I've been playing computer games regularly for about the last sixteen years, and I can happily say that I can play StarCraft and Diablo II all day. I love playing Unreal Tournament, System Shock II and the Final Fantasy genre.
Pac-man, on the other hand, bores the pants off me. Getting three billion points in it seems to be more an exercise in how to maintain concentration in the face of near-terminal monotony.
Space Invaders likewise. MAME held my interest for all of three days, before the nostalgia value wore off, and I started to wonder just how easily amused I must have been when I was fifteen.
There are _some_ games I'd like to play through again (the original C-64 Wizball springs to mind, although that's probably a little "post-classic"), but a lot of classic gaming seems to just be nostalgia for things that were only good because they were the best that could be done at the time.
Charles Miller (Woo, gonna lose karma over this one...) --
"Why couldn't there be a competing domain structure? It could use differently numbered TCP ports to communicate so as not to conflict with the existing one, and be somewhat gnutella-like to avoid centralization."
DNS is already "somewhat gnutella-like". The reason we have the TLD's that we do is because by convention, most DNS servers address the same set of root nameservers, which define what TLD's are available. It is a simple act to set up a root nameserver that will serve additional TLD's, the only hurdle you face is getting other people to add you to their nameserver configurations. People are doing this all the time, but I'm too lazy to find a URL pointing them out.
Some time in the not-so-distant future, I predict that countries will start setting up their own root nameservers, and enforcing their use within the country by legislation. This will be done for two reasons - firstly to stop the net being so reliant on servers within the USA for everyday operations, and secondly because it is the only way I can see to overthrow the current system, where countries are forced to pay a US organization for the privilege of having control of their own country-name TLD.
My via-Palo-Alto experience was when I was living in Perth. AARNET had just switched from Telstra to Optus, but my home box was hooked up to Telstra. The telcos hadn't even bothered to set up peering, they were just throwing everything at the international link. To some extent this still happens. The telcos have POPs in all major cities, but they only bother to peer with each other in Melbourne and Sydney, AFAIK, so the same packets would even today travel an 8000km round-trip in order to cover 15km of real-world distance.
ISPs complicate the issue even more, because routing in commercial ISPs is usually set up based on which link costs the most per megabyte, rather than which actually gives you the fastest route to your destination. If Telstra charges you 19c/meg incoming, but your satellite is feeding you at 7c/meg, you'll end up broadcasting your routes only via the satellite, then using the Telstra link for back-channeling. (Except now Telstra has a back-channeling charge, so unless you have an exemption you have to calculate your routes carefully to stay under the threshhold...)
This tends to disadvantage users, since satellites have a bit of a latency problem, but what they don't know doesn't hurt them.
The short summary is that routing is, and will continue to be totally insane until bandwidth gets cheap, and telcos get a clue. Oh, and then we could use the flying pigs to carry some of the packets!
You wouldn't need to take down the entire US for that. I'm willing to bet that you could bring down the Internet with two or three well-chosen routers (i.e. in a couple of different major backbone NOCs) set up to transmit bogus routes.
Firstly, there's a big difference between what Cringeley suggests (the ability to shut down ISPs) and shutting down all the backbones. Taking down all the ISPs in the USA, but leaving the backbones running would make the net several orders of magnitude faster for the rest of us. (and several orders of magnitude more boring) However, if the backbones were taken down...
Living in Sydney Australia, pretty much all of my routes go through the USA, except those to very close neighbours such as Malaysia and Indonesia. My routes to Japan and Taiwan go via the USA. South Africa is closer to Perth, Australia than I am. My packets to South Africa go to Perth, THEN to the US, THEN to.za.
Sometimes it's even worse than that. Back when I was at University, it was so bad that when I did a traceroute between two servers 15 minutes drive apart but on different backbones, the packets were going via California.
There are links between countries that could be used if the USA were to vanish, but these links are usually significantly underpowered. Most of the major content providers are in the USA, most of the packets go to and from the USA, so other countries tend to invest most of their money in fat pipes to North America. And since those fat pipes are already there, they may as well take care of some of the local traffic as well.
Between countries on the same continent, you're probably looking at a continuing stable network. But inter-continental links would most likely fall over and die.
Even if the underpowered inter-continental links could take it, you'd see a routing nightmare. BGP packets would be flying around in circles panicking, and any sane network administrator would lock him or herself in a small room and whimper until it was all over.
There's also other things to think of. How many of the root nameservers are outside the USA? How much traffic can they take? How would they cope with the prolonged absence of a.root-servers?
I'm going to lose karma for this one, I know it...
++"I guess this kind of joke is why women don't want to go into IT business"
This attitude annoys me no end. I'm all for political correctness, but this is the kind of blinkered, divorced from reality, humourless, reactionary comment that gives its detractors so much ammunition.
In the course of working several jobs, I have worked in male-dominated, female-dominated and evenly mixed workplaces. I've been exposed to men making sexual jokes about women, women making sexual jokes about men, and homosexuals making sexual jokes about each other. In my last job at an ISP, while there was only one female tech, the most common source of comments of the kind you're complaining about was from the receptionist and accounts staff, who were female.
If someone making a joke about the fact they're attracted to whatever gender they're attracted to offends you, then you're going to spend your life perpetually offended, wherever you work. Unless, of course, it's one of those things that's okay for women to do, but not for men.
Yes, it goes too far if people start wallpapering their offices with centerfolds, or if the joke is non-consentually at the expense of, or directed at, a colleague. But while I've seen that happen, I saw much more blatant examples of it outside IT than in, and I'd be willing to bet it's far more prevalent (proportional to women employed) outside the computer industry.
