Why does the laptop have to be always on? All that needs to be always on is the lojack system--which I assume would consume less power than the computer itself.
Actually, I'm not entirely sure about that. How much power does it take to broadcast your position, or whatever lojack does?
The big question I have is, would anyone really pay for this? When you pay $30k for a car, and expect to sell it for $15k in a few years, an extra $800 plus $100/year is no big deal. When you pay $3k for a laptop and expect to sell it for $300 in a few years, it doesn't make as much sense....
I think you're missing the point. Obviously C++ has as many problems as C, because you can compile almost any C code as C++ code. But his solution is to use std::string instead of char*. So try this:
std::string my_str; cin >> my_str;
It's not that C++ is a magical panacea, it's that using STL and the other standard C++ libraries, and sticking to their idioms, gives you protection against a lot of these problems.
Why? Well, the main idea in this case is that STL keeps track of the end of every string, not just the beginning. You can use the same idiom in C (instead of strcpy you use strncpy and pass in send-s), but the standard C libraries don't encourage you to think that way, and the C++ libraries do (and even automate a lot of the work).
Of course it doesn't prevent roll-your-own overflows. But why is this an issue? Why would someone code "while (*s1) *s2++=*s1++" instead of using strcpy?
Well, it could be a trojan, meant specifically to get around libsafe. But if you can get someone to run your trojan, you don't need to do this.
Or maybe there's some idiot who thinks that this loop will perform better than strcpy on every platform. Well, I've actually worked with people who think this way, but you can spot their code a mile away, and there are plenty of other reasons you wouldn't want to trust it....
As for you other issue (allowing overwrites of other parameters), that's true. So it doesn't fix all possible overwrite problems, just some of them. Does that mean you shouldn't use it?
OK, I'll come out in favor of Cyber Patrol. And this is not a troll.
I think American liberalism was weakened by its somewhat-successful attempts to coopt the moderates in the 70s (post-Watergate, Carter) and even more the 90s (DLC, Clinton).
Those of us who grew up as "liberal Democrats" in the 70s had Carter to look up to. Compare that to my grandmother, who had radical union leaders and anti-fascists. Is it any wonder that their first success was FDR, while ours was Clinton?
Maybe making Democratic party information "taboo" will lead to a revitalization of real progressive sentiment in the next generation of "liberal Democrats." It won't take a genius to realize how ridiculous the filtering software is, and by extension the system that created it. They'll either find a way around it or to get the information you're looking for elsewhere. If anything, it'll make liberal politics--and, by extension, politics in general--more interesting to kids.
If that doesn't happen, if the Democratic Party remains the party of welfare cuts, free trade, the war on drugs, and forcible repatriation of refugee immigrants, what do we need them for? We already have the Republicans for that.
Don't fret; I'm guessing things will go the same way they've gone in other fields: a fire-sale period, a drought, and then a new equilibrium that's not far from the original.
Look at turntables. After a couple of years of vinyl being declared "dead," you could find turntables for a third of their old price, record pressing equipment for a fifth--and records that were marked at $10 were in the 50 cent bin. DJ's had a field day.
Then came the backlash. Vinyl was cool. Suddenly, the only turntables being manufactured were high-end DJ decks, ultra-high-end audiophile equipment, and ultra-low-end throwaway trash, and people were selling used Technics 1200's for more than list price. The end result? A dozen new companies got into the business. The top-end models are still slightly more expensive than they were a decade ago, but there are decent turntables available for $100-$150, far better than anything in that price range in the old days. And where there used to be only one decent cartridge for under $50, there are now dozens of options. Sometimes the free market actually works....
Of course in the case of turntables, consumables are much less of an issue--a cartridge lasts a lot longer than a roll of film. So when the crash comes, stock up on film. But don't worry about the cameras and lenses.
Re:Shouldn't we consider this a "good" thing?
on
Athlons Sold Out
·
· Score: 1
And don't forget that Tyan and others who make motherboards for Athlons lose money, and those who make prefab computers out of Athlons lose money. (Or make the same money by selling more Intel boxes). This hurts the Athlon market as a whole.
It's just like Motorola's G4 shortage: It hurt Motorola and Apple at multiple levels. But in this case, Dell doesn't have to do what Apple did and cover for AMD; they can just push Intels harder.
"Second, traditional cameras, are likely to be just as fragile as a digital camera, except that they have more moving parts (shutter screens, rollers, etc.) that can seize up over time. I certainly wouldn't take this into consideration when buying a camera though -- unless I knew it was going to be handled rough."
Digital cameras certain _can_ be made less fragile than traditional cameras, and some models are. But the vast majority of cheap and midrange digitals out there are easier to break than cheap and midrange traditional cameras, and cost more to replace.
