20K L1 for P4, down from 32K since the Pentium MMX. Even the Itanium2 [geek.com] only has 32K.
Just for people who don't know, Intel reduced the amount of cache when they moved from the P3 to the P4. And hardware junkies know the performance hit that caused.
A seemingly unrelated sidenote: Intel wants to move to their IA-64 system, and, since it's not backwards-compatible, they're going to have to force a grass-roots popular movement to pull it off.
Perhaps they crippled the P4 to make the IA-64 processors look even faster to the general public?
In any case, I think the quality of the P4 is a sign that Intel wants to make its move soon. (Though losing $150 million, not to mention the context in which they lost it, may set back their schedule, giving AMD's 64-bit system a chance to catch on.)
That's also why it's cheap. At least for the end receiver.
You'd be amazed at the volume of hd space required to host the mail for two or three thousand accounts, if only for just two weeks per piece of mail.
I've even seen servers taken down, in a matter of days, because they were inadvertently accepting email when sent directly to their IP.
A couple of BBS operators I know (not the ones referenced by my sig) used to handle the email through their Worldgroup BBS software. They forgot to turn it off when they switched to a w32 system...
Funny that saying "I accept" when installing software is equivalent to signing a more recent agreement.
So Microsoft violates the big agreement, University sues, stating that Microsoft violated their agreement.
Microsoft argues that installing software is tantamount to rewriting the old agreement, and countersues for libel. (Saying that such a major company violated a major agreement will cause a major reduction in stock values.)
Easy...as soon as you increase the exposure time for your camera(to get a blur instead of nothing at all), or integrate any method where two CCDs start their exposures at slightly different times(To aid interpolation), you get slapped by the DMCA.
More copyright technology begets more desire for copyright-thwarting technologies (or technologies whos spinoffs include copyright-thwarting). More desire begets more bang-for-the-buck the RIAA gets for paying to get the DMCA passed. (I bet Clinton could be traced to such illicit dealings...but nobody's interested now.)
That in turn means what the DMCA was for all along: Develop all the copy-protection you want, it's now illegal for anyone to even develop technologies that can be used to break it.
I can see why he was moderated as "troll," but such moderation is the same level of demonstration of personal bias as what oliverthered was demonstrating.
But then, oliverthered has the right to free speech, doesn't he? And further, he was demonstrating his opinion that people should be allowed their own opions, even specifying that he wasn't in the same boat as them. (One would normally expect Christian Scientists to make his claim, not agnostics or athiests. Such people as the latter, in this case and cases like it, hold my highest esteem.
More on-topic:
Civil rights in a democracy boil down to one thing: How tolerant the majority of the actual voters are towards others' beliefs. If you don't like it, you can go to a dictated nation that supports America's first ten ammendments.
Personally, I hate the fact that competent medical physicians can be denied access to children, or anyone, who needs medical attention. Heck, I don't even have respect the people who believe that they should pray their way out of sickness.
God gave us the means to help ourselves, isn't that enough?
I hate to see it so soon, but I gaurantee everyone here that we'll soon start seeing ad campaigns against these things in public. They'll be pushed by these same, uninformed opinions.
There is absolutely zero advantage from a legacy perspective in using XMLQuery in applications where there is no legacy SQL database to work arround.
You never mentioned that you weren't dealing with SQL at all, just that you didn't see the sense in using XMLQuery.
Note that I'm not trying to make a convert of you.
The point behind using SQL is twofold: it's an open standard, meaning no patent or licensing issues, and it's a standard system, meaning your successor will be more able to understand what's going on. (Helpful in the same way as knowledge about RPN when looking at FORTH code.)
The point of using XMLQuery is that it's portable between servers. If you discover serious limitations PostgreSQL, you can just use MySQL instead. No rewriting of code.
It's also a standard, meaning that an SQL server could be written from the ground up, to make it fast, since it already knows what its communications framework will be, and can work with it quickly, not having had to adapt an existing system to it.
