I don't think fonts are often expensive just because it took a long time to create all the letters. It's probably more to it than that.
Good fonts are difficult to make. Making something for a title is easy, but it is a lot of difficult work to create a font in which large blocks of text are easy to read and which does this at many different sizes.
I have Microsoft's verdana font on my linux boxes, because it is an exceptionally good, easy-to-read font. I hope these new fonts are as good, so all I will need are Vera fonts and everything from Divide by Zero.
So they'll do more thorough inspections before reentry - but they still haven't addressed the issue of what to do if they actually find something wrong. As I understand it, there is no capacity to perform such repair work while in orbit.
This all bears repeating. NASA decided that repairing a tile in orbit is not viable because a repair effort will probably result in further damage to the surrounding tiles. Spacewalks are performed by astronauts with years of training specific to spacewalks. Most shuttle astronauts are not trained for this. When there are no EVAs planned, no one is trained and no suits are taken up. Shuttle EVAs are restricted to the bay or the arm, when the arm is carried, which it often isn't. Without a massive redesign, there is no way for a shuttle astronaut to even see the underside of the shuttle, or get there on an EVA, or anchor himself there to do any work.
The shuttle is a production model, not an experimental model. An experimental vessel would have all sorts of redundancies and extra equipment and measurements made for no reason except that you never know what would be useful. A production model is tuned to perform adequately under expected conditions with cost minimized. It is hard to say that this is bad, because it was one of the goals of the shuttle program.
I don't see many comments about the other story linked to about the Sports Illustrated cover. It took me several minutes to figure out what the largest alteration was. (And it wasn't mentioned.) It is that the picture was rotated. What's the difference? The player committing the facemasking is much more menacing in the cover because he appears larger due to the rotation. This fits the editorial slant (that the Ravens are meanies).
YOU CAN'T TRANSMIT AN ARBITRARILY LARGE AMOUNT OF DATA/SECOND ON A FINITE AMOUNT OF BANDWIDTH.
I can't believe how far down the page I had to read before someone said this.
Back of the envelope: Let's say you want approximately CD quality sound from your radio. I believe the parameters are 32k samples per second, 12 bits per sample, two channels, for 800kbits per second. To get this, your radio stations have to be about 800 kHz apart (within a factor of two or so). Of course, we don't get sound that good from our radios, but it gives you an idea.
Of course, if all you want are carrier waves with no signal on them, then they can be packed much more tightly.
The photons may not "interfere" (Reed means "scatter"), but if the finite-width frequency bands of two radio stations overlap, then there is interference.
I recognize that the Salon article may completely misrepresent Reed's work, and that the nonsense is theirs, not his.
What I found astounding was Dr. Reed's willingness to wave away two hundred years of well-established physics. Waves of the same frequency crossing the same point in space do interfere. How do I know? Because the very definition of interference is the effect they have.
What Reed is calling "interference" w.r.t photons is what a reputable scientist would call "scattering".
Religious descisions (for both the believer and the non-believer) are descisions of faith and experience. No amount of science will (or can) ever change this.
That is fine as long as you accept this:
Scientific decisions are decisions of observation and experimentation. No amount of religion will (or can) ever change this.
The oldest records of civilizations date back around 5000 years
The oldest living trees (determined by tree rings on the same tree - not radiocarbon dating) are around 5000 years as well. Though there is no reason trees can't live longer.
Flood stories exist in many (most?) world cultures
To account for problems with evolutionary theory, a new theory, Punctuated Equilibrium has gained prominence
(end quote)
You can always prove number 1, because you will define "civilization" as whatever existed 5000 years ago. If I talk about 30,000 year old cave paintings, you can just say that they don't signify civilization.
Since the last ice age ended around 10,000 years ago, no tree can be 20,000 years old because its climate would have changed too drastically for it to survive.
Flood stories exist in most world cultures because it rains on most world cultures.
You confuse the principle of a theory with the application of the theory. If a theory says that new species arise due to natural selection and evolution, that doesn't tell us anything about the population dynamics, rate of evolution, or why two populations may find interbreeding uninteresting. If I can't fix your television, that doesn't mean that there is an error in Maxwell's equations.