One point you didn't bring up in support of your point, by the way: "Most shops will allow you to shoot a roll (or, I guess, a flash card?) if you give them collateral, or know them well. If not, ask to shoot a few in the store." In the case of digital cameras, this doesn't cost them, or you, a dime. You can try out every digital camera in the store without having to buy 100 rolls of film. This gives the truly discerning consumer a lot more power.
Anyway, my conclusion is the same: One day, digital cameras will push traditional cameras into a niche, but not yet.
A cheap ($300) digital camera is (almost) as good as a cheap ($50) analog camera. A professional ($25000) digital camera is (almost) as good as a professional ($1500) analog camera. I think that proves my point. The price-performance just isn't there yet. And you can get a set of prints made from your memory stick or floppy just as you can from your 35, you just might have to drive a little farther and pay a little more.
Of course the consumables are higher for analog cameras--especially if your end goal is images on the computer.
This especially true in the middle range, where someone's going to be spending either $800 for an analog camera or $1000 for a digital camera, and the savings in film will make a huge difference.
But the kind of person who buys a FunShot disposable is not going to switch to digitals until the $5 digital disposable that the next poster mentioned is available. And the kind of person who takes pictures that are meant to be used as magazine covers isn't going to switch to digital until they're as good, as flexible, and much cheaper than $25,000.
Eventually, digital cameras will provide better price-performance than anything in the analog realm, and the majority of people will switch. Of course even then, traditional cameras will have their place. However, its place will be a niche.
Digital video may replace 1" video (much less VHS), but people will always keep using film; newer digital audio formats may eliminate CDs and MDs and DATs, but records will always be there. And in the future, traditional cameras will occupy a similar niche to film and vinyl.
Well, I think this is one of the interpretations he wanted you to take away, but I think he wanted it to be plausible that the terrorists were real, or that the terrorists were a fantasy of Sam's, or that they terrorists were a phantasmal "enemy" created by the system to explain the failures of the system, or that the terrorists were real but really weren't that effective and that the system built them up because it's easier to catch Mafiaboy than to address the real issues.
It's been a couple of years since I watched it carefully, but I think that all of these interpretations works.
Brazil isn't the first movie to do this, but it has become much more common since then. For example, two big recent movies--Fight Club and American Psycho--both lend themselves to a range of interpretations ranging from "It's all a fantasy in the head of a man who feels powerless and impotent in modern society" to "It's all true."
Well, not exactly. Obviously, some of what happens in each of those movies is fantasy. At the very least, the paperwork burial in Brazil, the explosions in more buildings than were targeted in Fight Club, and the over-the-top slaughter scene in American Psycho are at the very least distorted visions of what's really going on. But that doesn't mean that nothing that happens is real. And each person who walks out of the theater has a different perception about how much is real.
Maybe it's just a more sophisticated version of the classic Wizard of Oz ending, but it certainly gives you more to think about, and more to talk about--which is the point.
"One of the genius visions of this film is its depiction of the crazy patchwork of technology that runs our civilization."
This is an interesting way to put it. Other reviewers always refer to the crazy patchwork of technology that runs Brazil's civilization as a metaphor for the crazy patchwork that runs ours.
But actually, to a large extent, it is literally true. I don't know why I never thought about this before....
Well, there are still reasons for traditional cameras.
First, they're cheaper, and less fragile. Do you want to bring your $600 toy into the pit at an Atari Teenage Riot show, or would you rather carry a disposable camera?
Second, if your ultimate goal is to have prints to keep around, it's cheaper to develop a roll of film than the print out a digital image on a photo printer (with photo paper and ink). Plus, while it takes about the same amount of time, sometimes it's more convenient to just drop off the film, get lunch, and come back 23 minutes later than to spend that 23 minutes over your computer.
Third, there's quality. I'm not going to go into the old argument of the theoretical quality of analog vs. digital (records vs. CDs, for example), because they're mostly biased BS. But anyone can look at a picture taken with a consumer digital camera and a picture taken with an equivalent-priced analog camera and see the difference. And when you factor in interpolated digital zoom vs. optical zoom (since most digital cameras only do a small amount of optical zoom, whereas for the same price you could buy a good traditional camera and any zoom lens you want), it's even more dramatic.
I'm not saying that digital cameras don't have their place. But for the time being, traditional cameras have their place, too.
If you want to hear more from the "first group," the lists at http://www.open-rsc.org are still open, and people there continue to boldly beat the dead horse.
The best example of the "third group" can be found at http://www.kenfreed.org. Ken also has a huge collection of useful links (including all of the comments collected by Esther Dyson except mine, for some reason).
The name.space people can be found at http://name.space, but in case you happen to be among the 99.999% of the net that can't resolve that name, they're also http://name.space.xs2.net.
And somewhere, I believe there is an archive of all of the crazy ideas of Jeff Williams.
If anyone wants to have this arguments all over again, please go have them with the newdns archives.
Anyway, your history is a little off. I'll try to summarize the summary that I wrote for Esther Dyson a few years back....