Introducing XML Query in those cases simply ties an infrastructure that was independent of SQL lossage direct to SQL. So instead of working with an object oriented data model based on typed set theory we are back to 1960s style hacking arround with entity relation data models.
Working with any "standard" limits your design in some manner. Working with an object-oriented system instead of a data-relation system is a debate for theorists, whom you and I are obviously not. However, whenever I hear the phrase "object-oriented" or see something disparaged by its age, I hear bells and think, "Hype warning!"
I won't argue about the differences in performance between IMAP and POP3. I don't know much about that stuff.
To take the example it is as if someone had said that the VGA standard is so good that the O/S should require all screens to be 640x480 pixels so that we could be sure that an application would never produce a picture that was too big for a smaller display.
VGA required all screens to be 640x480. SVGA allowed for arbitrary resolutions. Both allowed for arbitrary monitor sizes and color depths.
Scaling has been, and will be for a long time, a developer issue. When I use GQView, I tell it to shrink to fit in the window. If I want the extra detail, I can scroll around.
Or to take the example of C it would be like saying that the only data model to be used would be structures referenced by unguarded pointers, we would never move on to Java or C#.
I specifically pointed out that C, with all its rough edges and raw power, was useful for small-to-medium-sized applications, where speed is critical. Large applications can afford the slightly slower C++, Objective C, and the massively slower Java. The value of object-oriented program structure becomes apparent from the developer perspective when you start dealing with massive, integrated applications.
The day I see a C++ - based benchmarking program offering "hard" data, I'll either laugh hysterically or cry a little. (Depends on if I'm depressed.:)
Take a look at the X86 CPU architecture. The SVGA standard. Take a look at C. (Please, no flames.) Even the use of silicon for electronics is a "legacy" standard.
C and the X86 instruction set have been around for over twenty years. SVGA has been around for about that long.
All three of these are doubtless "legacy" specifications.
The 386 instruction set has been the "least common denominator" for PC-based programs since the early, early 90s. There isn't a single 32-bit C-compliant program out there that can't be compiled for it. This is an example of interoperability which, as the poster pointed out, is the purpose of the W3C. Every standard exists to improve interoperability.
Heck, even "proprietary" formats have their own standard, so that at least one product can consistently work with it, no matter what part of that product is involved.
SVGA is the basis for communication between nearly every PC and monitor. Not only does it support modularity (your monitor will work with my machine), but it's open-ended. We haven't reached the limit to how much visual data can pass from your computer to your monitor. Since it's analog, the amount of data transferred depends on the capabilities of the hardware, not the standard. Digital formats can't compete yet, because you can only pack so much data into a finitely-resolutioned data stream.
With the way the SVGA standard is structured, you only have to worry about a small number of things.
C is excellent for fast, compact applications. C++ has its uses, but those are largely limited, IMHO, to large, integrated projects. For small, one-purpose utilities, C is usually faster. (And remember that repetitive processes go a lot faster when the individual cycle is only a little bit faster.) C has been extended a few times, (POSIX.1, POSIX.2, C99, and a couple of others I don't recall) but it is largely the same. And ANSI C (C89) is still the base standard, available across-the-board.
Silicon...where would we be today without it? It hasn't outlived its usefulness, at least not yet. Take the dead horse, for example. You can beat it all you want, and you'll be the one pushing it. You could breed a new horse, but then you have to wait for it to grow up, you have to "break" it, you have to get to know it, and, finally, you get to ride it. The whole process takes years.
And then there's the old saw, "If it ain't broke, don't fix it."
Sealand was declared sovereign before international borders were established. (It used to be that if you were outside the range of British cannons (2 1/2 miles), you weren't under their juristiction.)
Sorry...didn't mean to be anal about it. Anyway...
I can understand that it takes usecs for the wireless interfaces to sort themselves out, but what about continuing the routing process over the Internet? Are Internet ARP tables generally updated fast enough for this?
Or will there have to be a primary bridge every so often along the highway? (With primary bridges, you could reduce the number of times you have to update your location ala ARP.)