On a similar note, from a previous post by you in a different story:
"Guys tend to have few social skills, like giving feedback, discussing in a fair manner, self-reflecting. They don't care about themselves, their rooms, others."
How about you clear up your own rabid prejudices, before you start harping on the shortcomings of others?
"This just proves that Mozilla is too bloated as well."
When the Mozilla project started, the requirements were that the resulting software be highly portable, compliant with a whole slew of web standards, and be a superset of the functionality of Netscape Communicator. Thus, it was doomed to be bloated before it even began. It was doomed to be bloated because that's what people WANTED it to be.
To quote JWZ (although he was referring to the pre-opensource Mozilla), "Mozilla is big because your needs are big. Your needs are big because the Internet is big. There are lots of small, lean web browsers out there that, incidentally, do almost nothing useful. If that's what you need, you've got options..."
Mozilla's solution to its own bloat-by-design was, IMHO, rather elegant. It took stock of the what it was required to implement, and then used those technologies to build other components. This isn't anything particularly new, of course emacs invents itself from its lisp interpreter.
To fulfil its goal of standards compliance, Mozilla had to understand HTML, XML, Javascript, DOM and CSS. To maintain a suite of applications (browser, mail/news, html authoring tool) across widely differing platforms, it needed its own flexible, cross-platform UI and Networking libraries.
So Mozilla's self-reinvention as a platform, rather than an application was the result of necessity, and some pretty good thinking from the designers rather than just bloat for the sake of bloat. It was actually the most efficient way that it could perform the task it had been given.
When I have moderator points, which admittedly isn't often, I reserve "offtopic" for blatant cases where a post bears no relation to the surrounding posts. Posts that are meta-topic, or posts that are the result of natural thread-drift shouldn't be marked as 'offtopic'.
For example, this post has absolutely nothing to do with penguins, but it's perfectly at home in this particular thread. Having said this, of course, someone's going to come across and mod me down.:P
So my question is: how appropriate is it for people to use their website as a medium to raise consciousness and solicit donations for whatever cause they are currently supporting?
It is entirely appropriate. Slashdot's mandate is to post stories that interest the editors. It has no responsibility beyond that. "Hey, cool, the penguins need our help" is just as valid a story as "Hey, cool, the DeCSS authors need our help." If there's anyone out there who's so impressionable that they'll think "Hey, Hemos said I should donate, where's my cheque book!" then they deserve our pity.
Slashdot has become popular because the story selections that the editors make is generally of interest to its community of readers. That doesn't mean that every story has to interest every reader, I certainly don't read more than 25% of the stories myself. All it means is that on an average day, there are enough slashdot stories that interest me for me to return the next day. I read the penguin story because it was different.
If you watch a news-hour on TV, there's usually a cute animal story or something of that ilk. They're not there because they're news, they're there because they're _relief_ from news. Similarly, the penguin story is a relief from Yet Another Copyright Article, or Yet Another Obscure Technology.
Follow-up: does it matter if the purpose of the website typically excludes stories about that particular topic, moving though they may be?
You are working from an incorrect assumption. You're assuming that the "purpose" of this website is more concrete than it is. "News for Nerds" doesn't mean "News about technology", although there is a correlation between the two.
And another one: are there some issues which are so important that they should headline any/all forms of media, regardless of the subject matter usually addressed by that media? If so, is this one of those topics?
I doubt the injury of penguins will "headline" news worldwide. It headlines slashdot purely because the penguin-linux connection makes it a fun real-news/nerd-news crossover. Similarly, if there were a bunch of BSD daemons being held hostage in the middle-east, I expect that would show up on the slashdot front page too.
My greatest impression from reading the paper was one of a schizophrenic system. The differences between the two systems that form the basis for OSX seemed not to have been resolved, instead they were patched over with an additional layer of complexity, and a great deal of hope applied that the two different OS's at the core wouldn't misbehave and contradict each other.
Usually, this is the sort of thing that makes software developers run screaming down to the pub.
Larry Wall tells an amusing story of how he "hacked the licesnse war" by distributing Perl under the GPL and the Artistic License at the same time. Now if only I could remember the URL...
I don't think it'd be possible to run it under mod_perl. The CGI component of minivend is a C program (vlink?) that just forwards the http request to the Perl daemon running as the back-end. I pretty much came to the conclusion that the whole thing is a kludge. I was running it on a PII w/ 256Meg, and the delay was noticeable with a single client requesting pages. I shudder to think what sort of response serious customers are getting. (I don't work at that company any more). I'd seriously recommend trialling the software before committing yourself to implementing it anywhere. My recommendation is that if you're going the free software route, write your own implementation specifically for the site you're developing, in whatever language you're comfortable with. You won't lose nearly as much development time as you think, because the Minivend markup language is so evil, and your site won't run like a dog. Charles Miller --
I developed one or two sites in Minivend. I found it to be mind-numbingly slow, and its in-page markup/scripting language was worse than anything I've seen before or since, and that's including Net.Data. Minivend's scripting language is a series of hacks piled upon hacks piled upon hacks. Getting it to do even the simplest of things is an exercise in pain-endurance.
If it's a choice between Minivend and nothing, I'd probably go with nothing. All you really need to replace the vast majority of its 'features' is an in-page scripting language with decent session support.
I'm afraid, in this case, saying that it beats the closed-source solutions is just demonstrating that you've never compared its features or performance with what's being sold at the moment.
