Actually, the first multiplayer 3d shooter was probably a tank fighting game that was a GL demo for SGI systems. Kind of a neat game actually. It got better as GL got better. Preceeded even Wolfenstein 3D, but it required awesomely expensive hardware.:-)
Also, the multiplayer innovations in Quake were actually, to my knowledge, pioneered in nettrek for Unix boxes as well.
For a couple of other borrowed ideas...
Ultima Online - MUDs
Diablo - Rogue/Moria/Nethack/Angband
I like the PC games mentioned, and have a lot of respect for those who wrote them. But, sometimes it irritates me that people forget the genesis of the big ideas that went into them.
I haven't applied the label 'horrible' to Mozilla since 0.8, and 0.9.1 is actually pretty good. Aside from all the stupid media and Flash support in IE, I actually prefer Mozilla to IE now.
Also, VirtualAlloc there sounds and awful like like 'mmap'. Again, same basic idea, and Microsoft does it completely differently.
I know a fair amount about the insides of NT, and most design choices they made that are different than Unix's are worse.
Here are just two:
A FIFO VM?!?!? How stupid can you get? LIFO is much better, and while not really achievable, you can come closer than FIFO using a mark & sweep-like system (or perhaps there are better algorithms today).
WaitForMultipleObjects, you mean, every single semaphore and mutex call is an OS call now? No 10-20 cycle mutex grabs when there's no contention?
You misunderstand 'poll' completely. poll asks the OS to suspend your process until one of the indicated events happens, then you get to go respond to it. It's essentially the same thing.
Say, for example, that your dumping data into a socket. Under Unix, you write to the socket until the OS tells you that the socket buffer is full by setting the socket to non-blocking and writing until write returns EAGAIN as an error. Then you put the ability to write to that socket on the list of OS events you're interested in. Then, you go do whatever else it is you have to do. After you get done servicing everything you can service, you call poll and it blocks your process (possibly running others) until one of the indicated events happens and there's something else to service. Same basic paradigm.
Nice! So in other words, they used straight BSD sockets for their
implementation - which is NOT the way to get performance from Windows. You
need to use:
Asynchronous, Event based socket handling.
Completion ports.
Scatter/Gather buffering.
Polling is lousy no matter what way you do it. You'll lose most of your performance spent going round a small loop.
You're an idiot. They're using the 'poll' system call. If you bothered to read anything, you'd realize that 'poll' is the way to do asynchronous event based I/O under Unix. It's close to what 'WaitForMultipleObjects' does under NT.
They may use the sockets API, but as far as I know, that's the way to do TCP/IP under Windows. There are a few special calls to get NT 'handles' for your sockets so you can then do WaitForMultipleObjects based event based I/O handling. I'm betting this is exactly what they did.
As for scatter gather buffering, that depends a lot on your internal application architecture. I would agree that, in general, it's a good idea. I don't think their code would do scatter gather under Unix, and not under NT. Scatter gather is implemented nearly identically under both platforms.
Your comment shows a great deal of ignorance. It's a travesty that you were moderated to +5. *sigh*
The architecture they say performs the fastest, One-thread-many-tasks (asynchronous),
is exactly the one encouraged and supported by my
StreaModule system. I knew that things worked out this way, but I'm quite surprised to find such clear agreement by a third party. This idea doesn't really seem to crop up in many places.
It's like a multi-cpu machine, except instead of only sharing the bus, and possibly the L2 cache, it shares the L1 cache, the TLB, and a few other things. Also, the second CPU is on the same fleck of silicon as the first.
So, have they bothered with a 'select'-like statement, or is it still 3000+ threads for a server with 3000 cllients?
IMHO, Java is NOT industrial strength. It is fine for many things, but for what I want to do, C++ is still it. It will never be as fast as C++. All those nice run-time optimizations being applied to Java work for C++ too, if anybody would bother. Of course, since C++ is generally tons faster, nobody has yet.
As for worrying about all that icky memory stuff. I want to worry about it. My programs are faster and better designed for it. I think it would be highly amusing to plop down a Java programmer in an environment where careful memory management was crucial to successful execution. They wouldn't know their heads from their arses.