Based on some hints from Postel that IANA might consider adding alternate TLDs, a bunch of opportunists decided that they had the mandate to create their own TLDs (each new TLD owned by someone, as.com,.net, and.org were owned by NSI) and make IANA add them to the root servers.
However, IANA's intention was to spend a lot of time studying the best way to handle things, and the group they set up to study it (the group that eventually because the gtld-mou) went off in an entirely different direction. Their plan was for seven new TLDs to be managed by their organization, and anyone who signed up and paid enough money could become a registrar selling names under those TLDs. Which wasn't a bad idea, except for their insane structure and contractual setup--and the idea of putting DNS under ITU.
Meanwhile, the people who'd spent a bunch of money trying to be the first ones to claim exciting names like.biz,.xxx, etc. set up their own organizations, servers, and rules and decided that they'd take over the net gradually. If they could get just 5% of the net to use their root servers, the rest of the net would demand access to the new TLDs. And it might have worked if they'd gotten more than 0.05%....
Unfortunately, none of these people could get along. eDNS, uDNS, AlterNet... they ended up with a huge mess of infighting and lawsuits. Not to mention the little "joke" AlterNet pulled with the root servers....
While this was going on, the gtld-mou people decided they'd go through the governments of the world--and, when that didn't work, through the ITU.
A third group of people decided that the only way to decide this was through international Internet democracy, and set about coming up with the right way to build such a governance system.
Meanwhile, a company calling itself name.space decided they'd start their own root servers and allow anyone to register any two-level domain name with any TLD, whether it already existed or not. So Rob Malda could have rob.malda, or iam.rob.
Finally, a fifth group (mostly Europeans) wanted to force everyone who wasn't doing something truly international to use the country-code domains (.us,.uk,.mx), and anyone who was to use.int, completely eliminating.com,.net, and.org, not to mention.edu,.gov, and.mil (of course eliminating the inverse-lookup arpa domain breaks bind...).
Then the big question arose: Who actually has the authority to change the system? In the end, only the US government has any authority (they "own" IANA and Internic--and they can make NSI play ball). So some people from the first and third groups, hoping to stop the gtld-mou, went to the US government and began pressuring them to do something.
Somehow, Esther Dyson, in typical Esther Dyson fashion, managed to present herself as the world's foremost expert and the only person who could be trusted to muddle through this.
And in the end, we got ICANN. Backed by the US government, they took the basic gtld-mou idea of a core of registrars and applied it to.com,.net, and.org, and promised to start all over with studying how to expand the TLD space.
So now we get the same arguments all over again. Fun, eh?
Well, I'm sure iCab 1.9 shares a whole lot of code with 1.8, but the whole browser is only about as old as the Mozilla project.
In another post, I mentioned that the main reason Mozilla looks so unfinished is that it was an incredibly ambitious project--far more ambitious than IE 5.0 to 5.5, obviously.
I know that they have lots of justification for the current state of the project. In fact, I'm impressed that they've made it as good as they have.
But the "unwashed masses," as you put it, don't know that. We can all work to educate them, but I still think it's going to hurt public confidence in Netscape.
The Windows version I don't know as much about, but on the Mac, from version 3 to 4, everything feels very different. Everything that used to crash it no longer does (but new things do), there's a completely different set of HTML and layout bugs (with little overlap), the interface with the JVM is totally new it's smaller, it loads its code in a different way (all those hideous System libraries are gone...), etc.
I don't believe what Microsoft says just because they say it, but I think in the case of IE 4.0, they really did rewrite from scratch, just as they claim.
Comparing Office 6.0 to Office 98, I think the same is true there. That's two cases I know of where Microsoft did a complete rewrite (three if you count Outlook Express as a separate product), and throwing away their old code turned out to be a great idea. If only they'd apply that to the next version of Windows....
I'm not sure you want to call something a "real innovation" if every other OS already has it....
Anyway, it's interesting that you should use the Amiga as an example, because back when I was last doing serious BeOS development (this is in the days when BeOS for Intel was just a "sometime in the future" idea, so things may have changed...), a lot of developers were using the datatypes library ported from the Amiga.
It's actually a pretty cool library. If I build my app to use datatypes to draw images, and the user adds a PNG datatype library to her system, my app can magically use PNGs (except for some minor issues with selecting files by extension or MIME type...).
And I think that's the level this should be at--it shouldn't be built into the OS (or the windowing system). I should be able to download the PNG datatype library and install it, and maybe even choose among a few different PNG datatypes.
Anyway, it would be nice if linux had something like that, right? But doesn't it already? Isn't that the point of ImageMagick, and a few other libraries?
By the way, Windows doesn't natively have anything like this. If you use the Win32 bitmap APIs, you only get BMPs, and that's it.
But QuickTime (on both Windows and MacOS) will draw all kinds of images for you painlessly, and every time it's updated, your app magically gets new image types. And under Windows, there are ActiveX controls (including the IE control, but also some lighter-weight choices) that will give you the same thing.