(full disclosure: I work for a company that develops apps based on IBM Java stuff.)
<I>Just remember, folks. 15 years ago, IBM was the oppressor of the tech industry. The only reason they aren't as powerful today is because they got too big and unresponsive, which led them to make several bad decisions.</I>
Actually, the major reason they aren't as powerful today is because they lost an anti-trust suit. This is one of the things that people tend to forget in examining the Microsoft trial -- Bill Gates' billions are largely a consequence of IBM losing a very similar case. If it weren't for IBM's consent decree, they'd have been much more able to screw over the PC industry before it even started. Or at least, Microsoft would be a division of IBM, and we'd be running "PC-OS/2000".
<I>Look at what they're doing with Sun and Java. After working with Sun for 4 years, they refuse to ratify Sun's version of EJB, in effect leaving Sun out to dry. IBM was given access to all of Sun's java IP and this is how they repay them?</I>
You're misrepresenting things just a tad.
The relationship between IBM and Sun over Java wasn't one-way. IBM were responsible for a significant proportion of the Java2 Enterprise spec. They have done more to evangelise Java than any other company apart from Sun itself.
Sun don't like IBM because Sun want Sparc/Solaris to be the platform of choice for Java apps, and here comes IBM porting the language to their own line of Risc boxen, and making a JDK for Win32 and Linux that's faster than Sun's. This is competition done right. If you think about it, it's the same thing Compaq did to IBM in the PC hardware industry. Don't embrace and extend, make a superior product that's 100% compatible.
Sun's response was to "take their toys and go home" with regards to Java standardization, and to start charging license fees to people just for the right to say "this is J2EE". Sun left _themselves_ out to dry.
To get back on the subject: The GPL protects Linux from being screwed too radically by IBM. IBM are getting involved in Linux for the same reason they're getting involved in Java -- they've identified it as a technology that people want to use, so it's natural that they want to use it to sell hardware, and it's natural they want to port their proprietary products to run on it.
And if they're doing that (Although they're way behind on Websphere and VA Java.) And at the same time, it makes sense for them to improve the platform so that it will scale up to their larger boxen. They want to be able to say "Hey, your business is going great now, if you upgrade to this RS/6000, you can still run all your Linux stuff on it", but to do that effectively, Linux needs a few of the "enterprise-level" features that people running Wintel servers don't miss much, like Logical Volume Managers.
Which is cool enough. And in the meanwhile, if you have a boss who wants the warm fuzzies that come with buying something with a brand name, you can get a Netfinity, put Linux on it, and say it's an IBM solution. This is the thought behind things like their "Linux Server Proven" scheme -- you prove that your app runs on a Netfinity using their supported Linux configurations[1], they give you a sticker, and endorse you to their customers.
It takes us a step closer to the day when the Linux success stories _aren't_ all of the form "I switched to Linux over the weekend and my boss didn't notice!"
Charles Miller
[1] They support (off the top of my head) RedHat, SuSe, and I think Corel and TurboLinux. The rationale behind this is that three of them cover the "most popular distro" of North America, Europe and Japan (Asia?) respectively, and the fourth is probably there because it's "most suit-friendly". --
SSL is supported on Windows and Linux/Intel. Java is only supported on Windows currently, due to the fact that until recently, Windows was the only platform to have a JDK with an OJI plugin.
Shockwave is an abomination that should be wiped off the face of the web.
YM "five years" HTH. After the appeals court, it'll go to the Supreme court.
Five years is a hell of a long time in the software business. I mean, five years ago Linux was celebrating version 1.0, and the General Public was deciding whether they should upgrade from Win3.1 to Win95, and wondering if they should get on "That AOL thing." In five years, we saw Microsoft use its OS leverage to kill Netscape, and crap on who knows how many OEMs?
This means that Microsoft still have a lot of time to do a lot of damage. And if they think they've got a good chance of losing the case, it gives them five years to use their Monopoly to make sure that if they're broken up, both resulting companies end up as powerful as they can make them.
I mean, can you see any other rationale behind their new policy of no install media for OEM Win9x licenses? It's like they're saying "We've only got a few years of the gravy train left. Lets stretch it out as long as we can, and screw as many people as we can along the way."
The article linked to above is a pretty shaky foundation on which to base a criticism of JSP. It's obvious that the author is enamoured with his pet language, to the extent that he thinks saving a few keypresses equates to increased efficiency in the Real World.
Let's look at the criticisms one by one:
"Java code too tempting."
This is true of any scripting language. I've been forced to clean up amateur ASP projects that were 90% inline VBScript, 10% HTML, 0 VB objects. PHP3 was even worse, because you were almost forced to inline everything (which put me off PHP to the extent that I never bothered to look at whether v4 fixed that problem).
The proposed solution, however, removing the _ability_ to inline code, is far too restrictive. Why should my proof-of-concept JSP need to be coded for maintainability? If I've got a deadline tomorrow, I might _want_ to sacrifice good coding style, secure in the knowlege I can split the load properly between servlet and JSP over the next week, so long as the app is running. In the real world, there are too many situations where you want to inline code, although you should plan on fixing it later if the page is anything more than a one-shot.
"Java Code Required"
There is no real advantage to being able to type "$Request.ContextPath" instead of "Request.getContextPath()", unless you count three keypresses as critical to your project. They are two different syntaxes for the same thing. This is like saying that WebMacro is a bad thing because it forces you to use WebMacro.
The remaining complaints in that section are solved by using custom tags.