It's possible to do memory management in Java, despite the garbage collector, but it isn't as easy, and nobody feels they have to with the nice, warm, fuzzy garbage collector wrapping them up.
I think Java is fine for many things. I've watched its development and maturation with interest. I actually made a good stab at porting JVM beta 1 to my platform (UnixWare). It's just not the wonderful be-all and end-all language you make it out to be.
It also makes me extremely nervous that Sun still has such tight control of it.
BTW: If you have tendonitis, a lot of doctors will tell you to take Advil (or a generic version of it). That's because Advil is an anti-inflamitory drug that actually reduces the swelling in your wrist. Since the swolen tendons is what is causign all the friction (which causes more pain and more swelling), reducing it is a Good Thing. Other pain killers will 1: not reduce the swelling, and 2: ease the pain, allowing you to cheerfully type away and hurt yourself more.
Umm, aspirin has similar anti-inflammatory properties to ibuprofen (Advil). I believe that the painkiller in Aleve is also an anti-inflammatory.
Acetominaphine (Tylenol) is not a good anti-inflammatory, but it used for most pain relievers because of its perceived safety. I say perceived because acetominaphine does not have any warning signs (like aspirin's ringing in the ears) before fatal overdose, and a fatal overdose will flat out kill your liver. A dead liver is a recipe for a nasty, unpleasant death.
As for RSI being 'all in your head', I'm willing to believe it, though the article presents insufficient evidence. IMNSHO, that in no way diminishes the seriousness of the disease, or makes the pain less 'real' (since pain is a perception). It just suggests a different method of treatment. We, here in America have a strange disdain of suffering caused by some non-physical source.
It doesn't have to die. One of these days, in my copious free time, I will make some kind of implementation of Radia Perlman's idea for doing multicast into the Linux UDP/IP stack. Then, if ISPs and router vendors don't implement multicast, they'll just suffer higher traffic loads than they need to until they do.
I disagree. I do not consider using the various intel 'trap' instructions to constitute linking to the kernel. The major libraries everybody links against under Linux are all covered under the LGPL which is much more lenient in this regard than the GPL.
The LGPL is specifically designed to address the very concern that you have. If the author had wanted to allow his library to be used in this way, he would've put it under the LGPL.
Some Open Source companies will do well. It's only a matter of time. Cygnus was profitable before RedHat bought them. RedHat will most likely be profitable soon.
Also, the King amasses a great deal of wealth, and wealth is important, but that doesn't mean we should have monarchies.
The real problem here is the process by which funding is allocated on the basis of census data.
In the US, during the census push, they went on and on about how filling out the census would help your wise and powerful legislators appropriately apportion the largess of taxes to the people. To me, it completely damages the credibility of the census as an 'apolitical' process.
I don't want a 'father government' who I must tell the truth to so that he might wisely care for the needs of the individual family members he is somehow responsible for. The whole belief structure leading to the idea of needing an accurate census is an anathema to me.
If someone in my community wants my help, they can ask me. They shouldn't go running to the government to get them to take money from my pocket and put it in theirs. That kind of thing breeds dissension and distrust and destroys communities.
Cable providers are such pains about running services on your computer at home. They absolutely want you to be a passive consumer of information, no matter what.
I have a/29 from my ISP, and run various low bandwidth services (including being a primary DNS server for several domains) from my computer at home. I could never do that reliably with a cable modem.
My actual available bandwidth is more reliable as well. Especially if I go through my ISPs squid proxy for most stuff. The squid proxy is connected to the net at 45 MBps and the higher peak rates can smooth out network bumps to make my transfer rate a smooth 65k/sec. I wish they were more vocal about their squid proxy because then there'd be a better chance that stuff I want would be in it, and I wouldn't have to fetch from the Internet at large at all.
Of course, having a good ISP is a must. I wouldn't ever use QWorst as my ISP. I've been extremely pleased with VISI.
