Is the music industry really so dumb as to think that hardware and software solutions will really ever work?
Yes. They have many somewhat unethical DRM companies that smell money assuring them that yes, their scheme absolutely *will* work, or at least will work in a few years.
Frankly, I think they'd have better luck fingerprinting songs and refusing to play fingerprinted songs on a device that doesn't have rights to play it, and then shoot for a 2020 timeframe for having these legally required. Then you force every person who wants to listen to pirated music owning an illegal piece of hardware. Sure, some would, but it's like cable piracy. The losses are manageable.
The current schemes, which break if a single person anywhere manages to break the protection, are pretty silly.
Shipping songs lossily compressed (non-MP3 format) is a good first step. It makes MP3s or Oggs sound worse, doubly lossily compressed. Of course, trying to convince the consumer to buy worse-sounding music is going to be interesting...maybe if they introduce smaller media and market it as sexier, or find some other special features to add in. I dunno, maybe slap the instruments on separate tracks, so you can choose where to place the drums and whatnot when listening ("balance 5% to the left, tiny delay, slight amplitude cut on the snare drumps so they sound like they're off to the left"). Maybe embed lyrics in the new format.
It's a hell of a thorny problem, and you *know* consulting firms are doing anything they can to extract money from a wealthy and desperate industry.
That's one of the stupidest ad homonems I've ever seen on Slashdot, and that's saying something.
Frankly, I think a lot of people would hardly see that as an insult. If I say "most rabbis love long, complicated arguments about interpretations of religious law, though most people don't care in the least about such questions", I think you'd find a lot of rabbis and theologians that would agree. So?
Obviously, I really am not interested in spending hours of my time arguing over the particular classification of a package. It's pretty obvious that my viewpoint is not that of these folks. I'm simply pointing out that people on the Debian ML do do this, and that most people wouldn't. This guy is obsessed with the ideology, rather than the tech. You really can't argue that point.
If there weren't a number of geeks very concerned about things like licensing we wouldn't have Linux in the first place. We might have a nice kernel, but that's a long shot from a Free OS.
Besides the fact that that's completely irrelevant to what I said (I don't believe I said that people obsessed with legalities shouldn't exist), I think most people wouldn't care. What if FreeBSD was the big, popular, multi-distro OS that Linux is today. Would most of us *care* whether the kernel was BSD-licensed instead of GPL licensed? I sure as hell wouldn't.
Since it looks like this is getting into ideology, and you feel arguing ideological technicalities is fun, I'll play the game. I'll take ESR's viewpoint. ESR thinks that Open Source is better because it's *works better*. RMS thinks that Free software is better because it's ideologically distasteful to use non-Free software.
Logically, if open source software is superior, closed-source attempts to compete with it should fail (we'll assume perfect consumer information...naive model, but unless you really think that open source can be toppled by marketing...). So if MS takes the BSD kernel and tries to make a superior fork (closed source), they should lose out to the open source fork.
Therefore, I claim that the only people who claim that the GPL is necessary (rather than the BSD license) are those who (a) feel that open source software is inferior to closed source software, (b) feel that open source is so marginally better than closed source software that a simple influx of marketing will forever bar open source software from the market, or (c) are irrational.
Chew on that.
Debian's view is pretty simple: "If the software we use isn't Free, then someone can legally ask us to stop using it. Therefore, our operating system and its tools will always be Free, and no parts of it will ever depend on any software that is not Free."
I'm pretty sure that's *not* their view. If it is, it's wrong. There are plenty of non-Free licenses that someone cannot ask you to not use. Suppose you have to redistribute non-distributed modifications? That's incompatible with the GPL, so of course Stallman gets twitchy and calls it non-Free. But no one can ask you to "stop using the OS".
If that's not important to you keep using Red Hat, or Gentoo, or rolling your own.
Yeah, because Red Hat is just so anti-Free software.
I suppose you're the type to bitch about the ACLU being a bunch of extremists but posted a "Microsoft sucks" comment when they try to censor Slashdot, eh?
