron guns.
"Instead of sending the whole email content - and with it the ability to falsify email header information, why not just send the email header only - and require the originating server to hold the email content?"
I thought about this once. How do you ensure that the person coming back to get the mail off the originating server is indeed the person to whom the mail was sent? I don't think any certificate-based scheme is the answer, because key management is so complicated. At least in their current incarnations. (If there was something as simple as ssh-keygen that also posted the public key to the ISP, it might be workable. But there isn't, and there's the question of mail client support.)
Regarding the "It shows your IP address" that mabu pointed out, wouldn't it be trivial to proxy the "get the content" message through the ISP's MTA?
Required in what market? I should think that if there was such a demand, instead of chips putting out 70W, we would see VIA C3 or Transmeta-based offerings taking over the world.
Ever see an Evergreen 286-to-486 upgrade chip? A board with a 486 and some cache on it, and a 286-compatible set of pins underneath. You had to run a little program to enable the cache or it would perform worse than a 286, but once it was all set up, it made the fastest 286 on the block....
According to Scott Mueller's Upgrading and Repairing PCs, the early SX's were DX's with the FPU mysteriously disabled. Later SX's really had no FPU at all.
And in another strange twist, a "487" was a full-blown 486DX with a different pin layout that disabled the main processor.
I did something similar. Some version of Oracle and my scanner software conflicted, and uninstalling Oracle didn't replace the DLLs... and neither did a non-destructive reinstall of Windows. So I switched to Linux (it was a dual-boot box), and figured I'd scribble over the FAT and Windows would actually format it. But rather than hde1, I put the plain hde in the dd command.
NO OPERATING SYSTEM INSTALLED INSERT SYSTEM DISK AND PRESS ENTER
Windows 98SE will cheerfully install on a system with no partition table, but it cannot be booted afterwards. Thank God for freshmeat! Through them I found a static gpart (guess partitions) and restored everything... by sneakernetting it to another box, because no Linux tools would put the partition boundaries out in the middle of tracks! After that, I use fdisk and make my own partitions, and print the table when I'm done.
Nah, the default is whatever the sysadmin set up in/etc/skel/$SHELL_FILE. The real default looks like "%" for csh, or "bash-2.05b$ " for bash. I think both change to # for a root prompt.
Random note: bash uses its argv[0] in the prompt, so it's possible to see "init-2.05b# " if it's a really bad day....
Holy shit. If this is true, Python mops the floor with Perl on arbitrary precision. I have a Math::BigInt perl process doing only the long test which has been going on for 77:51 (min:sec) as I post this, whereas the entire benchmark in Python was over in 37:19. (I've constrained the accuracy/precision to ~67-bit range in Perl. It's still taking forever.)
OTOH, my Python results are somewhat dubious, since it comes up with 10 billion instead of ~777 million for that test, claiming -2940 seconds for it. Subtracting the other times, it really took 23:02.
So if a 5-year-old CPU can handle any task, why do they keep building these supercomputer clusters like the Earth Simulator or the G5 cluster? Why do we have video cards if a CPU can do everything?
The reason CPUs are fast enough to handle tasks we do today is that we're not throwing intensive problems at them. Imagine dispatching an AI to search your images by voice, then continuing a refreshing game of Final Fantasy Tactics on an emulator: the beefiest SMP box with multi-channel memory and RAID 10 would be screaming in pain.
I didn't see any mention of the numerical results obtained being checked against what they ought to be (or even against each other).
That last bit's pretty important. For me, Java and C produce 776627965 as the long result, whereas Python is giving me 10000000000. (And it's not some braindead failure like printing the wrong variable.) And Python came up with that result in -2940 seconds...
Dude! My MBR is only 512 bytes, and part of that is the primary partition table and boot signature, leaving (IIRC) 448 bytes for GRUB stage1. Stage 1's responsibility is loading the embedded stage 2 from the following 16 sectors of disk, which falls in a mysterious "reserved"[1] area (according to FreeBSD's installer), and stage 2 can actually understand the filesystem on the partition where its config is to actually do something useful.
x86 is vile.
[1] That reserved area was 31 sectors on my old 15GB IBM DTLA, and 63 sectors on my new 80GB Seagate.
Doubtful. If we are to believe Linus, the error numbers don't match up with SYSV's use of them. So I should think that makes Linux free and clear, and puts FreeBSD in the line of fire due to its SYSV ABI emulation.