Re:Don't know if it has been said yet...
on
To the Moon, Alice
·
· Score: 1
Interesting problem. Do you think, that if he solves the balancing problem that he'll have angular momentum problems once he gets high enough? Is 30 miles high enough? Will he accidentally hit the ocean or some city because the earth turned underneath him while his rocket was in the air?
Slashdot Mirror
Yep, bz. *grin*
I liked using the 'stealth' take and making creative use of cover and obstacles to sneak up on people.
I did it on purpose. :-) I just wanted the game to be something a little more dignified than "Ultimate Clash of the neferious Space slugs of hate".
Actually, the first multiplayer 3d shooter was probably a tank fighting game that was a GL demo for SGI systems. Kind of a neat game actually. It got better as GL got better. Preceeded even Wolfenstein 3D, but it required awesomely expensive hardware. :-)
Also, the multiplayer innovations in Quake were actually, to my knowledge, pioneered in nettrek for Unix boxes as well.
For a couple of other borrowed ideas...
I like the PC games mentioned, and have a lot of respect for those who wrote them. But, sometimes it irritates me that people forget the genesis of the big ideas that went into them.
I haven't applied the label 'horrible' to Mozilla since 0.8, and 0.9.1 is actually pretty good. Aside from all the stupid media and Flash support in IE, I actually prefer Mozilla to IE now.
You're right about LIFO. I apologize. I confused my acronyms. I meant LRU, not LIFO. :-(
How do I/O completion ports work exactly? And how are they better than WaitForMultipleObjects? Why does WaitForMultipleObjects have that limitation?
Also, VirtualAlloc there sounds and awful like like 'mmap'. Again, same basic idea, and Microsoft does it completely differently.
I know a fair amount about the insides of NT, and most design choices they made that are different than Unix's are worse.
Here are just two:
You misunderstand 'poll' completely. poll asks the OS to suspend your process until one of the indicated events happens, then you get to go respond to it. It's essentially the same thing.
Say, for example, that your dumping data into a socket. Under Unix, you write to the socket until the OS tells you that the socket buffer is full by setting the socket to non-blocking and writing until write returns EAGAIN as an error. Then you put the ability to write to that socket on the list of OS events you're interested in. Then, you go do whatever else it is you have to do. After you get done servicing everything you can service, you call poll and it blocks your process (possibly running others) until one of the indicated events happens and there's something else to service. Same basic paradigm.
You're an idiot. They're using the 'poll' system call. If you bothered to read anything, you'd realize that 'poll' is the way to do asynchronous event based I/O under Unix. It's close to what 'WaitForMultipleObjects' does under NT.
They may use the sockets API, but as far as I know, that's the way to do TCP/IP under Windows. There are a few special calls to get NT 'handles' for your sockets so you can then do WaitForMultipleObjects based event based I/O handling. I'm betting this is exactly what they did.
As for scatter gather buffering, that depends a lot on your internal application architecture. I would agree that, in general, it's a good idea. I don't think their code would do scatter gather under Unix, and not under NT. Scatter gather is implemented nearly identically under both platforms.
Your comment shows a great deal of ignorance. It's a travesty that you were moderated to +5. *sigh*
The architecture they say performs the fastest, One-thread-many-tasks (asynchronous), is exactly the one encouraged and supported by my StreaModule system. I knew that things worked out this way, but I'm quite surprised to find such clear agreement by a third party. This idea doesn't really seem to crop up in many places.
It's like a multi-cpu machine, except instead of only sharing the bus, and possibly the L2 cache, it shares the L1 cache, the TLB, and a few other things. Also, the second CPU is on the same fleck of silicon as the first.
So, have they bothered with a 'select'-like statement, or is it still 3000+ threads for a server with 3000 cllients?
IMHO, Java is NOT industrial strength. It is fine for many things, but for what I want to do, C++ is still it. It will never be as fast as C++. All those nice run-time optimizations being applied to Java work for C++ too, if anybody would bother. Of course, since C++ is generally tons faster, nobody has yet.
As for worrying about all that icky memory stuff. I want to worry about it. My programs are faster and better designed for it. I think it would be highly amusing to plop down a Java programmer in an environment where careful memory management was crucial to successful execution. They wouldn't know their heads from their arses.