Nothing is instantaneous. Twenty five years ago, computers performed orbit calculations "nearly instantaneously," compared to the pen and paper of twenty five years before.
The question is, does it adapt fast enough for me to stream data to a neighborhood, not only when I'm in range of it, but when I'm in range of a series of moble points which is flowing past the neighborhood in question? The fastest route will be constantly changing, and possible routes will constantly be passing in and out of range of my target.
Implementing functionality so specific to that application into the kernel (probably Linux, maybe *BSD, HURD, etc.) would be the programmatical equivalent of "-1 Flaimbait"
If you could make a dedicated device to act as a NIC, but with all networking brains internal, you'd be able to create your nationwide unbreakable network.
If you can store packets until a path is available (UUCP-like retrying until success), and probably make the packets larger (4K-like).
That's a whole new monster of a network. That's be cool.:)
Just realized: Anyone know how quickly 802.11 devices can connect, authenticate, route and transfer? (in that order...authentication optional.)
Can it be done without loosing running TCP/IP streams, or will we start seeing a prevalence of UDP/IP embedded protocols with connection-owner-tracking methods built in?
And without the UDP/IP embedded protocols, DHCP would/will be a real PITA.
Slashdot Mirror
It's sad that such a reference to The Rocketeer can go unnoticed.
Doh! And they were sitting on my desk at the time...
20K L1 for P4, down from 32K since the Pentium MMX. Even the Itanium2 [geek.com] only has 32K.
Just for people who don't know, Intel reduced the amount of cache when they moved from the P3 to the P4. And hardware junkies know the performance hit that caused.
A seemingly unrelated sidenote: Intel wants to move to their IA-64 system, and, since it's not backwards-compatible, they're going to have to force a grass-roots popular movement to pull it off.
Perhaps they crippled the P4 to make the IA-64 processors look even faster to the general public?
In any case, I think the quality of the P4 is a sign that Intel wants to make its move soon. (Though losing $150 million, not to mention the context in which they lost it, may set back their schedule, giving AMD's 64-bit system a chance to catch on.)
IIRC, that's about a 6-month budget for an "Public Radio" station.
Funny, the site seems to be slashdotted:
http://www.halfbakery.com/idea/TiVo_20Car_20Radio
Isaac Asimov's Moonbase is a story about nanomachines, their uses, and the violent public backlash. (Those people were called "nanoluddites")
They'd probably sell you a liquifying tool. Probably a cheap miniature blender.
They could give them away as sort of a promotion... "Buy this battery pack and we'll throw in a Fuelizer(tm) for free!"
That's also why it's cheap. At least for the end receiver.
You'd be amazed at the volume of hd space required to host the mail for two or three thousand accounts, if only for just two weeks per piece of mail.
I've even seen servers taken down, in a matter of days, because they were inadvertently accepting email when sent directly to their IP.
A couple of BBS operators I know (not the ones referenced by my sig) used to handle the email through their Worldgroup BBS software. They forgot to turn it off when they switched to a w32 system...
Funny that saying "I accept" when installing software is equivalent to signing a more recent agreement.
So Microsoft violates the big agreement, University sues, stating that Microsoft violated their agreement.
Microsoft argues that installing software is tantamount to rewriting the old agreement, and countersues for libel. (Saying that such a major company violated a major agreement will cause a major reduction in stock values.)
Oh, how the legal system is so...
Why does everything seem to loop back to the DMCA now?
Easy...as soon as you increase the exposure time for your camera(to get a blur instead of nothing at all), or integrate any method where two CCDs start their exposures at slightly different times(To aid interpolation), you get slapped by the DMCA.
More copyright technology begets more desire for copyright-thwarting technologies (or technologies whos spinoffs include copyright-thwarting). More desire begets more bang-for-the-buck the RIAA gets for paying to get the DMCA passed. (I bet Clinton could be traced to such illicit dealings...but nobody's interested now.)