"Simple Tasks are Hard"
Once again, "#parse "header.wm"" is just as complicated as "<%@include file="header.jsp"%>". The only difference is that the latter is a lot less likely to upset GUI web design tools, which makes it easier to keep the web-designers happy.
"Lousy Looping"
The JSP/taglib example of "foreach" is identical in complexity to the WebMacro version, once more with the advantage of being more friendly to existing web design applications.
"Useless Error Messages"
That's an implementation issue, not a problem with JSP's themselves. A decent JSP compiler should spot constructs that will cause errors in a compiled class, and flag them. If your JSP compiler is not decent, it's the fault of the people who wrote the compiler, not of the JSP spec.
As an aside, the setup I have here allows me to step through the JSP compilation process, and attach a debugger to running servlets and JSP's. Being able to set breakpoints, walk through methods etc. etc. makes it a lot easier to find the more obscure of my mistakes.
"Need a compiler"
So how do you run a dynamic site without a compiler (or an interpreter, if you want to be picky)? Osmosis? Telepathy?
It's not particularly hard to get hold of a Java compiler. They're available for free (beer) on most platforms (Solaris, NT, Linux, AIX, HP-UX, MacOS, we've even got one client running an application server off an AS/400). I'll admit, if you're not running on a platform with an available JDK, then JSP isn't for you.
"Wasted space"
The disk-space argument isn't even worth mentioning. You're never going to have so many JSP's that it'll cost you more than a few dollars worth of hard drive space, if that.
Sure, Java uses up a lot of memory. But that's Java itself, not something specific to JSP. I don't know what system he was using to test on, but the average size of a JSP on the production web-app closest to hand is 6-7K, not 30+K. The JSP compiler doesn't load all the HTML as static class data, it gets pulled off the disk at the start of the service() method, and is available for garbage-collection as soon as the request is done.
IBM flirt with Free Software occasionally, but it's not the way they generally do business. They give a lot of stuff away, in a free beer sense, but only because it can get you in a position to buy some of the stuff that makes them money - things like DB2, etc. But in a Free Speech sense, the only things that spring to mind are their collaborations with the Apache group in XSLT stuff, and the Jikes compiler.
More power to them, I guess. I'm not RMS. Whether a company frees its code is entirely up to the company. Their half-hearted Linux support disappoints me, on the other hand. Websphere's still a major version behind, and you can't get VAJava Enterprise Edition at all.
Alphaworks is a special case. The licenses are stricter than production products because they're _not_ production products. It's alpha software. You get to play with it, but on the understanding that next month it may suddenly have become a commercial product, or it may no longer be being developed at all. So the site is neat if you want to scrounge new toys from the bleeding edge, but don't use it for anything serious.
[offtopic] My favourite IBM free-beer package at the moment is their developer program where you can book in to spend time testing your apps on Big IBM Boxen. It looks pretty cool for those of us who can't afford an AS/400, or who don't have the resources to do huge load tests. --
This was an injunction. You take out injunctions after you've told the guy to stop some other way. Ebay got no damages at all, all they did was assert their right to say "Don't do that again."
Noone's going to be able to sue you because you visited a site, and then learned after the fact that you weren't allowed to. All they can do is tell you not to visit it again.
The "using CPU cycles" is not a new legal argument. Back when there were no specific laws against computer cracking, I believe one of the earliest convictions came from proving that a cracker "stole electricity" by using cycles on the machine they'd broken into.
You'll first note, from following the provided link, that this was simply an injunction. There was no mention of damages. There is nothing in this that sets any precedent against webcrawlers, or deep linking. All the case says is that yes, the owners of a webserver may choose how that server is used.
Bidders' Edge were doing something to Ebay's site that Ebay didn't want them to do. Ebay asked them to stop. Bidders' Edge didn't want to stop. So it went to court. The court told Bidders' Edge that what they were doing constituted a use of the Ebay system (their CPU cycles) against the wishes of its owner, and ordered them to stop.
Now what, exactly, is wrong with that? Do we really want to set the alternative precedent, which is that as soon as you put a public resource on the net, you have no say in how that resource is used? I can imagine all the IRC script kiddies drooling at the thought of that one. "Hey! It's a public server, my clonebots were only a few hundred out of the thousands of users online, so I can't have been using a significant amount of the net..."
Just because a resource has been made available to the general public, DOES NOT mean that the owners of that resource shouldn't retain every right to be able to bar people from using it, for whatever reason.
The ability to say "You are no longer permitted to use my server in that way, stop it." is an essential one. It means, for example, that even though slashdot is open for public posting, Rob, Jeff et. al. still have the right to bar someone who they feel is abusing that right from their servers.
Charles Miller --
Re:Why designed for one platform
on
AtheOS
·
· Score: 2
Linux was initially designed solely for Intel. Linus made it perfectly clear that Linux was dependant on certain features of the i386 architecture, and would not work on anything else. Download an older (1.0) kernel and see it for yourself.
Before it was ported the first few times, Linux was hideously non-portable. The first few porting efforts made the core kernel itself more portable, by separating the i386-specific code from the cross-platform code, which in turn encouraged the next set of ports.
So if you want Atheos on [platform of choice], then do what the PowerPC / Sparc / Alpha / MIPS / 68k owners did when they wanted to run Linux:
Slashdot Mirror
Before this debate goes any further, everyone read the
Charles Miller
--
I've been playing computer games regularly for about the last sixteen years, and I can happily say that I can play StarCraft and Diablo II all day. I love playing Unreal Tournament, System Shock II and the Final Fantasy genre.