Did you go to the more complete test suite linked on the page? Did gamma really work right? Did alpha transparency without a background (in other words, compositing with the page itself) work? Because I tested IE 5.0 under Win98, and 4.0, 4.5, and 5.0 under MacOS 9, and none of them did. Plus, the image that Netscape tried to load forever actually crashed MacOS 9 with IE 4.0.
I suppose it's possible that later versions of IE have actually gone backwards in PNG compatibility, but I doubt it.
Still, all versions of IE did much better than Netscape 4.7. On the other hand, Mozilla did better than either. As did Opera and iCab. So much for IE, eh?
But not a single browser I have available passed everything (well, lynx did fine passing the PNGs off to an outside viewer...). Which means the problem is serious.
This is an interesting idea. It would probably be easier to modify existing software to use "GIF2000" than to add PNG support, if the format change were kept incredibly simple (ideally, just link with a new library...).
A few minor changes to the spec allows a GIF00a format which is as backwards-compatible as possible (a GIF87a viewer could intelligently skip the images, and a lot of code can be reused). I've got the details below.
Now go ahead and code a reference implementation. If you decode LZW, zlib, and uncompressed, but only encode zlib and uncompressed, you should be patent-free.
A quick summary:
Use "GIF00a" instead of "GIF87a" or "GIF89a" in the header. Take over the 2-bit "Reserved" slot in the "Packed Fields" in the Index Descriptor for a compression method specifier: 0=LZW, 1=zlib, 2=reserved, 3=uncompressed.
In the Table Based Image Data, always use a code size of 0 for zlib or uncompressed data. As with LZW data, break the zlib or uncompressed stream into blocks of up to 255 bytes; prepend each with a block size byte; and append a 0 byte to the end of the last block. The blocks, when you assemble them and strip off the size bytes, make up a zlib data stream, or an uncompressed image.
For zlib data, use the exact same rules as PNG: no pre-defined codes, type-8 (deflate) compression only, and a maximum sliding window size of 32768. As with LZW images, a zlib image which reaches the terminator without specifying a complete zlib datastream is corrupt.
Sections to change: (change)17.c.ii, (change)20.c, (insert)20.c.ix, (change)20.e, (change)22.a, (change)22.c, (change)22.c.i, (insert)Appendix G, (insert)Appendix H
The TIFF format may be open, but it uses LZW, just like GIF, which means it has the same problems. (Also, there are all kinds of problems with the TIFF spec, but that would be getting way off-topic.)
As for "implementing essentially a JPEG driver," I'm not sure what that means. Any application can use libjpeg or libpng to take a JPEG or PNG and use it. It's really not that hard.
You might want to have wrapper classes that fit in nicely with your foundation (MFC, BeKit, Gtk--, PowerPlant, etc.), but they're pretty easy to write and can be easily supplied by a third party. For example, on one Windows/MFC project, I replaced all the BMPs with PNGs just by swapping in a (free) third-party library with a class called CImage and using it in place of MFC's CBitmap.
Anyway, any application that I work on that needs to display images, I push for PNG as the format to use, and having news stories like this to show to the suits often helps. But on web pages I design, I often find myself having to use GIFs because I want the page to work with most browsers. Sad, but what other choice do I have?
I don't know what you've been gathering, but you might want to read the article. UniSys can charge people to use (LZW-compressed) GIFs on a web page, and is doing so. That's what all the fuss is about.
This is a good question. Let's look at it. Why does anyone pay for Windows (other than the fact that it came "free" with their computer--meaning they already paid for it whether they use it or not)?
Well, there are a large number of things Windows does better than linux. Even if linux is clearly better overall, Windows is much better for playing games, watching ASF movies, reading the stupid Word-formatted memos that everyone emails around on the Exchange mail server at work, browsing the web, working on images, composing music, etc. If vmware didn't exist, I'd actually have to boot to Windows reasonably often (ditto for mol on my PPC machines).
So, why does anyone use GIF instead of PNG?
Again, there are things it does better. For one thing, it shows up in every browser. Transparency works in almost every browser. It does animations. There are better tools for manipulating GIFs.
In other words, while the PNG format is clearly far superior to the GIF format, there are reasons people keep using GIFs.
Slashdot Mirror
Why does the laptop have to be always on? All that needs to be always on is the lojack system--which I assume would consume less power than the computer itself.
Actually, I'm not entirely sure about that. How much power does it take to broadcast your position, or whatever lojack does?
The big question I have is, would anyone really pay for this? When you pay $30k for a car, and expect to sell it for $15k in a few years, an extra $800 plus $100/year is no big deal. When you pay $3k for a laptop and expect to sell it for $300 in a few years, it doesn't make as much sense....