Behe is argumenting from ignorance throughout the whole book, saying that since we cannot figure out how a complex system could have evolved through natural selection, it must be the work of a supreme being.
That should be "since he cannot figure out..."
This sounds like "I have invented a new encryption scheme. I can't break it, so it must be secure."
The discoverer must propose new laws of nature to explain an observation
I think people are missing the fundamental point, that you need to apply these warning signs together.
It's stupidly obvious that if you throw out any proposed new laws of nature, then you will never have any new laws of nature or make any progress!
The important combination here is proposing new laws based on new suspect data. If the data is questionable (at the limits of detection) and new pysical laws must be invented to explain it, then you are on dangerous ground.
And Parks isn't even saying you can't propose such things and work on verifying them later, he is saying you shouldn't rely on the validity of such claims in a court of law.
3. Plenty of real scientific research happens at the limits of detection. As I recall, Einstein's relativity was an example of this at the time he proposed it. Quantum physics and the outer limits of astronomy are further examples.
To add to one responder's example of the photoelectric effect, I would point out that the emission spectra from different atoms were well known and easy to measure, not to mention that the fact that atoms exist at all in a stable state cannot be explained without QM.
Einstein's special relativity served to unify mechanics and E&M. The inconsistency between mechanics and E&M was well known, and many people were working on theories to solve the problem and experiments to test the theories. Verifying the correctness of the details of SR may have involved difficult measurements, but the theory was developed to explain a gross discrepency in existing physics.
I think your argument that "the conditions could just as readily be applied to certain particularly brilliant scientific breakthroughs" comes from a lack of knowledge of science.
This may be because of the way history is recorded and studied. If ten people were working on something, and one got it done first, we don't hear about the other nine. Long before calculus was invented, mathematicians knew that the area under x^n was obtained by evaluating 1/(n+1)*x^(n+1) at the end points. Darwin finally published his work on evolution only because someone else had worked it out independently, and Darwin wanted credit for being first. The famous equations from special relativity weren't worked out by Einstein, but he was the one who figured out the right way to interpret and apply them. The formula for the energy levels in the hydrogen atom was known before QM could predict it.
These warning signs are more accurate than you think.
Trade secrets: Beyond a doubt there are piles of things in the source code that could be considered trade secrets. One way to protect trade secrets is to make certain that they are widely available but not legally available. In the cynic's view (i.e. mine) M$ wants the code to be leaked by China.
If the code is illegally leaked, it is very easy for M$ to accuse other products (future Linux apps?) of using illegally acquired trade secrets. How can the authors, living in countries around the world, prove that none of them have ever seen illegally leaked material?
Based on what I have read about the development of the clone of the IBM BIOS, it appears that the burden of proof de facto lies on the defendant to show that they are not using trade secrets illegally.
This may give M$ a very big gun to point at any colloboratively developed code that they don't care for.
It should be stripped down to only those features that I like, arranged in a manner that helps me work the way I want to. Nothing more, nothing less.
How that affects *you* is *your* problem.
You have hit on the central absurdity of the way this discussion is phrased. What "should" an interface be like. This implies that there exists some universal value judgement that can be used to rate interface features as good or bad. There isn't, so people just blather on and nothing meaningful is said. So what values do people use in determining "should"? Most often on/. I see the criteria as being how some feature will help drive M$ out of business. If you read the ravings about usability at joelonsoftware, you realize that his underlying value system is how marketable the software will be. That is very important to the people writing the software, but is completely irrelevent to the people using the software.
Interfaces to programs I use should be highly configurable, and I don't care about the rest. Other people can use whatever they want, as long as they don't complain to me about it. Actually, I should rephrase. Interfaces to most programs I use will be highly configurable, because if they aren't, I won't use them. (A million thanks to the fvwm guys.)
Whaddya mean outlaw "forged" headers? Most email I send had "forged" headers on it, because I am not sending it from a mail server. So, duh, I put in a "forged" From: line so replies go to the mail server, rather than to a machine that doesn't even listen on the SMTP port. What about masquarading in sendmail, will that be illegal too?