Actually, yes. You can't justify the ACLU's extreme viewpoints by arguing that the particular (IMHO egregious) case of MS censoring Slashdot is out of line. You're taking a slippery slope argument ("well, you don't support the ACLU, so you must support Microsoft censoring Slashdot").
didn't we just have a NetBSD package article [slashdot.org]? Is this news? We don't know how many FreeBSD and OpenBSD packages there are, how many Fink packages or Gentoo packages there are? I don't want to like like a troll but it doesn't seem nessisary
Good point. I demand a Black Cat Linux package count.
When NetBSD finally gets ported to the handheld abacus that Sir Issac Newton used (I made that up).
Forget Slashdot; the Associated Press would pick that one up in a heartbeat.
You're right, though. Even a story detailing the last package packaged each month for NetBSD would be more useful. Someone might discover a new software package.
I hope that you will elect me as the next Debian Project Leader because I am the person best suited to motivating and coordinating the Debian community.
I hold a Master degree in Philosophy and have recently completed a Master of Science in Psychology.
Yup. I'd say that this guy pretty much fits the Debian mold.
I suspect that someone like Linus would simply ignore the long, drawn out threads on licensing that the Debian team loves so dearly. ("Well, *this* package should be in nonfree because it depends on another package that is dual-licensed but has all new contributions donated only to the nonfree license version...").
In which way is a static address not a huge, protocol-independent, world-readable cookie?
It is. But it provides some benefit (a static place to contact me). Using my MAC as the bottom portion of my IP doesn't benefit me at all, and is a drawback.
It also tells the world what type of system you're running (router, Mac, x86 box, SPARC, etc).
Unlike an IP, the MAC bits stay the same from provider to provider and from location to location (admittedly, mostly an issue to laptop owners). This is particularly nasty for laptops that travel from home to secure business locations -- and yes, this is not abnormal in the business world.
It hands out the MAC to anyone on the Internet, which can be nice for MAC-related attacks if a hacker can compromise a nearby system...
As a non-privacy-related but nasty issue, my IP changes if my Ethernet card breaks and I get a new one. People running a server will love that (and "IP numbers unassociated with MACs" become a premium item to sell to business accounts).
Finally, I can *get* a new IP number if I want one today. If my ISP has a policy (and has routers that depend on) my IP ending in my MAC, I'm stuck with it.
The horror.
Heh. I just watched Apocolypse Now for the first time.
Of course there are ways to get a new address from DHCP, most obviously changing the MAC address.
You can't do that on any card that I'm familiar with, though I'm sure there are some that you can finagle into pulling that off on. The Linux approach of "changing the MAC" just kicks the card into promisc mode and then does software filtering when listening for frames with the right MAC. It wastes CPU time.
Object pools are a bad idea in Java, but I can't even count the number of C/C++ programmers who I've heard say thats what's needed to "fix" Java.
I didn't say Java needs to be "fixed", just that it's not a high performance language. And the person who told me that he ended up having to use object pools on his project (urging my comment) to get enough performance was a PhD specializing in computer language design.
Uh, huh. As evidenced by all those high quality horizontal Java apps. Oh, wait.
Considering Java has 54% of the software development market [fawcette.com], I'd say there's probably more Java out there than you think.
Dammit, I said "horizontal Java apps", not vertical market stuff. Yes, most development is in vertical markets (even if most *installed copies* of software and the most money are in horizontal markets). The needs of vertical markets generally put emphasis on short development time, not performance. If you're paying at least tens of thousands of dollars for your code, you don't have a problem putting a few extra thousand into hardware to shorten dev time. That doesn't support the performance argument.
we'd all be writing in assembler
While I agree that there's a point of dimishing returns in dev time/performance, I'd like to point out that it's currently pretty damn hard to outperform a compiler. Maybe for a few instructions, you can pull it off. I remember Rasterman (Enlightenment, imlib) writing how he couldn't manage to outcode gcc any more with hand-tweaked assembly.
However, consider this. How often do you buy or seriously upgrade a computer? Every three to five years is pretty par for the course, and probably on the conservative side. So you're buying a new computer that's got maybe at *most* eight times the processor power, and maybe a little less than than in memory increase. Java software that I've used tends to be in that ballpark next to the closest C equivalents (particularly in memory usage). So basically, you're throwing out that last upgrade you did if you're chosing to use Java-based software.