I think you're missing the point. Obviously C++ has as many problems as C, because you can compile almost any C code as C++ code. But his solution is to use std::string instead of char*. So try this:
std::string my_str; cin >> my_str;
It's not that C++ is a magical panacea, it's that using STL and the other standard C++ libraries, and sticking to their idioms, gives you protection against a lot of these problems.
Why? Well, the main idea in this case is that STL keeps track of the end of every string, not just the beginning. You can use the same idiom in C (instead of strcpy you use strncpy and pass in send-s), but the standard C libraries don't encourage you to think that way, and the C++ libraries do (and even automate a lot of the work).
Of course it doesn't prevent roll-your-own overflows. But why is this an issue? Why would someone code "while (*s1) *s2++=*s1++" instead of using strcpy?
Well, it could be a trojan, meant specifically to get around libsafe. But if you can get someone to run your trojan, you don't need to do this.
Or maybe there's some idiot who thinks that this loop will perform better than strcpy on every platform. Well, I've actually worked with people who think this way, but you can spot their code a mile away, and there are plenty of other reasons you wouldn't want to trust it....
As for you other issue (allowing overwrites of other parameters), that's true. So it doesn't fix all possible overwrite problems, just some of them. Does that mean you shouldn't use it?
American political parties blocked: Democratic, Green, Reform.
American political parties not blocked: Republican, Libertarian, American Socialist.
So apparently moderate leftists aren't acceptable, but socialists are.... Now that's interesting.
OK, I'll come out in favor of Cyber Patrol. And this is not a troll.
I think American liberalism was weakened by its somewhat-successful attempts to coopt the moderates in the 70s (post-Watergate, Carter) and even more the 90s (DLC, Clinton).
Those of us who grew up as "liberal Democrats" in the 70s had Carter to look up to. Compare that to my grandmother, who had radical union leaders and anti-fascists. Is it any wonder that their first success was FDR, while ours was Clinton?
Maybe making Democratic party information "taboo" will lead to a revitalization of real progressive sentiment in the next generation of "liberal Democrats." It won't take a genius to realize how ridiculous the filtering software is, and by extension the system that created it. They'll either find a way around it or to get the information you're looking for elsewhere. If anything, it'll make liberal politics--and, by extension, politics in general--more interesting to kids.
If that doesn't happen, if the Democratic Party remains the party of welfare cuts, free trade, the war on drugs, and forcible repatriation of refugee immigrants, what do we need them for? We already have the Republicans for that.
Don't fret; I'm guessing things will go the same way they've gone in other fields: a fire-sale period, a drought, and then a new equilibrium that's not far from the original.
Look at turntables. After a couple of years of vinyl being declared "dead," you could find turntables for a third of their old price, record pressing equipment for a fifth--and records that were marked at $10 were in the 50 cent bin. DJ's had a field day.
Then came the backlash. Vinyl was cool. Suddenly, the only turntables being manufactured were high-end DJ decks, ultra-high-end audiophile equipment, and ultra-low-end throwaway trash, and people were selling used Technics 1200's for more than list price. The end result? A dozen new companies got into the business. The top-end models are still slightly more expensive than they were a decade ago, but there are decent turntables available for $100-$150, far better than anything in that price range in the old days. And where there used to be only one decent cartridge for under $50, there are now dozens of options. Sometimes the free market actually works....
Of course in the case of turntables, consumables are much less of an issue--a cartridge lasts a lot longer than a roll of film. So when the crash comes, stock up on film. But don't worry about the cameras and lenses.
And don't forget that Tyan and others who make motherboards for Athlons lose money, and those who make prefab computers out of Athlons lose money. (Or make the same money by selling more Intel boxes). This hurts the Athlon market as a whole.
It's just like Motorola's G4 shortage: It hurt Motorola and Apple at multiple levels. But in this case, Dell doesn't have to do what Apple did and cover for AMD; they can just push Intels harder.
"Second, traditional cameras, are likely to be just as fragile as a digital camera, except that they have more moving parts (shutter screens, rollers, etc.) that can seize up over time. I certainly wouldn't take this into consideration when buying a camera though -- unless I knew it was going to be handled rough."
Digital cameras certain _can_ be made less fragile than traditional cameras, and some models are. But the vast majority of cheap and midrange digitals out there are easier to break than cheap and midrange traditional cameras, and cost more to replace.
One point you didn't bring up in support of your point, by the way: "Most shops will allow you to shoot a roll (or, I guess, a flash card?) if you give them collateral, or know them well. If not, ask to shoot a few in the store." In the case of digital cameras, this doesn't cost them, or you, a dime. You can try out every digital camera in the store without having to buy 100 rolls of film. This gives the truly discerning consumer a lot more power.
Anyway, my conclusion is the same: One day, digital cameras will push traditional cameras into a niche, but not yet.