The only headers that should be preserved are perhaps the Received: lines which show that route that the message has taken. Still, I can think of a legitimate reason to muck with these - if a company network has a sufficiently complicated internal structure, these headers might reveal some information that they don't want widely available.
I find it disturbing that there is so little comment on perhaps the biggest issue here, the effect the NYT reporting had on the process Mr. Mitnick was due. If the NYT--the "paper of record"--was so inaccurate about the facts of this case, how can we trust any of its content?
The NYT once did a massive piece of re-reporting on the Wen Ho Lee case. They admitted that their contemporaneous coverage was highly inaccurate, and presented a lengthy thoroughly-researched article to report the facts correctly. This took a few years, as I recall.
It sounds like Kevin deserves one of these quasi-retractions.
People have to go to broadband. Too much of the web is becoming unusable over a modem. This is probably mostly due to graphics. How many commercial sites do you see have a gif of the word "site map" or whatever rather than the ascii text? Almost all of them. Slow connections also sometimes result in the download hanging. No big deal? It is if you are trying to buy airline tickets 2 weeks in advance where the prices and availability change every few minutes and a disrupted session can cost you hundreds of dollars.
Most people who have the option wait to do some of their surfing at work on a fast connection. For people who don't have that option, they must pay for a faster connection or accept that some web content and services will be barely usable or unavailable.
The Enterprise was never meant to go into space and is only for training inside the atmosphere.
They were talking about the Enterprise on NPR this morning. The Enterprise was built to prove that the shuttle was physically capable of landing safely. It would be carried up piggy-back on a plane (747?) and let loose so it could glide in for a landing.
They also said that the ship was named Enterprise at the behest of hordes of Trekkies.
Has anyone ever explored a sort of "salvage use" copyright? In other words, I don't have a huge problem with Disney wanting to control a copyright on Mickey, because they're still actively using him in their business.
I think there is a confusion of two different things throughout these discussions, trademark and copyright.
Copyright means you can't duplicate Steamboat Micky and sell the copies in the stall at the mall.
Trademark means you can't sell mugs with your own drawing of Mickey Mouse on them.
I don't see that the lapse of the copyright on Steamboat Mickey would give anyone the right to create and sell new work with Mickey Mouse, because it would violate Disney's trademark.
What is the difference? There are laws about how trademarks must be used and defended to remain valid. This seems to tie into the "in active use" idea people here want to apply to copyrights.
As an aside, I went looking for works by Mark Twain, one of the greatest American authors. Most of it is out of print and unavailable. Too bad. And then I found it on Project Gutenburg's collection of public domain material. Hooray! If it had been under copyright, the works would not have been available. Who does that help?
His crazy plot for Ep III sounds interesting...I don't believe that Lucas would be smart enough (or would let his ego deflate enough) to use this terrific plot that has been presented infront of him.
One of Brin's Salon articles from 1999 contains this quote: "How come we never see Yoda take on an enemy with a light saber? Come on master, fire it up and battle a Sith Lord! That's a battle I'd pay to see!" Maybe Lucas does pay attention.
He certainly dropped that stupid midichlorian crap quickly enough.
Or are you advocating lousy design? The Web does not, and the Web should not look the same to everyone. If you want complex graphical presentation, go with PDF.
No, you're missing the point. Use PDF when you need fixed presentation, when you want to control every pixel on the page.
I am simply talking about using a variety of fonts, colors and so on to improve the site. It is certainly reasonable if I want to put "Hi Mom" in green italics on my home page (so she'll notice). I don't care about exact placement, I just want it in green italics if supported and if someone has set their browser italics font to adobe-courier it won't kill me. Let's say I also want "Hi Dad" in purple. The markup approach would be to define classes for maternalgreeting, paternalgreeting, and so on in the CSS to describe the format, and this is too much work to do something this simple.
Note that I am not saying markup is bad, or never useful or important. I am just saying that markup and CSS are sometimes a clunky, inelegant way to describe a web page.