Development costs, maintainence costs, deployment costs, portability, security....in all these other respects, Java beats C/C++ hands down.
I'll agree when it comes to development costs. Maintenance costs can go either way -- minimal maintenance ("library foo broke") is certainly better with Java. I've found that if you need to add a not-originally-anticipated feature to code down the road, you're much better off with C -- Java (and even C++) put an emphasis on heavy design at the start, which can result in pretty code -- as long as you don't have to change your design much. I hate the term "extreme programming", but the constant revision that it references fits C much better than Java. Deployment costs -- I'd have to say C/C++ wins if you're not using natively-compiled code (because you have an extra dependency on a JVM), and otherwise the two are the same. I'll give you portability and security (though most of the security problems in C code stem from *old* C code with statically-sized arrays...modern C tends to look more object-oriented and have better library functions available to avoid sticky situations -- take a look at glib, which I love dearly).
I tend to be cynical about Java. Java has several notable characteristics. It's quite inefficient when it comes to CPU cycles and memory usage. It's very easy to write distributed apps with it. It has a lot of libraries for common business-task related code. It's also developed and promoted by Sun, a company that makes money by selling hardware. So, yes, Java "scales" well. It's very attractive for business apps, so businesses buy it and use it...and then pay out the nose for the associated hardware platform. More load? Just buy more hardware!
I thought that something inherent to the JVM design made tail recursion difficult, if not impossible. I know an ML fan who was complaining about how unsuitable the JVM was as a target for his language because of the inability to do tail recursion.
Might be something to do with the security model...
Oh, bull and shit. I heard that same line from Sun for *years*. First it was "Yeah, Java may *seem* to be slow as hell, but that's because VMs aren't mature enough". Then it was "Yeah, Java may *seem* to be slow as hell, but that's because you're not using a JIT/native code compiler". Then it was "Yeah, Java may *seem* to be slow as hell, but that's because you're using yesteryear's computer and don't have enough memory". Now it's "Yeah, Java may *seem* to be slow as hell, but that's because you're using bad programming techniques." It's pretty clear to everyone involved except (a) people that have committed to the "Java can be fast" line and (b) Sun salespeople that Java *is* slow as hell. Doesn't mean that it doesn't have good uses, but it's not for high-performance stuff.
We've re-written C applicatons in Java and actually made them run faster by improving the architecture.
Yes. And you could have done the same rearchitecting of the code in C, and had even better performance.
If you try to write a Java app using C/C++ programming techniques, it will suck.
Oh, like having actual allocation and deallocation going on? Yeah, you can get better Java performance if you create all your objects in a huge pool and then just hand around references to the thing. It's also a pain in the ass.
If you take the time to learn how Java is supposed to work, you'll be 10 times more productive and create a product of equal (or better) quality.
Uh, huh. As evidenced by all those high quality horizontal Java apps. Oh, wait.
The only place I see Java being popular is in vertical market stuff, where you need fast development time. Java does speed up development, I'll give it that. It has a nice library set, and a clean syntax.
But when it comes to speed...no, Java is a dog, no matter how you paint it. And uses more RAM than just about anything else.
Why would anyone want to write a serious "enterprise" application in Java vs. say C++??
(a) Most of the time, computers are powerful enough for the task, and using Java shortens development time
(b) Most programmers are pretty piss-poor (especially "business", "database" and "web" programmers), and using Java makes things easier for them.
(c) Freaking big set of libraries.
Obviously, if your thing is writing codecs or high-load servers, Java isn't all that appropriate. But if you're writing a lightweight client or a distributed app, it can be convenient.
That being said, I've never seen an application where I wouldn't prefer to use an equivalent C program.
GC nuts say that you can get good performance with GC. I'm pretty sure that it isn't GC that makes Java so slow -- Ocaml uses GC, and it runs about as quickly as C.
i have heard of ipv6 and have a vague idea of what it is, but could someone elaberate?