Pac-man, on the other hand, bores the pants off me. Getting three billion points in it seems to be more an exercise in how to maintain concentration in the face of near-terminal monotony.
Space Invaders likewise. MAME held my interest for all of three days, before the nostalgia value wore off, and I started to wonder just how easily amused I must have been when I was fifteen.
There are _some_ games I'd like to play through again (the original C-64 Wizball springs to mind, although that's probably a little "post-classic"), but a lot of classic gaming seems to just be nostalgia for things that were only good because they were the best that could be done at the time.
Charles Miller
(Woo, gonna lose karma over this one...)
--
"Why couldn't there be a competing domain structure? It could use differently numbered TCP ports to communicate so as not to conflict with the existing one, and be somewhat gnutella-like to avoid centralization."
DNS is already "somewhat gnutella-like". The reason we have the TLD's that we do is because by convention, most DNS servers address the same set of root nameservers, which define what TLD's are available. It is a simple act to set up a root nameserver that will serve additional TLD's, the only hurdle you face is getting other people to add you to their nameserver configurations. People are doing this all the time, but I'm too lazy to find a URL pointing them out.
Some time in the not-so-distant future, I predict that countries will start setting up their own root nameservers, and enforcing their use within the country by legislation. This will be done for two reasons - firstly to stop the net being so reliant on servers within the USA for everyday operations, and secondly because it is the only way I can see to overthrow the current system, where countries are forced to pay a US organization for the privilege of having control of their own country-name TLD.
Charles Miller
--
My via-Palo-Alto experience was when I was living in Perth. AARNET had just switched from Telstra to Optus, but my home box was hooked up to Telstra. The telcos hadn't even bothered to set up peering, they were just throwing everything at the international link. To some extent this still happens. The telcos have POPs in all major cities, but they only bother to peer with each other in Melbourne and Sydney, AFAIK, so the same packets would even today travel an 8000km round-trip in order to cover 15km of real-world distance.
ISPs complicate the issue even more, because routing in commercial ISPs is usually set up based on which link costs the most per megabyte, rather than which actually gives you the fastest route to your destination. If Telstra charges you 19c/meg incoming, but your satellite is feeding you at 7c/meg, you'll end up broadcasting your routes only via the satellite, then using the Telstra link for back-channeling. (Except now Telstra has a back-channeling charge, so unless you have an exemption you have to calculate your routes carefully to stay under the threshhold...)
This tends to disadvantage users, since satellites have a bit of a latency problem, but what they don't know doesn't hurt them.
The short summary is that routing is, and will continue to be totally insane until bandwidth gets cheap, and telcos get a clue. Oh, and then we could use the flying pigs to carry some of the packets!
Charles Miller
--
You wouldn't need to take down the entire US for that. I'm willing to bet that you could bring down the Internet with two or three well-chosen routers (i.e. in a couple of different major backbone NOCs) set up to transmit bogus routes.
Charles Miller
--
Firstly, there's a big difference between what Cringeley suggests (the ability to shut down ISPs) and shutting down all the backbones. Taking down all the ISPs in the USA, but leaving the backbones running would make the net several orders of magnitude faster for the rest of us. (and several orders of magnitude more boring) However, if the backbones were taken down...
.za.
Living in Sydney Australia, pretty much all of my routes go through the USA, except those to very close neighbours such as Malaysia and Indonesia. My routes to Japan and Taiwan go via the USA. South Africa is closer to Perth, Australia than I am. My packets to South Africa go to Perth, THEN to the US, THEN to
Sometimes it's even worse than that. Back when I was at University, it was so bad that when I did a traceroute between two servers 15 minutes drive apart but on different backbones, the packets were going via California.
There are links between countries that could be used if the USA were to vanish, but these links are usually significantly underpowered. Most of the major content providers are in the USA, most of the packets go to and from the USA, so other countries tend to invest most of their money in fat pipes to North America. And since those fat pipes are already there, they may as well take care of some of the local traffic as well.
Between countries on the same continent, you're probably looking at a continuing stable network. But inter-continental links would most likely fall over and die.
Even if the underpowered inter-continental links could take it, you'd see a routing nightmare. BGP packets would be flying around in circles panicking, and any sane network administrator would lock him or herself in a small room and whimper until it was all over.
There's also other things to think of. How many of the root nameservers are outside the USA? How much traffic can they take? How would they cope with the prolonged absence of a.root-servers?
Charles Miller
--
I'm going to lose karma for this one, I know it...
This attitude annoys me no end. I'm all for political correctness, but this is the kind of blinkered, divorced from reality, humourless, reactionary comment that gives its detractors so much ammunition.
In the course of working several jobs, I have worked in male-dominated, female-dominated and evenly mixed workplaces. I've been exposed to men making sexual jokes about women, women making sexual jokes about men, and homosexuals making sexual jokes about each other. In my last job at an ISP, while there was only one female tech, the most common source of comments of the kind you're complaining about was from the receptionist and accounts staff, who were female.
If someone making a joke about the fact they're attracted to whatever gender they're attracted to offends you, then you're going to spend your life perpetually offended, wherever you work. Unless, of course, it's one of those things that's okay for women to do, but not for men.
Yes, it goes too far if people start wallpapering their offices with centerfolds, or if the joke is non-consentually at the expense of, or directed at, a colleague. But while I've seen that happen, I saw much more blatant examples of it outside IT than in, and I'd be willing to bet it's far more prevalent (proportional to women employed) outside the computer industry.
On a similar note, from a previous post by you in a different story:
How about you clear up your own rabid prejudices, before you start harping on the shortcomings of others?