OK, summarizing the replies:
A cheap ($300) digital camera is (almost) as good as a cheap ($50) analog camera. A professional ($25000) digital camera is (almost) as good as a professional ($1500) analog camera. I think that proves my point. The price-performance just isn't there yet. And you can get a set of prints made from your memory stick or floppy just as you can from your 35, you just might have to drive a little farther and pay a little more.
Of course the consumables are higher for analog cameras--especially if your end goal is images on the computer.
This especially true in the middle range, where someone's going to be spending either $800 for an analog camera or $1000 for a digital camera, and the savings in film will make a huge difference.
But the kind of person who buys a FunShot disposable is not going to switch to digitals until the $5 digital disposable that the next poster mentioned is available. And the kind of person who takes pictures that are meant to be used as magazine covers isn't going to switch to digital until they're as good, as flexible, and much cheaper than $25,000.
Eventually, digital cameras will provide better price-performance than anything in the analog realm, and the majority of people will switch. Of course even then, traditional cameras will have their place. However, its place will be a niche.
Digital video may replace 1" video (much less VHS), but people will always keep using film; newer digital audio formats may eliminate CDs and MDs and DATs, but records will always be there. And in the future, traditional cameras will occupy a similar niche to film and vinyl.
Spoilers for Brazil and for other movies below:
Well, I think this is one of the interpretations he wanted you to take away, but I think he wanted it to be plausible that the terrorists were real, or that the terrorists were a fantasy of Sam's, or that they terrorists were a phantasmal "enemy" created by the system to explain the failures of the system, or that the terrorists were real but really weren't that effective and that the system built them up because it's easier to catch Mafiaboy than to address the real issues.
It's been a couple of years since I watched it carefully, but I think that all of these interpretations works.
Brazil isn't the first movie to do this, but it has become much more common since then. For example, two big recent movies--Fight Club and American Psycho--both lend themselves to a range of interpretations ranging from "It's all a fantasy in the head of a man who feels powerless and impotent in modern society" to "It's all true."
Well, not exactly. Obviously, some of what happens in each of those movies is fantasy. At the very least, the paperwork burial in Brazil, the explosions in more buildings than were targeted in Fight Club, and the over-the-top slaughter scene in American Psycho are at the very least distorted visions of what's really going on. But that doesn't mean that nothing that happens is real. And each person who walks out of the theater has a different perception about how much is real.
Maybe it's just a more sophisticated version of the classic Wizard of Oz ending, but it certainly gives you more to think about, and more to talk about--which is the point.
"One of the genius visions of this film is its depiction of the crazy patchwork of technology that runs our civilization."
This is an interesting way to put it. Other reviewers always refer to the crazy patchwork of technology that runs Brazil's civilization as a metaphor for the crazy patchwork that runs ours.
But actually, to a large extent, it is literally true. I don't know why I never thought about this before....
Well, there are still reasons for traditional cameras.
First, they're cheaper, and less fragile. Do you want to bring your $600 toy into the pit at an Atari Teenage Riot show, or would you rather carry a disposable camera?
Second, if your ultimate goal is to have prints to keep around, it's cheaper to develop a roll of film than the print out a digital image on a photo printer (with photo paper and ink). Plus, while it takes about the same amount of time, sometimes it's more convenient to just drop off the film, get lunch, and come back 23 minutes later than to spend that 23 minutes over your computer.
Third, there's quality. I'm not going to go into the old argument of the theoretical quality of analog vs. digital (records vs. CDs, for example), because they're mostly biased BS. But anyone can look at a picture taken with a consumer digital camera and a picture taken with an equivalent-priced analog camera and see the difference. And when you factor in interpolated digital zoom vs. optical zoom (since most digital cameras only do a small amount of optical zoom, whereas for the same price you could buy a good traditional camera and any zoom lens you want), it's even more dramatic.
I'm not saying that digital cameras don't have their place. But for the time being, traditional cameras have their place, too.
First, http://www.newdom.com/archive has almost-complete archives of the original IIIA list (also at http://www.iiia.org/lists/newdom) and the most important successors (ar.com and vrx.net/newdom.com).
You may also want to read http://www.newdom.com/archive/iana- meeting.html, about the incident that started the whole thing.
If you want to hear more from the "first group," the lists at http://www.open-rsc.org are still open, and people there continue to boldly beat the dead horse.
The gtld-mou ("second group") site is http://www.gtld-mou.org. You can find the archive to their discussion list at http://www.gtld-mou.org/gtld-disc uss/mail-archive.
The best example of the "third group" can be found at http://www.kenfreed.org. Ken also has a huge collection of useful links (including all of the comments collected by Esther Dyson except mine, for some reason).
The comments to the NTIA are scattered around their website, but start at http://www.ntia.doc.gov/ ntiahome/domainname/domainhome.htm.
The name.space people can be found at http://name.space, but in case you happen to be among the 99.999% of the net that can't resolve that name, they're also http://name.space.xs2.net.