A revision of IPv4. The big things it adds (well, that I care about) are:
* More QoS stuff. No one used the IPv4 stuff that was already there, but maybe someone will change their mind, and we'll have tiered bandwidth packages someday ("I want 50 megs of high-prio data/week, 5 gigs of regular/week, and 50 gigs of low-prio data/week...if I exhaust my quota, just kick the packets down to the next prio level") * IPSec built in. All connections can be encrypted, if both hosts feel like it. * Bigger address space. This lets organizations get rid of stupid shit like DHCP/bootp with non-static IPs and NAT. Basically, everyone who wants one can have a static address.
We aren't using it all over because Cisco routers are overpriced, and companies that spent lots of money on an IPv4 router don't want to do the same for an IPv6 router. It is not used much in the US, because of the huge address space allocated to the US. IPv6 is more commonly used in Japan. There are also a number of people tunneling networks of IPv6 machines together over IPv6, which is what things like the 6bone were designed to do.
There aren't really any downs to IPv6 other than the replacement costs. Possibly privacy issues -- there's been interest in using your MAC address as the last bits of your IPv6 address, which seems incredibly stupid to me -- like one huge, protocol-independent, world-readable cookie, but whatever.
So, please, please, PLEASE stop complaining about something that was supposed to be going away from the very beginning!!!
Wakko, we're all dying. Sooner or later, I'm going to be dead, and you're going to be dead.
So what say I just blow your head off right now?
No? You don't like that? You don't like the precise timing, even though you *knew* that sooner or later, you had to die?
Maybe you can understand the viewpoint of the people complaining.
IPv4 won't be around forever. IPv6 probably won't be around forever either -- IP doesn't natively provide for some things like bandwidth allocation or ad hoc networks. The web browser, the operating system you're using...they're all going to be gone, unsupported, decaying, unused and finally forgotten before too long. That doesn't mean you want to say goodbye just yet.
What firewall admin in their right mind would allow users to do end-to-end encryption through a firewall without being able to control the traffic??
Never heard of VPNs?
Besides, you can set up IPSec on IPv4 if you want.
Besides, there's no shortage of IP addresses if IANA would get off their ass and allocate them.
Routing tables are finitely-sized. You can't just run around slicing everything up finely and handing out three addresses here, seven there. Having a routable address with not-hideously-expensive routers means some address space waste.
There are huge class A network yet to be touched and more and more businesses are just finding NAT'ing is easier and more secure anyway.
NAT *easier*? That's a new one.
As for more secure, you can get the same degree of impaired functionality by simply telling your organization firewall "no inbound connections".
Why pay ARIN for address space when you can NAT several thousand people to one or two IP addresses?
But then you'd have a loud computer standing next to your tv and have a kludgy interface that probably would make you have a keyboard there too.
I've always felt that TVs look terrible. Blurry and icky, shadowing...bleh. DVDs on the computer monitor, that's the way to go. Then I have my big leather computer chair with maybe some snacks...mmm....
I strongly suspect that the C standard provides no guarantees about whether free() coallesces free chunks or not, regardless of whether some implementations do so.
They also helped me for free when I couldn't get Visual Studio.NET to install. About a thirty minute call there, including my callback to the support engineer.
Why did it take thirty minutes to figure out how to install VS.NET?
Come on, without fans of programs, the world would be lots less exciting.
Although it's not impossible to do, I still maintain that admins should patch their systems, but you don't have to rush. I don't see script kiddies exploting this one in the coming time yet. And besides, my data isn't worth crap either, so I'm harly a target.
Yeah, I used to say the same thing until I had a box get broken into.
Slashdot Mirror
Is the music industry really so dumb as to think that hardware and software solutions will really ever work?
Yes. They have many somewhat unethical DRM companies that smell money assuring them that yes, their scheme absolutely *will* work, or at least will work in a few years.
Frankly, I think they'd have better luck fingerprinting songs and refusing to play fingerprinted songs on a device that doesn't have rights to play it, and then shoot for a 2020 timeframe for having these legally required. Then you force every person who wants to listen to pirated music owning an illegal piece of hardware. Sure, some would, but it's like cable piracy. The losses are manageable.
The current schemes, which break if a single person anywhere manages to break the protection, are pretty silly.