Charles Miller
--
"This just proves that Mozilla is too bloated as well."
When the Mozilla project started, the requirements were that the resulting software be highly portable, compliant with a whole slew of web standards, and be a superset of the functionality of Netscape Communicator. Thus, it was doomed to be bloated before it even began. It was doomed to be bloated because that's what people WANTED it to be.
To quote JWZ (although he was referring to the pre-opensource Mozilla), "Mozilla is big because your needs are big. Your needs are big because the Internet is big. There are lots of small, lean web browsers out there that, incidentally, do almost nothing useful. If that's what you need, you've got options..."
Mozilla's solution to its own bloat-by-design was, IMHO, rather elegant. It took stock of the what it was required to implement, and then used those technologies to build other components. This isn't anything particularly new, of course emacs invents itself from its lisp interpreter.
To fulfil its goal of standards compliance, Mozilla had to understand HTML, XML, Javascript, DOM and CSS. To maintain a suite of applications (browser, mail/news, html authoring tool) across widely differing platforms, it needed its own flexible, cross-platform UI and Networking libraries.
So Mozilla's self-reinvention as a platform, rather than an application was the result of necessity, and some pretty good thinking from the designers rather than just bloat for the sake of bloat. It was actually the most efficient way that it could perform the task it had been given.
Charles Miller
--
When I have moderator points, which admittedly isn't often, I reserve "offtopic" for blatant cases where a post bears no relation to the surrounding posts. Posts that are meta-topic, or posts that are the result of natural thread-drift shouldn't be marked as 'offtopic'.
:P
For example, this post has absolutely nothing to do with penguins, but it's perfectly at home in this particular thread. Having said this, of course, someone's going to come across and mod me down.
Charles Miller
--
So my question is: how appropriate is it for people to use their website as a medium to raise consciousness and solicit donations for whatever cause they are currently supporting?
It is entirely appropriate. Slashdot's mandate is to post stories that interest the editors. It has no responsibility beyond that. "Hey, cool, the penguins need our help" is just as valid a story as "Hey, cool, the DeCSS authors need our help." If there's anyone out there who's so impressionable that they'll think "Hey, Hemos said I should donate, where's my cheque book!" then they deserve our pity.
Slashdot has become popular because the story selections that the editors make is generally of interest to its community of readers. That doesn't mean that every story has to interest every reader, I certainly don't read more than 25% of the stories myself. All it means is that on an average day, there are enough slashdot stories that interest me for me to return the next day. I read the penguin story because it was different.
If you watch a news-hour on TV, there's usually a cute animal story or something of that ilk. They're not there because they're news, they're there because they're _relief_ from news. Similarly, the penguin story is a relief from Yet Another Copyright Article, or Yet Another Obscure Technology.
Follow-up: does it matter if the purpose of the website typically excludes stories about that particular topic, moving though they may be?
You are working from an incorrect assumption. You're assuming that the "purpose" of this website is more concrete than it is. "News for Nerds" doesn't mean "News about technology", although there is a correlation between the two.
And another one: are there some issues which are so important that they should headline any/all forms of media, regardless of the subject matter usually addressed by that media? If so, is this one of those topics?
I doubt the injury of penguins will "headline" news worldwide. It headlines slashdot purely because the penguin-linux connection makes it a fun real-news/nerd-news crossover. Similarly, if there were a bunch of BSD daemons being held hostage in the middle-east, I expect that would show up on the slashdot front page too.
Charles Miller--
My greatest impression from reading the paper was one of a schizophrenic system. The differences between the two systems that form the basis for OSX seemed not to have been resolved, instead they were patched over with an additional layer of complexity, and a great deal of hope applied that the two different OS's at the core wouldn't misbehave and contradict each other.
Usually, this is the sort of thing that makes software developers run screaming down to the pub.
Charles Miler
--
Larry Wall tells an amusing story of how he "hacked the licesnse war" by distributing Perl under the GPL and the Artistic License at the same time. Now if only I could remember the URL...
Charles Miller
--
I don't think it'd be possible to run it under mod_perl. The CGI component of minivend is a C program (vlink?) that just forwards the http request to the Perl daemon running as the back-end. I pretty much came to the conclusion that the whole thing is a kludge. I was running it on a PII w/ 256Meg, and the delay was noticeable with a single client requesting pages. I shudder to think what sort of response serious customers are getting. (I don't work at that company any more). I'd seriously recommend trialling the software before committing yourself to implementing it anywhere. My recommendation is that if you're going the free software route, write your own implementation specifically for the site you're developing, in whatever language you're comfortable with. You won't lose nearly as much development time as you think, because the Minivend markup language is so evil, and your site won't run like a dog. Charles Miller
--
I developed one or two sites in Minivend. I found it to be mind-numbingly slow, and its in-page markup/scripting language was worse than anything I've seen before or since, and that's including Net.Data. Minivend's scripting language is a series of hacks piled upon hacks piled upon hacks. Getting it to do even the simplest of things is an exercise in pain-endurance.
If it's a choice between Minivend and nothing, I'd probably go with nothing. All you really need to replace the vast majority of its 'features' is an in-page scripting language with decent session support.
I'm afraid, in this case, saying that it beats the closed-source solutions is just demonstrating that you've never compared its features or performance with what's being sold at the moment.
Charles Miller
--
(full disclosure: I work for a company that develops apps based on IBM Java stuff.)