And further, macros in the header files (ctype) would be compiled into the C code, so I doubt the "ABI" for those can be copyrighted, given that it's not interfacing with anything once it reaches binary form. "API", on the other hand, would be problematic.
It hasn't been done because nobody wants to see a 10-year lawsuit over who owns "www.sex" when a forged letter is sent to a negligent registrar who nonetheless disclaims all responsibility for anything they claim to do.
This may have been cloaked as a "We'd like to pretend there is no porn online" meme.
They chose to implement L1 as a second "Yes" button, so as long as you never make mistakes, you can play 1-handed. It was most handy for playing while on the phone.
Depends on the game. For things like Final Fantasy it wouldn't be too bad, but Super Metroid would be nigh impossible. I've been playing these on keyboard, albeit 2-handed, through ZSNES, and it's very difficult to use the grappling beam in Metroid even with both hands. Even things like Descent which were designed with a keyboard in mind took two hands to operate.
Slashdot Mirror
I looked into writing a JPEG2000 pixbuf loader for GTK2 once. The only open-source-ish implementation I could find was JasPer, which had too many licensing requirements to be useful. I tried to find the specifications, but not even the baseline specs are free. So I said "Screw this" and abandoned the idea of ever supporting JPEG2000 in anything, and I suspect I'm not alone.
"Not to familiar with PNG, does it also use wavelet compression?"
Nope. PNG uses lossless LZ77 encoding, also used by gzip.
"As for the P4/1.8GHz story try this for a test : Install MySQL on your linux PC and create a database with a table of about 5-6GB. Run alter table on it. Wait for it CRUMBLE TO DUST as it hits past 2GBs. Then get a Sun."
No way would I ever use a toy like MySQL if I had the option of Oracle, DB2, or even PostgreSQL. News flash for you: real software uses the 64-bit I/O calls.
Pretty sure it was The Trees, but it could've just as well been Freewill or A Passage to Bangkok. Sometime in my very last days of high school (1999) over a 14.4 modem (1996). Indeed, it took my parents 3 years to get around to getting online.
I don't have it anymore, as my system wasn't purchased until a year and a half later, and the 1GB drive in the old box got overcrowded in the meantime.
Those instances of Slashdotter are not necessarily the same person. Furthermore, Microsoft would be totally braindead to make a business plan from Slashdot comments. BillG probably isn't that dumb, despite what we wish for ;-)
Yep, not even the Color-I Plus. No power LED for us!
My dad's had to repair it a couple times, but we still run it for video gaming. Man, that thing's had more stuff connected to it....
* I believe the Ohio Scientific with a huge 8K RAM used a different monitor, and the C= 64 was the original reason this one was purchased. But I'm too young to remember anything before the Amigas very well.
** To run the Amigas, my dad built a custom cable and added a plug to the monitor to hook the Amiga RGB output up more-or-less directly to the electron guns.
"Instead of sending the whole email content - and with it the ability to falsify email header information, why not just send the email header only - and require the originating server to hold the email content?"
I thought about this once. How do you ensure that the person coming back to get the mail off the originating server is indeed the person to whom the mail was sent? I don't think any certificate-based scheme is the answer, because key management is so complicated. At least in their current incarnations. (If there was something as simple as ssh-keygen that also posted the public key to the ISP, it might be workable. But there isn't, and there's the question of mail client support.)
Regarding the "It shows your IP address" that mabu pointed out, wouldn't it be trivial to proxy the "get the content" message through the ISP's MTA?
Required in what market? I should think that if there was such a demand, instead of chips putting out 70W, we would see VIA C3 or Transmeta-based offerings taking over the world.
"In the given environment, it performs the same, so why would anyone use or even want the more expensive chip?"
Because it has bigger numbers, so it *must* be better! (See also: Intel, Megahertz Myth)
Ever see an Evergreen 286-to-486 upgrade chip? A board with a 486 and some cache on it, and a 286-compatible set of pins underneath. You had to run a little program to enable the cache or it would perform worse than a 286, but once it was all set up, it made the fastest 286 on the block....
According to Scott Mueller's Upgrading and Repairing PCs, the early SX's were DX's with the FPU mysteriously disabled. Later SX's really had no FPU at all.
And in another strange twist, a "487" was a full-blown 486DX with a different pin layout that disabled the main processor.