Shipping songs lossily compressed (non-MP3 format) is a good first step. It makes MP3s or Oggs sound worse, doubly lossily compressed. Of course, trying to convince the consumer to buy worse-sounding music is going to be interesting...maybe if they introduce smaller media and market it as sexier, or find some other special features to add in. I dunno, maybe slap the instruments on separate tracks, so you can choose where to place the drums and whatnot when listening ("balance 5% to the left, tiny delay, slight amplitude cut on the snare drumps so they sound like they're off to the left"). Maybe embed lyrics in the new format.
It's a hell of a thorny problem, and you *know* consulting firms are doing anything they can to extract money from a wealthy and desperate industry.
Put up a monument to this day -- someone proposed a Slashdot boycott that might *happen*!
That's one of the stupidest ad homonems I've ever seen on Slashdot, and that's saying something.
Frankly, I think a lot of people would hardly see that as an insult. If I say "most rabbis love long, complicated arguments about interpretations of religious law, though most people don't care in the least about such questions", I think you'd find a lot of rabbis and theologians that would agree. So?
Obviously, I really am not interested in spending hours of my time arguing over the particular classification of a package. It's pretty obvious that my viewpoint is not that of these folks. I'm simply pointing out that people on the Debian ML do do this, and that most people wouldn't. This guy is obsessed with the ideology, rather than the tech. You really can't argue that point.
If there weren't a number of geeks very concerned about things like licensing we wouldn't have Linux in the first place. We might have a nice kernel, but that's a long shot from a Free OS.
Besides the fact that that's completely irrelevant to what I said (I don't believe I said that people obsessed with legalities shouldn't exist), I think most people wouldn't care. What if FreeBSD was the big, popular, multi-distro OS that Linux is today. Would most of us *care* whether the kernel was BSD-licensed instead of GPL licensed? I sure as hell wouldn't.
Since it looks like this is getting into ideology, and you feel arguing ideological technicalities is fun, I'll play the game. I'll take ESR's viewpoint. ESR thinks that Open Source is better because it's *works better*. RMS thinks that Free software is better because it's ideologically distasteful to use non-Free software.
Logically, if open source software is superior, closed-source attempts to compete with it should fail (we'll assume perfect consumer information...naive model, but unless you really think that open source can be toppled by marketing...). So if MS takes the BSD kernel and tries to make a superior fork (closed source), they should lose out to the open source fork.
Therefore, I claim that the only people who claim that the GPL is necessary (rather than the BSD license) are those who (a) feel that open source software is inferior to closed source software, (b) feel that open source is so marginally better than closed source software that a simple influx of marketing will forever bar open source software from the market, or (c) are irrational.
Chew on that.
Debian's view is pretty simple: "If the software we use isn't Free, then someone can legally ask us to stop using it. Therefore, our operating system and its tools will always be Free, and no parts of it will ever depend on any software that is not Free."
I'm pretty sure that's *not* their view. If it is, it's wrong. There are plenty of non-Free licenses that someone cannot ask you to not use. Suppose you have to redistribute non-distributed modifications? That's incompatible with the GPL, so of course Stallman gets twitchy and calls it non-Free. But no one can ask you to "stop using the OS".
If that's not important to you keep using Red Hat, or Gentoo, or rolling your own.
Yeah, because Red Hat is just so anti-Free software.
I suppose you're the type to bitch about the ACLU being a bunch of extremists but posted a "Microsoft sucks" comment when they try to censor Slashdot, eh?
Actually, yes. You can't justify the ACLU's extreme viewpoints by arguing that the particular (IMHO egregious) case of MS censoring Slashdot is out of line. You're taking a slippery slope argument ("well, you don't support the ACLU, so you must support Microsoft censoring Slashdot").
There, that ideology-arguing enough for you?
didn't we just have a NetBSD package article [slashdot.org]? Is this news? We don't know how many FreeBSD and OpenBSD packages there are, how many Fink packages or Gentoo packages there are? I don't want to like like a troll but it doesn't seem nessisary
Good point. I demand a Black Cat Linux package count.
When NetBSD finally gets ported to the handheld abacus that Sir Issac Newton used (I made that up).
Forget Slashdot; the Associated Press would pick that one up in a heartbeat.