And somewhere, I believe there is an archive of all of the crazy ideas of Jeff Williams.
If anyone wants to have this arguments all over again, please go have them with the newdns archives.
.com, .net, and .org were owned by NSI) and make IANA add them to the root servers.
.biz, .xxx, etc. set up their own organizations, servers, and rules and decided that they'd take over the net gradually. If they could get just 5% of the net to use their root servers, the rest of the net would demand access to the new TLDs. And it might have worked if they'd gotten more than 0.05%....
.uk, .mx), and anyone who was to use .int, completely eliminating .com, .net, and .org, not to mention .edu, .gov, and .mil (of course eliminating the inverse-lookup arpa domain breaks bind...).
.com, .net, and .org, and promised to start all over with studying how to expand the TLD space.
Anyway, your history is a little off. I'll try to summarize the summary that I wrote for Esther Dyson a few years back....
Based on some hints from Postel that IANA might consider adding alternate TLDs, a bunch of opportunists decided that they had the mandate to create their own TLDs (each new TLD owned by someone, as
However, IANA's intention was to spend a lot of time studying the best way to handle things, and the group they set up to study it (the group that eventually because the gtld-mou) went off in an entirely different direction. Their plan was for seven new TLDs to be managed by their organization, and anyone who signed up and paid enough money could become a registrar selling names under those TLDs. Which wasn't a bad idea, except for their insane structure and contractual setup--and the idea of putting DNS under ITU.
Meanwhile, the people who'd spent a bunch of money trying to be the first ones to claim exciting names like
Unfortunately, none of these people could get along. eDNS, uDNS, AlterNet... they ended up with a huge mess of infighting and lawsuits. Not to mention the little "joke" AlterNet pulled with the root servers....
While this was going on, the gtld-mou people decided they'd go through the governments of the world--and, when that didn't work, through the ITU.
A third group of people decided that the only way to decide this was through international Internet democracy, and set about coming up with the right way to build such a governance system.
Meanwhile, a company calling itself name.space decided they'd start their own root servers and allow anyone to register any two-level domain name with any TLD, whether it already existed or not. So Rob Malda could have rob.malda, or iam.rob.
Finally, a fifth group (mostly Europeans) wanted to force everyone who wasn't doing something truly international to use the country-code domains (.us,
Then the big question arose: Who actually has the authority to change the system? In the end, only the US government has any authority (they "own" IANA and Internic--and they can make NSI play ball). So some people from the first and third groups, hoping to stop the gtld-mou, went to the US government and began pressuring them to do something.
Somehow, Esther Dyson, in typical Esther Dyson fashion, managed to present herself as the world's foremost expert and the only person who could be trusted to muddle through this.
And in the end, we got ICANN. Backed by the US government, they took the basic gtld-mou idea of a core of registrars and applied it to
So now we get the same arguments all over again. Fun, eh?
Hey, I have an x86 box with a 32MB video card, and only 30MB of system RAM....
Well, I'm sure iCab 1.9 shares a whole lot of code with 1.8, but the whole browser is only about as old as the Mozilla project.
In another post, I mentioned that the main reason Mozilla looks so unfinished is that it was an incredibly ambitious project--far more ambitious than IE 5.0 to 5.5, obviously.
I know that they have lots of justification for the current state of the project. In fact, I'm impressed that they've made it as good as they have.
But the "unwashed masses," as you put it, don't know that. We can all work to educate them, but I still think it's going to hurt public confidence in Netscape.
The Windows version I don't know as much about, but on the Mac, from version 3 to 4, everything feels very different. Everything that used to crash it no longer does (but new things do), there's a completely different set of HTML and layout bugs (with little overlap), the interface with the JVM is totally new it's smaller, it loads its code in a different way (all those hideous System libraries are gone...), etc.
I don't believe what Microsoft says just because they say it, but I think in the case of IE 4.0, they really did rewrite from scratch, just as they claim.
Comparing Office 6.0 to Office 98, I think the same is true there. That's two cases I know of where Microsoft did a complete rewrite (three if you count Outlook Express as a separate product), and throwing away their old code turned out to be a great idea. If only they'd apply that to the next version of Windows....
Yes, GIF animations are mostly annoying. But people still use them, and they're obviously not going to stop any time soon. So it's still a concern.
I'm not sure you want to call something a "real innovation" if every other OS already has it....
Anyway, it's interesting that you should use the Amiga as an example, because back when I was last doing serious BeOS development (this is in the days when BeOS for Intel was just a "sometime in the future" idea, so things may have changed...), a lot of developers were using the datatypes library ported from the Amiga.
It's actually a pretty cool library. If I build my app to use datatypes to draw images, and the user adds a PNG datatype library to her system, my app can magically use PNGs (except for some minor issues with selecting files by extension or MIME type...).
And I think that's the level this should be at--it shouldn't be built into the OS (or the windowing system). I should be able to download the PNG datatype library and install it, and maybe even choose among a few different PNG datatypes.