And the title "Z advocates lousy markup" refers to his use of h2 for an element which is not a header.
We always hear that everything would be alright if people used modern browsers, good CSS and so on. But there's a problem here. HTML is a soi-disant markup language (hee hee, I am so pretentious). From personal experience, I can say that a markup language (like latex) is far superior to manual formatting for technical publications. This is partly because there are a small number of elements in the design and the format is highly structured, so it makes much more sense to identify something as a chapter heading in the body and describe how a chapter heading should look in a format document.
Now we get to the web where in many cases the presentation is creative and graphical. Zeldman gives an example a site's "Join Now" text being inside font tags which are inside a table. He suggests replacing this with <h2>Join Now</h2> and add CSS stuff for h2. This is wrong! The text is not a heading, and <h2> should not be used. It is simply some text. The correct markup way to do this is to define a class p.joinnow in the CSS, and in the body use <P CLASS="joinnow">Join Now</P>.
This is utter insanity. You could end up with a separate class for almost every element. A markup language just doesn't make sense for complex graphical presentation where most elements are one-of-a-kind. It is just wrong. Most of the web doesn't fall into this category, and html-css is the right way to describe the content, but some of it does, and trying to squish it into a markup language causes many of the problems we see now.
You don't need to reverse the stack direction. All the matters is that arrays/strings are written in the same direction the stack grows. In other words, leave the stack the same and write the strings backward. An overflow would run off the stack, and it could not overwrite the return address of the routine and force execution of malicious code. Instead, it would just run into empty space reserved for the stack.
It can't be that easy, can it? (It wouldn't stop all possible stack smashing attacks, but many.)
The argument about 'unix' having become generic failed in both cases.
The report indicates that they are not allowed to addres the question of whether UNIX (UNIX is a registered trademark of The Open Group in the United States and other countries) has become generic. All they can consider is whether The Open Group has a trademark on UNIX (UNIX is a registered trademark of The Open Group in the United States and other countries).
This is similar to the issue of patents: When someone is accused of violating a patent, they cannot argue that the patent is invalid. Getting a patent declared invalid is a completely separate legal procedure. (I can see that there is a source of confusion for USAns here. When someone is accused of violating a law, if the appeals go high enough it is possible for the law to be declared unconstitutional, but appeal of a patent violation will never consider the validity of the patent.)
While it is not relevant to the case, I think that there is a good argument that UNIX (UNIX is a registered trademark of The Open Group in the United States and other countries) has become a generic term. There is no such thing as "A UNIX" (UNIX is a registered trademark of The Open Group in the United States and other countries). The Open Group specfically states that a valid use of "UNIX" (UNIX is a registered trademark of The Open Group in the United States and other countries) is in an advertisement for "MyProduct word processor for the UNIX* operating system". ("It is acceptable to use an asterisk in place of the trademark symbol where the medium used... cannot reproduce the... symbols"). This shows the absurdity of their claim, for there is no program that works on all UNIX* operating systems (hello world excepted). I submit without justification that no reasonably useful program can be written just to the standard; i.e it compiles on all certified UNIX* systems with no ifdefs.
Furthermore, The Open Group certifications listed on the web site apparently show that none of the BSDs are UNIX* systems. But there is NCR UNIX, so we have UNIX* for cash registers, woohoo! I think that they can trademark their certifications, i.e. "UNIX 98", but term UNIX* itself is moving close to being generic.
MHO, the Mozilla developers made a very bad decision when they decided to create their own GUI toolkit from scratch rather than rely on the interface of each operating system Mozilla ran on. Sure, Mozilla's controls look the same on Mac OS X as they do on Windows and Linux and Be and OS/2 and OpenVMS... but who cares?
Microsoft cares a whole lot.
Why did the browser war start? Because MS was terrified of netscape becoming a platform itself. The softies imagined a future in which people did stuff by using netscape to interact with servers. If this happened, it wouldn't matter what the OS was as long as it ran netscape. This is what made MS flip out.
So they killed netscape, but moz is rising from the ashes, with a much more versatile interface. Get people used to moz/NS, and they'll be happy on windoze, linux boxes, X terminals...