You're right, though. Even a story detailing the last package packaged each month for NetBSD would be more useful. Someone might discover a new software package.
And besides, among the notable additions were "xmms-funtimedancer." How did NetBSD users survive without a funtime dancer?
:-)
They haven't been. NetBSD has been dying for lack of a funtime dancer.
If it exists for Linux, it's probable that it's packaged in RPM format.
Of course, for both deb and rpm, there's the excellent checkinstall
I hope that you will elect me as the next Debian Project Leader because I am the person best suited to motivating and coordinating the Debian community.
I hold a Master degree in Philosophy and have recently completed a Master of Science in Psychology.
Yup. I'd say that this guy pretty much fits the Debian mold.
I suspect that someone like Linus would simply ignore the long, drawn out threads on licensing that the Debian team loves so dearly. ("Well, *this* package should be in nonfree because it depends on another package that is dual-licensed but has all new contributions donated only to the nonfree license version...").
In which way is a static address not a huge, protocol-independent, world-readable cookie?
It is. But it provides some benefit (a static place to contact me). Using my MAC as the bottom portion of my IP doesn't benefit me at all, and is a drawback.
It also tells the world what type of system you're running (router, Mac, x86 box, SPARC, etc).
Unlike an IP, the MAC bits stay the same from provider to provider and from location to location (admittedly, mostly an issue to laptop owners). This is particularly nasty for laptops that travel from home to secure business locations -- and yes, this is not abnormal in the business world.
It hands out the MAC to anyone on the Internet, which can be nice for MAC-related attacks if a hacker can compromise a nearby system...
As a non-privacy-related but nasty issue, my IP changes if my Ethernet card breaks and I get a new one. People running a server will love that (and "IP numbers unassociated with MACs" become a premium item to sell to business accounts).
Finally, I can *get* a new IP number if I want one today. If my ISP has a policy (and has routers that depend on) my IP ending in my MAC, I'm stuck with it.
The horror.
Heh. I just watched Apocolypse Now for the first time.
Of course there are ways to get a new address from DHCP, most obviously changing the MAC address.
You can't do that on any card that I'm familiar with, though I'm sure there are some that you can finagle into pulling that off on. The Linux approach of "changing the MAC" just kicks the card into promisc mode and then does software filtering when listening for frames with the right MAC. It wastes CPU time.
I can't understand anybody complaining when they've got three years notice...
Grr...okay, let's extend the analogy.
"Evil Adrian, I will blow your head off in three years."
See?
Object pools are a bad idea in Java, but I can't even count the number of C/C++ programmers who I've heard say thats what's needed to "fix" Java.
I didn't say Java needs to be "fixed", just that it's not a high performance language. And the person who told me that he ended up having to use object pools on his project (urging my comment) to get enough performance was a PhD specializing in computer language design.
Uh, huh. As evidenced by all those high quality horizontal Java apps. Oh, wait.
Considering Java has 54% of the software development market [fawcette.com], I'd say there's probably more Java out there than you think.
Dammit, I said "horizontal Java apps", not vertical market stuff. Yes, most development is in vertical markets (even if most *installed copies* of software and the most money are in horizontal markets). The needs of vertical markets generally put emphasis on short development time, not performance. If you're paying at least tens of thousands of dollars for your code, you don't have a problem putting a few extra thousand into hardware to shorten dev time. That doesn't support the performance argument.
we'd all be writing in assembler
While I agree that there's a point of dimishing returns in dev time/performance, I'd like to point out that it's currently pretty damn hard to outperform a compiler. Maybe for a few instructions, you can pull it off. I remember Rasterman (Enlightenment, imlib) writing how he couldn't manage to outcode gcc any more with hand-tweaked assembly.
However, consider this. How often do you buy or seriously upgrade a computer? Every three to five years is pretty par for the course, and probably on the conservative side. So you're buying a new computer that's got maybe at *most* eight times the processor power, and maybe a little less than than in memory increase. Java software that I've used tends to be in that ballpark next to the closest C equivalents (particularly in memory usage). So basically, you're throwing out that last upgrade you did if you're chosing to use Java-based software.
Development costs, maintainence costs, deployment costs, portability, security....in all these other respects, Java beats C/C++ hands down.