Anyway, it would be nice if linux had something like that, right? But doesn't it already? Isn't that the point of ImageMagick, and a few other libraries?
By the way, Windows doesn't natively have anything like this. If you use the Win32 bitmap APIs, you only get BMPs, and that's it.
But QuickTime (on both Windows and MacOS) will draw all kinds of images for you painlessly, and every time it's updated, your app magically gets new image types. And under Windows, there are ActiveX controls (including the IE control, but also some lighter-weight choices) that will give you the same thing.
Did you go to the more complete test suite linked on the page? Did gamma really work right? Did alpha transparency without a background (in other words, compositing with the page itself) work? Because I tested IE 5.0 under Win98, and 4.0, 4.5, and 5.0 under MacOS 9, and none of them did. Plus, the image that Netscape tried to load forever actually crashed MacOS 9 with IE 4.0.
I suppose it's possible that later versions of IE have actually gone backwards in PNG compatibility, but I doubt it.
Still, all versions of IE did much better than Netscape 4.7. On the other hand, Mozilla did better than either. As did Opera and iCab. So much for IE, eh?
But not a single browser I have available passed everything (well, lynx did fine passing the PNGs off to an outside viewer...). Which means the problem is serious.
This is an interesting idea. It would probably be easier to modify existing software to use "GIF2000" than to add PNG support, if the format change were kept incredibly simple (ideally, just link with a new library...).
A few minor changes to the spec allows a GIF00a format which is as backwards-compatible as possible (a GIF87a viewer could intelligently skip the images, and a lot of code can be reused). I've got the details below.
Now go ahead and code a reference implementation. If you decode LZW, zlib, and uncompressed, but only encode zlib and uncompressed, you should be patent-free.
A quick summary:
Use "GIF00a" instead of "GIF87a" or "GIF89a" in the header. Take over the 2-bit "Reserved" slot in the "Packed Fields" in the Index Descriptor for a compression method specifier: 0=LZW, 1=zlib, 2=reserved, 3=uncompressed.
In the Table Based Image Data, always use a code size of 0 for zlib or uncompressed data. As with LZW data, break the zlib or uncompressed stream into blocks of up to 255 bytes; prepend each with a block size byte; and append a 0 byte to the end of the last block. The blocks, when you assemble them and strip off the size bytes, make up a zlib data stream, or an uncompressed image.
For zlib data, use the exact same rules as PNG: no pre-defined codes, type-8 (deflate) compression only, and a maximum sliding window size of 32768. As with LZW images, a zlib image which reaches the terminator without specifying a complete zlib datastream is corrupt.
Sections to change:
(change)17.c.ii, (change)20.c, (insert)20.c.ix, (change)20.e, (change)22.a, (change)22.c, (change)22.c.i, (insert)Appendix G, (insert)Appendix H
The TIFF format may be open, but it uses LZW, just like GIF, which means it has the same problems. (Also, there are all kinds of problems with the TIFF spec, but that would be getting way off-topic.)
As for "implementing essentially a JPEG driver," I'm not sure what that means. Any application can use libjpeg or libpng to take a JPEG or PNG and use it. It's really not that hard.
You might want to have wrapper classes that fit in nicely with your foundation (MFC, BeKit, Gtk--, PowerPlant, etc.), but they're pretty easy to write and can be easily supplied by a third party. For example, on one Windows/MFC project, I replaced all the BMPs with PNGs just by swapping in a (free) third-party library with a class called CImage and using it in place of MFC's CBitmap.
Anyway, any application that I work on that needs to display images, I push for PNG as the format to use, and having news stories like this to show to the suits often helps. But on web pages I design, I often find myself having to use GIFs because I want the page to work with most browsers. Sad, but what other choice do I have?
I don't know what you've been gathering, but you might want to read the article. UniSys can charge people to use (LZW-compressed) GIFs on a web page, and is doing so. That's what all the fuss is about.
It has to end it .COM or .EXE? Can't it just be chmod a+x, or have file type APPL?
This is a good question. Let's look at it. Why does anyone pay for Windows (other than the fact that it came "free" with their computer--meaning they already paid for it whether they use it or not)?
Well, there are a large number of things Windows does better than linux. Even if linux is clearly better overall, Windows is much better for playing games, watching ASF movies, reading the stupid Word-formatted memos that everyone emails around on the Exchange mail server at work, browsing the web, working on images, composing music, etc. If vmware didn't exist, I'd actually have to boot to Windows reasonably often (ditto for mol on my PPC machines).
So, why does anyone use GIF instead of PNG?
Again, there are things it does better. For one thing, it shows up in every browser. Transparency works in almost every browser. It does animations. There are better tools for manipulating GIFs.
In other words, while the PNG format is clearly far superior to the GIF format, there are reasons people keep using GIFs.