<I>Just remember, folks. 15 years ago, IBM was the oppressor of the tech industry. The only reason they aren't as powerful today is because they got too big and unresponsive, which led them to make several bad decisions.</I>
Actually, the major reason they aren't as powerful today is because they lost an anti-trust suit. This is one of the things that people tend to forget in examining the Microsoft trial -- Bill Gates' billions are largely a consequence of IBM losing a very similar case. If it weren't for IBM's consent decree, they'd have been much more able to screw over the PC industry before it even started. Or at least, Microsoft would be a division of IBM, and we'd be running "PC-OS/2000".
<I>Look at what they're doing with Sun and Java. After working with Sun for 4 years, they refuse to ratify Sun's version of EJB, in effect leaving Sun out to dry. IBM was given access to all of Sun's java IP and this is how they repay them?</I>
You're misrepresenting things just a tad.
The relationship between IBM and Sun over Java wasn't one-way. IBM were responsible for a significant proportion of the Java2 Enterprise spec. They have done more to evangelise Java than any other company apart from Sun itself.
Sun don't like IBM because Sun want Sparc/Solaris to be the platform of choice for Java apps, and here comes IBM porting the language to their own line of Risc boxen, and making a JDK for Win32 and Linux that's faster than Sun's. This is competition done right. If you think about it, it's the same thing Compaq did to IBM in the PC hardware industry. Don't embrace and extend, make a superior product that's 100% compatible.
Sun's response was to "take their toys and go home" with regards to Java standardization, and to start charging license fees to people just for the right to say "this is J2EE". Sun left _themselves_ out to dry.
To get back on the subject: The GPL protects Linux from being screwed too radically by IBM. IBM are getting involved in Linux for the same reason they're getting involved in Java -- they've identified it as a technology that people want to use, so it's natural that they want to use it to sell hardware, and it's natural they want to port their proprietary products to run on it.
And if they're doing that (Although they're way behind on Websphere and VA Java.) And at the same time, it makes sense for them to improve the platform so that it will scale up to their larger boxen. They want to be able to say "Hey, your business is going great now, if you upgrade to this RS/6000, you can still run all your Linux stuff on it", but to do that effectively, Linux needs a few of the "enterprise-level" features that people running Wintel servers don't miss much, like Logical Volume Managers.
Which is cool enough. And in the meanwhile, if you have a boss who wants the warm fuzzies that come with buying something with a brand name, you can get a Netfinity, put Linux on it, and say it's an IBM solution. This is the thought behind things like their "Linux Server Proven" scheme -- you prove that your app runs on a Netfinity using their supported Linux configurations[1], they give you a sticker, and endorse you to their customers.
It takes us a step closer to the day when the Linux success stories _aren't_ all of the form "I switched to Linux over the weekend and my boss didn't notice!"
Charles Miller
[1] They support (off the top of my head) RedHat, SuSe, and I think Corel and TurboLinux. The rationale behind this is that three of them cover the "most popular distro" of North America, Europe and Japan (Asia?) respectively, and the fourth is probably there because it's "most suit-friendly".
--
They're getting there.
SSL is supported on Windows and Linux/Intel. Java is only supported on Windows currently, due to the fact that until recently, Windows was the only platform to have a JDK with an OJI plugin.
Shockwave is an abomination that should be wiped off the face of the web.
Have a nice day.
Charles Miller
--
YM "five years" HTH. After the appeals court, it'll go to the Supreme court.
Five years is a hell of a long time in the software business. I mean, five years ago Linux was celebrating version 1.0, and the General Public was deciding whether they should upgrade from Win3.1 to Win95, and wondering if they should get on "That AOL thing." In five years, we saw Microsoft use its OS leverage to kill Netscape, and crap on who knows how many OEMs?
This means that Microsoft still have a lot of time to do a lot of damage. And if they think they've got a good chance of losing the case, it gives them five years to use their Monopoly to make sure that if they're broken up, both resulting companies end up as powerful as they can make them.
I mean, can you see any other rationale behind their new policy of no install media for OEM Win9x licenses? It's like they're saying "We've only got a few years of the gravy train left. Lets stretch it out as long as we can, and screw as many people as we can along the way."
Scary, huh?
Charles Miller
--
Pretty much everyone. From what I remember of the game, there were so few changes that they should have sold it as an add-on pack for Doom 1.
Charles Miller
--
The article linked to above is a pretty shaky foundation on which to base a criticism of JSP. It's obvious that the author is enamoured with his pet language, to the extent that he thinks saving a few keypresses equates to increased efficiency in the Real World.
Let's look at the criticisms one by one:
"Java code too tempting."
This is true of any scripting language. I've been forced to clean up amateur ASP projects that were 90% inline VBScript, 10% HTML, 0 VB objects. PHP3 was even worse, because you were almost forced to inline everything (which put me off PHP to the extent that I never bothered to look at whether v4 fixed that problem).
The proposed solution, however, removing the _ability_ to inline code, is far too restrictive. Why should my proof-of-concept JSP need to be coded for maintainability? If I've got a deadline tomorrow, I might _want_ to sacrifice good coding style, secure in the knowlege I can split the load properly between servlet and JSP over the next week, so long as the app is running. In the real world, there are too many situations where you want to inline code, although you should plan on fixing it later if the page is anything more than a one-shot.
"Java Code Required"
There is no real advantage to being able to type "$Request.ContextPath" instead of "Request.getContextPath()", unless you count three keypresses as critical to your project. They are two different syntaxes for the same thing. This is like saying that WebMacro is a bad thing because it forces you to use WebMacro.
The remaining complaints in that section are solved by using custom tags.