I'll agree when it comes to development costs. Maintenance costs can go either way -- minimal maintenance ("library foo broke") is certainly better with Java. I've found that if you need to add a not-originally-anticipated feature to code down the road, you're much better off with C -- Java (and even C++) put an emphasis on heavy design at the start, which can result in pretty code -- as long as you don't have to change your design much. I hate the term "extreme programming", but the constant revision that it references fits C much better than Java. Deployment costs -- I'd have to say C/C++ wins if you're not using natively-compiled code (because you have an extra dependency on a JVM), and otherwise the two are the same. I'll give you portability and security (though most of the security problems in C code stem from *old* C code with statically-sized arrays...modern C tends to look more object-oriented and have better library functions available to avoid sticky situations -- take a look at glib, which I love dearly).
I tend to be cynical about Java. Java has several notable characteristics. It's quite inefficient when it comes to CPU cycles and memory usage. It's very easy to write distributed apps with it. It has a lot of libraries for common business-task related code. It's also developed and promoted by Sun, a company that makes money by selling hardware. So, yes, Java "scales" well. It's very attractive for business apps, so businesses buy it and use it...and then pay out the nose for the associated hardware platform. More load? Just buy more hardware!
I thought that something inherent to the JVM design made tail recursion difficult, if not impossible. I know an ML fan who was complaining about how unsuitable the JVM was as a target for his language because of the inability to do tail recursion.
Might be something to do with the security model...
Java will run as fast or as slow as you make it.
Oh, bull and shit. I heard that same line from Sun for *years*. First it was "Yeah, Java may *seem* to be slow as hell, but that's because VMs aren't mature enough". Then it was "Yeah, Java may *seem* to be slow as hell, but that's because you're not using a JIT/native code compiler". Then it was "Yeah, Java may *seem* to be slow as hell, but that's because you're using yesteryear's computer and don't have enough memory". Now it's "Yeah, Java may *seem* to be slow as hell, but that's because you're using bad programming techniques." It's pretty clear to everyone involved except (a) people that have committed to the "Java can be fast" line and (b) Sun salespeople that Java *is* slow as hell. Doesn't mean that it doesn't have good uses, but it's not for high-performance stuff.
We've re-written C applicatons in Java and actually made them run faster by improving the architecture.
Yes. And you could have done the same rearchitecting of the code in C, and had even better performance.
If you try to write a Java app using C/C++ programming techniques, it will suck.
Oh, like having actual allocation and deallocation going on? Yeah, you can get better Java performance if you create all your objects in a huge pool and then just hand around references to the thing. It's also a pain in the ass.
If you take the time to learn how Java is supposed to work, you'll be 10 times more productive and create a product of equal (or better) quality.
Uh, huh. As evidenced by all those high quality horizontal Java apps. Oh, wait.
The only place I see Java being popular is in vertical market stuff, where you need fast development time. Java does speed up development, I'll give it that. It has a nice library set, and a clean syntax.
But when it comes to speed...no, Java is a dog, no matter how you paint it. And uses more RAM than just about anything else.
Why would anyone want to write a serious "enterprise" application in Java vs. say C++??
(a) Most of the time, computers are powerful enough for the task, and using Java shortens development time
(b) Most programmers are pretty piss-poor (especially "business", "database" and "web" programmers), and using Java makes things easier for them.
(c) Freaking big set of libraries.
Obviously, if your thing is writing codecs or high-load servers, Java isn't all that appropriate. But if you're writing a lightweight client or a distributed app, it can be convenient.
That being said, I've never seen an application where I wouldn't prefer to use an equivalent C program.
GC nuts say that you can get good performance with GC. I'm pretty sure that it isn't GC that makes Java so slow -- Ocaml uses GC, and it runs about as quickly as C.
i have heard of ipv6 and have a vague idea of what it is, but could someone elaberate?
A revision of IPv4. The big things it adds (well, that I care about) are:
* More QoS stuff. No one used the IPv4 stuff that was already there, but maybe someone will change their mind, and we'll have tiered bandwidth packages someday ("I want 50 megs of high-prio data/week, 5 gigs of regular/week, and 50 gigs of low-prio data/week...if I exhaust my quota, just kick the packets down to the next prio level")
* IPSec built in. All connections can be encrypted, if both hosts feel like it.