It's possible to do memory management in Java, despite the garbage collector, but it isn't as easy, and nobody feels they have to with the nice, warm, fuzzy garbage collector wrapping them up.
I think Java is fine for many things. I've watched its development and maturation with interest. I actually made a good stab at porting JVM beta 1 to my platform (UnixWare). It's just not the wonderful be-all and end-all language you make it out to be.
It also makes me extremely nervous that Sun still has such tight control of it.
And you don't think this would get any worse as time went on, and the feature became more accepted?
Umm, aspirin has similar anti-inflammatory properties to ibuprofen (Advil). I believe that the painkiller in Aleve is also an anti-inflammatory.
Acetominaphine (Tylenol) is not a good anti-inflammatory, but it used for most pain relievers because of its perceived safety. I say perceived because acetominaphine does not have any warning signs (like aspirin's ringing in the ears) before fatal overdose, and a fatal overdose will flat out kill your liver. A dead liver is a recipe for a nasty, unpleasant death.
As for RSI being 'all in your head', I'm willing to believe it, though the article presents insufficient evidence. IMNSHO, that in no way diminishes the seriousness of the disease, or makes the pain less 'real' (since pain is a perception). It just suggests a different method of treatment. We, here in America have a strange disdain of suffering caused by some non-physical source.
I doubt it would react to a magnetic field, but it would react strongly to an electric field. Statics would be a problem.
This, of course, is a bad mistake. As soon as one person cracks it, for all intents and purposes, everybody has.
It doesn't have to die. One of these days, in my copious free time, I will make some kind of implementation of Radia Perlman's idea for doing multicast into the Linux UDP/IP stack. Then, if ISPs and router vendors don't implement multicast, they'll just suffer higher traffic loads than they need to until they do.
I disagree. I do not consider using the various intel 'trap' instructions to constitute linking to the kernel. The major libraries everybody links against under Linux are all covered under the LGPL which is much more lenient in this regard than the GPL.
The LGPL is specifically designed to address the very concern that you have. If the author had wanted to allow his library to be used in this way, he would've put it under the LGPL.
Some Open Source companies will do well. It's only a matter of time. Cygnus was profitable before RedHat bought them. RedHat will most likely be profitable soon.
Also, the King amasses a great deal of wealth, and wealth is important, but that doesn't mean we should have monarchies.
This is close to what I said in a different comment, though you manage to say it much less antogonistically. :-)
The real problem here is the process by which funding is allocated on the basis of census data.
In the US, during the census push, they went on and on about how filling out the census would help your wise and powerful legislators appropriately apportion the largess of taxes to the people. To me, it completely damages the credibility of the census as an 'apolitical' process.
I don't want a 'father government' who I must tell the truth to so that he might wisely care for the needs of the individual family members he is somehow responsible for. The whole belief structure leading to the idea of needing an accurate census is an anathema to me.
If someone in my community wants my help, they can ask me. They shouldn't go running to the government to get them to take money from my pocket and put it in theirs. That kind of thing breeds dissension and distrust and destroys communities.
No, it is flamebait.
Ouch! Bad pun!
Cable providers are such pains about running services on your computer at home. They absolutely want you to be a passive consumer of information, no matter what.
I have a /29 from my ISP, and run various low bandwidth services (including being a primary DNS server for several domains) from my computer at home. I could never do that reliably with a cable modem.
My actual available bandwidth is more reliable as well. Especially if I go through my ISPs squid proxy for most stuff. The squid proxy is connected to the net at 45 MBps and the higher peak rates can smooth out network bumps to make my transfer rate a smooth 65k/sec. I wish they were more vocal about their squid proxy because then there'd be a better chance that stuff I want would be in it, and I wouldn't have to fetch from the Internet at large at all.
Of course, having a good ISP is a must. I wouldn't ever use QWorst as my ISP. I've been extremely pleased with VISI.
Interesting problem. Do you think, that if he solves the balancing problem that he'll have angular momentum problems once he gets high enough? Is 30 miles high enough? Will he accidentally hit the ocean or some city because the earth turned underneath him while his rocket was in the air?