Exactly, software should work, in default mode, like the other software on that platform. That is fundamental UI that the open-source community feels perfectly happy to neglect.
A foolish consistency is the hobgloblin of little minds. This is where the confusion between usability and marketability comes in. To follow your idea to the extreme, mozilla will not be a success unless it looks and acts exactly like IE. What purpose is there in having mozilla if it is made indistinguishable from IE?
It is certainly clear that a program for windows, lacking some spectacular feature, will sell only if it follows the arbitrary conventions of the windows interface. But no one is trying to make a profit selling mozilla for windows.
Back to foolish consistency: a program should follow the conventions of other software only if it does not decrease the usability. As an example, the most useful menu in emacs is the buffer menu; everything else is either seldom used or more easily accessed from the keyboard. This menu has been moved in version 21 from its prime location at the left so that the File menu can be in its "conventional" location. Maybe some people are comforted by this bit of familiarity, but this should not be confused with usability.
Remember, the moz interface is easily and infinitely malleable. This is a program where people could test out hundreds new ideas on interface design, now that it is (mostly) stable.
Slashdot Mirror
I have Microsoft's verdana font on my linux boxes, because it is an exceptionally good, easy-to-read font. I hope these new fonts are as good, so all I will need are Vera fonts and everything from Divide by Zero.
The shuttle is a production model, not an experimental model. An experimental vessel would have all sorts of redundancies and extra equipment and measurements made for no reason except that you never know what would be useful. A production model is tuned to perform adequately under expected conditions with cost minimized. It is hard to say that this is bad, because it was one of the goals of the shuttle program.
I don't see many comments about the other story linked to about the Sports Illustrated cover. It took me several minutes to figure out what the largest alteration was. (And it wasn't mentioned.) It is that the picture was rotated. What's the difference? The player committing the facemasking is much more menacing in the cover because he appears larger due to the rotation. This fits the editorial slant (that the Ravens are meanies).
Back of the envelope: Let's say you want approximately CD quality sound from your radio. I believe the parameters are 32k samples per second, 12 bits per sample, two channels, for 800kbits per second. To get this, your radio stations have to be about 800 kHz apart (within a factor of two or so). Of course, we don't get sound that good from our radios, but it gives you an idea.
Of course, if all you want are carrier waves with no signal on them, then they can be packed much more tightly.
The photons may not "interfere" (Reed means "scatter"), but if the finite-width frequency bands of two radio stations overlap, then there is interference.
I recognize that the Salon article may completely misrepresent Reed's work, and that the nonsense is theirs, not his.
Scientific decisions are decisions of observation and experimentation. No amount of religion will (or can) ever change this.
Since the last ice age ended around 10,000 years ago, no tree can be 20,000 years old because its climate would have changed too drastically for it to survive.
Flood stories exist in most world cultures because it rains on most world cultures.
You confuse the principle of a theory with the application of the theory. If a theory says that new species arise due to natural selection and evolution, that doesn't tell us anything about the population dynamics, rate of evolution, or why two populations may find interbreeding uninteresting. If I can't fix your television, that doesn't mean that there is an error in Maxwell's equations.
This sounds like "I have invented a new encryption scheme. I can't break it, so it must be secure."
It's stupidly obvious that if you throw out any proposed new laws of nature, then you will never have any new laws of nature or make any progress!
The important combination here is proposing new laws based on new suspect data. If the data is questionable (at the limits of detection) and new pysical laws must be invented to explain it, then you are on dangerous ground.
And Parks isn't even saying you can't propose such things and work on verifying them later, he is saying you shouldn't rely on the validity of such claims in a court of law.
Einstein's special relativity served to unify mechanics and E&M. The inconsistency between mechanics and E&M was well known, and many people were working on theories to solve the problem and experiments to test the theories. Verifying the correctness of the details of SR may have involved difficult measurements, but the theory was developed to explain a gross discrepency in existing physics.
I think your argument that "the conditions could just as readily be applied to certain particularly brilliant scientific breakthroughs" comes from a lack of knowledge of science.