That in turn means what the DMCA was for all along: Develop all the copy-protection you want, it's now illegal for anyone to even develop technologies that can be used to break it.
If you'd like to recommend to your superiors to look here for Linux-centric employees...
I can see why he was moderated as "troll," but such moderation is the same level of demonstration of personal bias as what oliverthered was demonstrating.
But then, oliverthered has the right to free speech, doesn't he? And further, he was demonstrating his opinion that people should be allowed their own opions, even specifying that he wasn't in the same boat as them. (One would normally expect Christian Scientists to make his claim, not agnostics or athiests. Such people as the latter, in this case and cases like it, hold my highest esteem.
More on-topic:
Civil rights in a democracy boil down to one thing: How tolerant the majority of the actual voters are towards others' beliefs. If you don't like it, you can go to a dictated nation that supports America's first ten ammendments.
Personally, I hate the fact that competent medical physicians can be denied access to children, or anyone, who needs medical attention. Heck, I don't even have respect the people who believe that they should pray their way out of sickness.
God gave us the means to help ourselves, isn't that enough?
Hope I get oliverthred's moderation in metamod.
I hate to see it so soon, but I gaurantee everyone here that we'll soon start seeing ad campaigns against these things in public. They'll be pushed by these same, uninformed opinions.
Sigh...such is free speech.
Yup. But how important is it that we consider Mercury a planet and Titan a moon?
In all the SF I remember reading, if it had an atmosphere, it was colonizable. If it didn't, it was minable. Seems like a logical distinction to me.
You're trolling now, but I'll bite.
:)
There is absolutely zero advantage from a legacy perspective in using XMLQuery in applications where there is no legacy SQL database to work arround.
You never mentioned that you weren't dealing with SQL at all, just that you didn't see the sense in using XMLQuery.
Note that I'm not trying to make a convert of you.
The point behind using SQL is twofold: it's an open standard, meaning no patent or licensing issues, and it's a standard system, meaning your successor will be more able to understand what's going on. (Helpful in the same way as knowledge about RPN when looking at FORTH code.)
The point of using XMLQuery is that it's portable between servers. If you discover serious limitations PostgreSQL, you can just use MySQL instead. No rewriting of code.
It's also a standard, meaning that an SQL server could be written from the ground up, to make it fast, since it already knows what its communications framework will be, and can work with it quickly, not having had to adapt an existing system to it.
Introducing XML Query in those cases simply ties an infrastructure that was independent of SQL lossage direct to SQL. So instead of working with an object oriented data model based on typed set theory we are back to 1960s style hacking arround with entity relation data models.
Working with any "standard" limits your design in some manner. Working with an object-oriented system instead of a data-relation system is a debate for theorists, whom you and I are obviously not. However, whenever I hear the phrase "object-oriented" or see something disparaged by its age, I hear bells and think, "Hype warning!"
I won't argue about the differences in performance between IMAP and POP3. I don't know much about that stuff.
To take the example it is as if someone had said that the VGA standard is so good that the O/S should require all screens to be 640x480 pixels so that we could be sure that an application would never produce a picture that was too big for a smaller display.
VGA required all screens to be 640x480. SVGA allowed for arbitrary resolutions. Both allowed for arbitrary monitor sizes and color depths.
Scaling has been, and will be for a long time, a developer issue. When I use GQView, I tell it to shrink to fit in the window. If I want the extra detail, I can scroll around.
Or to take the example of C it would be like saying that the only data model to be used would be structures referenced by unguarded pointers, we would never move on to Java or C#.
I specifically pointed out that C, with all its rough edges and raw power, was useful for small-to-medium-sized applications, where speed is critical. Large applications can afford the slightly slower C++, Objective C, and the massively slower Java. The value of object-oriented program structure becomes apparent from the developer perspective when you start dealing with massive, integrated applications.
The day I see a C++ - based benchmarking program offering "hard" data, I'll either laugh hysterically or cry a little. (Depends on if I'm depressed.
Why don't we just define a "planet" as "an object that maintains an atmosphere via its gravity well."?