"Simple Tasks are Hard"
Once again, "#parse "header.wm"" is just as complicated as "<%@include file="header.jsp"%>". The only difference is that the latter is a lot less likely to upset GUI web design tools, which makes it easier to keep the web-designers happy.
"Lousy Looping"
The JSP/taglib example of "foreach" is identical in complexity to the WebMacro version, once more with the advantage of being more friendly to existing web design applications.
"Useless Error Messages"
That's an implementation issue, not a problem with JSP's themselves. A decent JSP compiler should spot constructs that will cause errors in a compiled class, and flag them. If your JSP compiler is not decent, it's the fault of the people who wrote the compiler, not of the JSP spec.
As an aside, the setup I have here allows me to step through the JSP compilation process, and attach a debugger to running servlets and JSP's. Being able to set breakpoints, walk through methods etc. etc. makes it a lot easier to find the more obscure of my mistakes.
"Need a compiler"
So how do you run a dynamic site without a compiler (or an interpreter, if you want to be picky)? Osmosis? Telepathy?
It's not particularly hard to get hold of a Java compiler. They're available for free (beer) on most platforms (Solaris, NT, Linux, AIX, HP-UX, MacOS, we've even got one client running an application server off an AS/400). I'll admit, if you're not running on a platform with an available JDK, then JSP isn't for you.
"Wasted space"
The disk-space argument isn't even worth mentioning. You're never going to have so many JSP's that it'll cost you more than a few dollars worth of hard drive space, if that.
Sure, Java uses up a lot of memory. But that's Java itself, not something specific to JSP. I don't know what system he was using to test on, but the average size of a JSP on the production web-app closest to hand is 6-7K, not 30+K. The JSP compiler doesn't load all the HTML as static class data, it gets pulled off the disk at the start of the service() method, and is available for garbage-collection as soon as the request is done.
Anyway, enough counter-rant.
Charles Miller
--
IBM flirt with Free Software occasionally, but it's not the way they generally do business. They give a lot of stuff away, in a free beer sense, but only because it can get you in a position to buy some of the stuff that makes them money - things like DB2, etc. But in a Free Speech sense, the only things that spring to mind are their collaborations with the Apache group in XSLT stuff, and the Jikes compiler.
More power to them, I guess. I'm not RMS. Whether a company frees its code is entirely up to the company. Their half-hearted Linux support disappoints me, on the other hand. Websphere's still a major version behind, and you can't get VAJava Enterprise Edition at all.
Alphaworks is a special case. The licenses are stricter than production products because they're _not_ production products. It's alpha software. You get to play with it, but on the understanding that next month it may suddenly have become a commercial product, or it may no longer be being developed at all. So the site is neat if you want to scrounge new toys from the bleeding edge, but don't use it for anything serious.
[offtopic] My favourite IBM free-beer package at the moment is their developer program where you can book in to spend time testing your apps on Big IBM Boxen. It looks pretty cool for those of us who can't afford an AS/400, or who don't have the resources to do huge load tests.
--
And this isn't flamebait. United States Haters go elsewhere. No, it's a troll. The "ETHERNET" reference was a dead giveaway. Charles Miller
--
This was an injunction. You take out injunctions after you've told the guy to stop some other way. Ebay got no damages at all, all they did was assert their right to say "Don't do that again."
Noone's going to be able to sue you because you visited a site, and then learned after the fact that you weren't allowed to. All they can do is tell you not to visit it again.
Charles Miller
--
The "using CPU cycles" is not a new legal argument. Back when there were no specific laws against computer cracking, I believe one of the earliest convictions came from proving that a cracker "stole electricity" by using cycles on the machine they'd broken into.
You'll first note, from following the provided link, that this was simply an injunction. There was no mention of damages. There is nothing in this that sets any precedent against webcrawlers, or deep linking. All the case says is that yes, the owners of a webserver may choose how that server is used.
Bidders' Edge were doing something to Ebay's site that Ebay didn't want them to do. Ebay asked them to stop. Bidders' Edge didn't want to stop. So it went to court. The court told Bidders' Edge that what they were doing constituted a use of the Ebay system (their CPU cycles) against the wishes of its owner, and ordered them to stop.
Now what, exactly, is wrong with that? Do we really want to set the alternative precedent, which is that as soon as you put a public resource on the net, you have no say in how that resource is used? I can imagine all the IRC script kiddies drooling at the thought of that one. "Hey! It's a public server, my clonebots were only a few hundred out of the thousands of users online, so I can't have been using a significant amount of the net..."
Charles Miller
--
Just because a resource has been made available to the general public, DOES NOT mean that the owners of that resource shouldn't retain every right to be able to bar people from using it, for whatever reason.
The ability to say "You are no longer permitted to use my server in that way, stop it." is an essential one. It means, for example, that even though slashdot is open for public posting, Rob, Jeff et. al. still have the right to bar someone who they feel is abusing that right from their servers.
Charles Miller
--
Linux was initially designed solely for Intel. Linus made it perfectly clear that Linux was dependant on certain features of the i386 architecture, and would not work on anything else. Download an older (1.0) kernel and see it for yourself.
Before it was ported the first few times, Linux was hideously non-portable. The first few porting efforts made the core kernel itself more portable, by separating the i386-specific code from the cross-platform code, which in turn encouraged the next set of ports.
So if you want Atheos on [platform of choice], then do what the PowerPC / Sparc / Alpha / MIPS / 68k owners did when they wanted to run Linux:
Port it.
Charles Miller
--