This may be because of the way history is recorded and studied. If ten people were working on something, and one got it done first, we don't hear about the other nine. Long before calculus was invented, mathematicians knew that the area under x^n was obtained by evaluating 1/(n+1)*x^(n+1) at the end points. Darwin finally published his work on evolution only because someone else had worked it out independently, and Darwin wanted credit for being first. The famous equations from special relativity weren't worked out by Einstein, but he was the one who figured out the right way to interpret and apply them. The formula for the energy levels in the hydrogen atom was known before QM could predict it.
These warning signs are more accurate than you think.
Trade secrets: Beyond a doubt there are piles of things in the source code that could be considered trade secrets. One way to protect trade secrets is to make certain that they are widely available but not legally available. In the cynic's view (i.e. mine) M$ wants the code to be leaked by China.
If the code is illegally leaked, it is very easy for M$ to accuse other products (future Linux apps?) of using illegally acquired trade secrets. How can the authors, living in countries around the world, prove that none of them have ever seen illegally leaked material?
Based on what I have read about the development of the clone of the IBM BIOS, it appears that the burden of proof de facto lies on the defendant to show that they are not using trade secrets illegally.
This may give M$ a very big gun to point at any colloboratively developed code that they don't care for.
You have hit on the central absurdity of the way this discussion is phrased. What "should" an interface be like. This implies that there exists some universal value judgement that can be used to rate interface features as good or bad. There isn't, so people just blather on and nothing meaningful is said. So what values do people use in determining "should"? Most often on /. I see the criteria as being how some feature will help drive M$ out of business. If you read the ravings about usability at joelonsoftware, you realize that his underlying value system is how marketable the software will be. That is very important to the people writing the software, but is completely irrelevent to the people using the software.
Interfaces to programs I use should be highly configurable, and I don't care about the rest. Other people can use whatever they want, as long as they don't complain to me about it. Actually, I should rephrase. Interfaces to most programs I use will be highly configurable, because if they aren't, I won't use them. (A million thanks to the fvwm guys.)
The only headers that should be preserved are perhaps the Received: lines which show that route that the message has taken. Still, I can think of a legitimate reason to muck with these - if a company network has a sufficiently complicated internal structure, these headers might reveal some information that they don't want widely available.
The NYT once did a massive piece of re-reporting on the Wen Ho Lee case. They admitted that their contemporaneous coverage was highly inaccurate, and presented a lengthy thoroughly-researched article to report the facts correctly. This took a few years, as I recall.
It sounds like Kevin deserves one of these quasi-retractions.
Most people who have the option wait to do some of their surfing at work on a fast connection. For people who don't have that option, they must pay for a faster connection or accept that some web content and services will be barely usable or unavailable.
They also said that the ship was named Enterprise at the behest of hordes of Trekkies.
I think there is a confusion of two different things throughout these discussions, trademark and copyright.
Copyright means you can't duplicate Steamboat Micky and sell the copies in the stall at the mall.
Trademark means you can't sell mugs with your own drawing of Mickey Mouse on them.
I don't see that the lapse of the copyright on Steamboat Mickey would give anyone the right to create and sell new work with Mickey Mouse, because it would violate Disney's trademark.
What is the difference? There are laws about how trademarks must be used and defended to remain valid. This seems to tie into the "in active use" idea people here want to apply to copyrights.
As an aside, I went looking for works by Mark Twain, one of the greatest American authors. Most of it is out of print and unavailable. Too bad. And then I found it on Project Gutenburg's collection of public domain material. Hooray! If it had been under copyright, the works would not have been available. Who does that help?
Especially since we all post to \. using our real names...
One of Brin's Salon articles from 1999 contains this quote: "How come we never see Yoda take on an enemy with a light saber? Come on master, fire it up and battle a Sith Lord! That's a battle I'd pay to see!" Maybe Lucas does pay attention.
He certainly dropped that stupid midichlorian crap quickly enough.
No, you're missing the point. Use PDF when you need fixed presentation, when you want to control every pixel on the page.