* Bigger address space. This lets organizations get rid of stupid shit like DHCP/bootp with non-static IPs and NAT. Basically, everyone who wants one can have a static address.
We aren't using it all over because Cisco routers are overpriced, and companies that spent lots of money on an IPv4 router don't want to do the same for an IPv6 router. It is not used much in the US, because of the huge address space allocated to the US. IPv6 is more commonly used in Japan. There are also a number of people tunneling networks of IPv6 machines together over IPv6, which is what things like the 6bone were designed to do.
There aren't really any downs to IPv6 other than the replacement costs. Possibly privacy issues -- there's been interest in using your MAC address as the last bits of your IPv6 address, which seems incredibly stupid to me -- like one huge, protocol-independent, world-readable cookie, but whatever.
So, please, please, PLEASE stop complaining about something that was supposed to be going away from the very beginning!!!
Wakko, we're all dying. Sooner or later, I'm going to be dead, and you're going to be dead.
So what say I just blow your head off right now?
No? You don't like that? You don't like the precise timing, even though you *knew* that sooner or later, you had to die?
Maybe you can understand the viewpoint of the people complaining.
IPv4 won't be around forever. IPv6 probably won't be around forever either -- IP doesn't natively provide for some things like bandwidth allocation or ad hoc networks. The web browser, the operating system you're using...they're all going to be gone, unsupported, decaying, unused and finally forgotten before too long. That doesn't mean you want to say goodbye just yet.
One of the main problems with it is security.
This should be good.
What firewall admin in their right mind would allow users to do end-to-end encryption through a firewall without being able to control the traffic??
Never heard of VPNs?
Besides, you can set up IPSec on IPv4 if you want.
Besides, there's no shortage of IP addresses if IANA would get off their ass and allocate them.
Routing tables are finitely-sized. You can't just run around slicing everything up finely and handing out three addresses here, seven there. Having a routable address with not-hideously-expensive routers means some address space waste.
There are huge class A network yet to be touched and more and more businesses are just finding NAT'ing is easier and more secure anyway.
NAT *easier*? That's a new one.
As for more secure, you can get the same degree of impaired functionality by simply telling your organization firewall "no inbound connections".
Why pay ARIN for address space when you can NAT several thousand people to one or two IP addresses?
Because NAT is a PITA and an utter hack?
But then you'd have a loud computer standing next to your tv and have a kludgy interface that probably would make you have a keyboard there too.
I've always felt that TVs look terrible. Blurry and icky, shadowing...bleh. DVDs on the computer monitor, that's the way to go. Then I have my big leather computer chair with maybe some snacks...mmm....
I strongly suspect that the C standard provides no guarantees about whether free() coallesces free chunks or not, regardless of whether some implementations do so.
I was about to post a nasty post about how crummy Microsoft is -- a *third party* is charging money to fix their lousy memory management system.
Then I remembered Ram Doubler on the Mac.
Proprietary OSes. Sigh.
They also helped me for free when I couldn't get Visual Studio .NET to install. About a thirty minute call there, including my callback to the support engineer.
.NET?
Why did it take thirty minutes to figure out how to install VS
Geez, it's only a mail server.
And it's only an editor.
And it's only an operating system.
Come on, without fans of programs, the world would be lots less exciting.
Although it's not impossible to do, I still maintain that admins should patch their systems, but you don't have to rush. I don't see script kiddies exploting this one in the coming time yet. And besides, my data isn't worth crap either, so I'm harly a target.
Yeah, I used to say the same thing until I had a box get broken into.
Sarcasm can be insightful.
Send the above post to the qmail/postfix mailing lists and see what happens.
:-)
I can tell you that postfix definitely works with procmail.
I can't understand why any general-purpose distros still ship sendmail. Qmail is good too, though I prefer postfix.
Sendmail takes (on my system) a thousand-line config file just to have sane settings for the modern world. It has a horrendous security history.
Postfix has non-dumb defaults, is quite secure, and I cannot see why anyone wouldn't use it.