It'd sure clear up a lot.
Take a look at the X86 CPU architecture. The SVGA standard. Take a look at C. (Please, no flames.) Even the use of silicon for electronics is a "legacy" standard.
C and the X86 instruction set have been around for over twenty years. SVGA has been around for about that long.
All three of these are doubtless "legacy" specifications.
The 386 instruction set has been the "least common denominator" for PC-based programs since the early, early 90s. There isn't a single 32-bit C-compliant program out there that can't be compiled for it. This is an example of interoperability which, as the poster pointed out, is the purpose of the W3C. Every standard exists to improve interoperability.
Heck, even "proprietary" formats have their own standard, so that at least one product can consistently work with it, no matter what part of that product is involved.
SVGA is the basis for communication between nearly every PC and monitor. Not only does it support modularity (your monitor will work with my machine), but it's open-ended. We haven't reached the limit to how much visual data can pass from your computer to your monitor. Since it's analog, the amount of data transferred depends on the capabilities of the hardware, not the standard. Digital formats can't compete yet, because you can only pack so much data into a finitely-resolutioned data stream.
With the way the SVGA standard is structured, you only have to worry about a small number of things.
C is excellent for fast, compact applications. C++ has its uses, but those are largely limited, IMHO, to large, integrated projects. For small, one-purpose utilities, C is usually faster. (And remember that repetitive processes go a lot faster when the individual cycle is only a little bit faster.) C has been extended a few times, (POSIX.1, POSIX.2, C99, and a couple of others I don't recall) but it is largely the same. And ANSI C (C89) is still the base standard, available across-the-board.
Silicon...where would we be today without it? It hasn't outlived its usefulness, at least not yet. Take the dead horse, for example. You can beat it all you want, and you'll be the one pushing it. You could breed a new horse, but then you have to wait for it to grow up, you have to "break" it, you have to get to know it, and, finally, you get to ride it. The whole process takes years.
And then there's the old saw, "If it ain't broke, don't fix it."
Without a mention of the fact in its foreward, obviously. :)
Sealand was declared sovereign before international borders were established. (It used to be that if you were outside the range of British cannons (2 1/2 miles), you weren't under their juristiction.)
That and a few hundred other comments on this previous (Napster + Sealand) story on Slashdot.
Sorry...didn't mean to be anal about it. Anyway...
I can understand that it takes usecs for the wireless interfaces to sort themselves out, but what about continuing the routing process over the Internet? Are Internet ARP tables generally updated fast enough for this?
Or will there have to be a primary bridge every so often along the highway? (With primary bridges, you could reduce the number of times you have to update your location ala ARP.)
Nothing is instantaneous. Twenty five years ago, computers performed orbit calculations "nearly instantaneously," compared to the pen and paper of twenty five years before.
:)
The question is, does it adapt fast enough for me to stream data to a neighborhood, not only when I'm in range of it, but when I'm in range of a series of moble points which is flowing past the neighborhood in question? The fastest route will be constantly changing, and possible routes will constantly be passing in and out of range of my target.
Implementing functionality so specific to that application into the kernel (probably Linux, maybe *BSD, HURD, etc.) would be the programmatical equivalent of "-1 Flaimbait"
If you could make a dedicated device to act as a NIC, but with all networking brains internal, you'd be able to create your nationwide unbreakable network.
If you can store packets until a path is available (UUCP-like retrying until success), and probably make the packets larger (4K-like).
That's a whole new monster of a network. That's be cool.
It's called "netsplit."
...seems to require a relatively high density.
:)
One phrase: Rush-hour/traffic-jam entertainment.
Just realized: Anyone know how quickly 802.11 devices can connect, authenticate, route and transfer? (in that order...authentication optional.)
Can it be done without loosing running TCP/IP streams, or will we start seeing a prevalence of UDP/IP embedded protocols with connection-owner-tracking methods built in?
And without the UDP/IP embedded protocols, DHCP would/will be a real PITA.