I am simply talking about using a variety of fonts, colors and so on to improve the site. It is certainly reasonable if I want to put "Hi Mom" in green italics on my home page (so she'll notice). I don't care about exact placement, I just want it in green italics if supported and if someone has set their browser italics font to adobe-courier it won't kill me. Let's say I also want "Hi Dad" in purple. The markup approach would be to define classes for maternalgreeting, paternalgreeting, and so on in the CSS to describe the format, and this is too much work to do something this simple.
Note that I am not saying markup is bad, or never useful or important. I am just saying that markup and CSS are sometimes a clunky, inelegant way to describe a web page.
And the title "Z advocates lousy markup" refers to his use of h2 for an element which is not a header.
Now we get to the web where in many cases the presentation is creative and graphical. Zeldman gives an example a site's "Join Now" text being inside font tags which are inside a table. He suggests replacing this with <h2>Join Now</h2> and add CSS stuff for h2. This is wrong! The text is not a heading, and <h2> should not be used. It is simply some text. The correct markup way to do this is to define a class p.joinnow in the CSS, and in the body use <P CLASS="joinnow">Join Now</P>.
This is utter insanity. You could end up with a separate class for almost every element. A markup language just doesn't make sense for complex graphical presentation where most elements are one-of-a-kind. It is just wrong. Most of the web doesn't fall into this category, and html-css is the right way to describe the content, but some of it does, and trying to squish it into a markup language causes many of the problems we see now.
It can't be that easy, can it? (It wouldn't stop all possible stack smashing attacks, but many.)
This is similar to the issue of patents: When someone is accused of violating a patent, they cannot argue that the patent is invalid. Getting a patent declared invalid is a completely separate legal procedure. (I can see that there is a source of confusion for USAns here. When someone is accused of violating a law, if the appeals go high enough it is possible for the law to be declared unconstitutional, but appeal of a patent violation will never consider the validity of the patent.)
While it is not relevant to the case, I think that there is a good argument that UNIX (UNIX is a registered trademark of The Open Group in the United States and other countries) has become a generic term. There is no such thing as "A UNIX" (UNIX is a registered trademark of The Open Group in the United States and other countries). The Open Group specfically states that a valid use of "UNIX" (UNIX is a registered trademark of The Open Group in the United States and other countries) is in an advertisement for "MyProduct word processor for the UNIX* operating system". ("It is acceptable to use an asterisk in place of the trademark symbol where the medium used... cannot reproduce the... symbols"). This shows the absurdity of their claim, for there is no program that works on all UNIX* operating systems (hello world excepted). I submit without justification that no reasonably useful program can be written just to the standard; i.e it compiles on all certified UNIX* systems with no ifdefs.
Furthermore, The Open Group certifications listed on the web site apparently show that none of the BSDs are UNIX* systems. But there is NCR UNIX, so we have UNIX* for cash registers, woohoo! I think that they can trademark their certifications, i.e. "UNIX 98", but term UNIX* itself is moving close to being generic.
Microsoft cares a whole lot.
Why did the browser war start? Because MS was terrified of netscape becoming a platform itself. The softies imagined a future in which people did stuff by using netscape to interact with servers. If this happened, it wouldn't matter what the OS was as long as it ran netscape. This is what made MS flip out.
So they killed netscape, but moz is rising from the ashes, with a much more versatile interface. Get people used to moz/NS, and they'll be happy on windoze, linux boxes, X terminals...
Oh, someone does care very much indeed.
It is certainly clear that a program for windows, lacking some spectacular feature, will sell only if it follows the arbitrary conventions of the windows interface. But no one is trying to make a profit selling mozilla for windows.
Back to foolish consistency: a program should follow the conventions of other software only if it does not decrease the usability. As an example, the most useful menu in emacs is the buffer menu; everything else is either seldom used or more easily accessed from the keyboard. This menu has been moved in version 21 from its prime location at the left so that the File menu can be in its "conventional" location. Maybe some people are comforted by this bit of familiarity, but this should not be confused with usability.
Remember, the moz interface is easily and infinitely malleable. This is a program where people could test out hundreds new ideas on interface design, now that it is (mostly) stable.