Well, there *is* that POSIX subsystem collecting dust somewhere under NT....
I did something similar. Some version of Oracle and my scanner software conflicted, and uninstalling Oracle didn't replace the DLLs... and neither did a non-destructive reinstall of Windows. So I switched to Linux (it was a dual-boot box), and figured I'd scribble over the FAT and Windows would actually format it. But rather than hde1, I put the plain hde in the dd command.
NO OPERATING SYSTEM INSTALLED
INSERT SYSTEM DISK AND PRESS ENTER
Windows 98SE will cheerfully install on a system with no partition table, but it cannot be booted afterwards. Thank God for freshmeat! Through them I found a static gpart (guess partitions) and restored everything... by sneakernetting it to another box, because no Linux tools would put the partition boundaries out in the middle of tracks! After that, I use fdisk and make my own partitions, and print the table when I'm done.
Nah, the default is whatever the sysadmin set up in /etc/skel/$SHELL_FILE. The real default looks like "%" for csh, or "bash-2.05b$ " for bash. I think both change to # for a root prompt.
Random note: bash uses its argv[0] in the prompt, so it's possible to see "init-2.05b# " if it's a really bad day....
Holy shit. If this is true, Python mops the floor with Perl on arbitrary precision. I have a Math::BigInt perl process doing only the long test which has been going on for 77:51 (min:sec) as I post this, whereas the entire benchmark in Python was over in 37:19. (I've constrained the accuracy/precision to ~67-bit range in Perl. It's still taking forever.)
OTOH, my Python results are somewhat dubious, since it comes up with 10 billion instead of ~777 million for that test, claiming -2940 seconds for it. Subtracting the other times, it really took 23:02.
So if a 5-year-old CPU can handle any task, why do they keep building these supercomputer clusters like the Earth Simulator or the G5 cluster? Why do we have video cards if a CPU can do everything?
The reason CPUs are fast enough to handle tasks we do today is that we're not throwing intensive problems at them. Imagine dispatching an AI to search your images by voice, then continuing a refreshing game of Final Fantasy Tactics on an emulator: the beefiest SMP box with multi-channel memory and RAID 10 would be screaming in pain.
That last bit's pretty important. For me, Java and C produce 776627965 as the long result, whereas Python is giving me 10000000000. (And it's not some braindead failure like printing the wrong variable.) And Python came up with that result in -2940 seconds...
Only if the ISP is clueful. Some (alltel.net) won't relay mail even if the sender is currently on one of their own dialup lines.
Dude! My MBR is only 512 bytes, and part of that is the primary partition table and boot signature, leaving (IIRC) 448 bytes for GRUB stage1. Stage 1's responsibility is loading the embedded stage 2 from the following 16 sectors of disk, which falls in a mysterious "reserved"[1] area (according to FreeBSD's installer), and stage 2 can actually understand the filesystem on the partition where its config is to actually do something useful.
x86 is vile.
[1] That reserved area was 31 sectors on my old 15GB IBM DTLA, and 63 sectors on my new 80GB Seagate.
Doubtful. If we are to believe Linus, the error numbers don't match up with SYSV's use of them. So I should think that makes Linux free and clear, and puts FreeBSD in the line of fire due to its SYSV ABI emulation.
And further, macros in the header files (ctype) would be compiled into the C code, so I doubt the "ABI" for those can be copyrighted, given that it's not interfacing with anything once it reaches binary form. "API", on the other hand, would be problematic.
It hasn't been done because nobody wants to see a 10-year lawsuit over who owns "www.sex" when a forged letter is sent to a negligent registrar who nonetheless disclaims all responsibility for anything they claim to do.
This may have been cloaked as a "We'd like to pretend there is no porn online" meme.
Arrrgh! I will never be able to read things like that without thinking "Cold Fusion light bulbs".
Ha. Go back and read the eerily similar section of the GPL.
They chose to implement L1 as a second "Yes" button, so as long as you never make mistakes, you can play 1-handed. It was most handy for playing while on the phone.
Depends on the game. For things like Final Fantasy it wouldn't be too bad, but Super Metroid would be nigh impossible. I've been playing these on keyboard, albeit 2-handed, through ZSNES, and it's very difficult to use the grappling beam in Metroid even with both hands. Even things like Descent which were designed with a keyboard in mind took two hands to operate.