I'm no fan of Michael, but who's the idiot who moderated this as insightful? Just for the record since some people apparently think this guy has a point, this post is complete bullshit and most likely a troll.
Is anyone using any critical thinking? How can a particular combination of bits on a CD crash your computer, much less "cause damage to your computer"?
If your computer crashes based on a bad CD, then get a new CD-ROM drive because it's a piece of crap.
Assuming Sony is not doing anything physically wrong to the disk (like making it too thick or something absurd), there is no story here.
I don't really keep track of RMS, so I don't have the articles where I've seen him say it that baldly. But read his own writings, and you'll see that it's not hard to see him believing it.
The biggest problem with RMS is that he has said several times that not only does he think all software should be free, but you should be required by law to make your software free. There is no room in his philosophy for people to choose what type of software they want to use.
you must replace your propriety software with free software, software that lets you have those freedoms.
When he says "must", he means it as in, "you will be required to use free software."
Is there anything worse than a zealot who requires everyone to conform to his beliefs in the name of "freedom"?
I don't have a problem with RMS living his life the way he wants to live it. I have a big problem with his shoving his version of "freedom" down my throat. If I want to use closed source, proprietary software, then dammit RMS stay the hell out of my machine.
These are my last posts in this thread. I grow weary of debating this. I'm not even sure why I'm bothering.
No you werent right.
Sorry, but you are totally wrong. Unfortunately, I can't find my copy of the C standard to quote directly from it. However, at this link, I quote:
Minimum Type Limits
Any compiler conforming to the Standard must also respect the following limits with respect to the range of values any particular type may accept. Note that these are lower limits: an implementation is free to exceed any or all of these. Note also that the minimum range for a char is dependent on whether or not a char is considered to be signed or unsigned.
Type Minimum Range
signed char -127 to +127
unsigned char 0 to 255
short int -32767 to +32767
unsigned short int 0 to 65535
int -32767 to +32767
unsigned int 0 to 65535
long int -2147483647 to +2147483647
unsigned long int 0 to 4294967295
Type Minimum Precision
float 6 digits
double 10 digits
long double 10 digits
The Standard also specifies that these limits should be present as preprocessor macros in the header file.id.
And by the way, C++ has the same limits. I quote from Bjarne Stroustrap, The C++ Programming Langauge, Second Edition, Page 50: "In addition, it is guaranteed that a char has at least 8 bits, a short at least 16 bits, and a long at least 32 bits".
There, is that good enough for you?
Dont use 64 bit arithmetic. Use bitwise operators. Thats pretty simple.
It's pretty hard to do a compare operation with bitwise operators. And even if one could, I shouldn't have to.
Now ask yourself this: why define these if they are standard? Could it be, maybe, someone decided that they could be variable depending on the platform you are using? This would explain why a C integer's on a MIPS processor are smaller than those on a i386 processor?
What part of "at least" don't you understand? Standard C guarantees that a byte is at least 8 bits, a short is at least 16 bits, and a long is at least 32 bits. If you don't believe me, please go read the C standard or post a reference. I am right, you are wrong about this.
This is nice for C making small footprints but bad for crossplatform capability.
No, this is nice for performance. It makes it more difficult to achieve cross-platform compatibility, but it's defined that way specifically for cross-platform compatibility. I can write programs in C that are completely portable, yet are still efficient across a large majority of platforms.
It's not even that difficult; it mostly takes experience to know how to do things so that they'll be portable.
Okay, take a 32 bit signed variable in C and go past its maximum range. You will get a negative number. Wow, it works just like Java! Go past the bounds on an unsigned integer and what do your get? A mess.
The point is that I did NOT go past the limits of an unsigned 32 bit variable. I think you don't understand what sign extension means. What that means is that assigning a 32 bit integer of, say, 0xffffffff to a 64 bit integer will get you 0xffffffffffffffff instead of 0x00000000ffffffff. That's why you have to mask it off if you want to simulate unsigned variables.
If you dont want it to become a negative number, use a larger datatype.
Which is why Java is brain damaged as a general purpose language. Yes, you can kludge around the lack of unsigned data types, but I shouldn't have to do very slow 64 bit arithmetic just to be able to deal with a 32 bit unsigned value.
But if you are expecting to read in, say a unsigned int from a fopen, what size is the uint datatype on the different platforms running Unix?
It doesn't matter. If I read in 32 bits, the holder variables just needs to be at least 32 bits. Where you get into trouble is reading an unsigned number and assigning it to a signed variable longer than 32 bits.
Not true. What is the word datatype size on a Palm or that on a Mainframe?
Sheesh, man, please go read the C standard before spouting BS. A long variable is at least 32 bits (actually, they specify a range of values). A short is at least 16 bits. A char is at least 8 bits.
What? There is no unsigned values in Java, so according to your own words masking is STILL not needed.
No. Read what I wrote. Better yet, try an experiment. Assign a long variable (32 bit) to 4,000,000,000. Assign that to a 64 bit variable. Watch it get butchered into a negative value, because Java treats it as a signed value, which is then signed extended.
If you want to "pretend" a 32 bit value is unsigned and assigned it to a 64 bit value, you have to do "a64 = b32 & 0xffffffff".
You are most likely thinking IPs are 32 bits rather than the sum of four bytes with each byte representing an address class of A,B,C, and D.
Sorry, you are wrong. The address classes happen to mask to single bytes, but the subnet masking does not.
The problem occurs in that if they ever run out of addresses and make more, they would add another byte.
What?? What part of your anatomy did you pull that out of? IPv4 is, by definition and unchanging, 32 bits. The next version of TCP/IP is IPv6, which uses 128 bit addresses. Of course, at that point we no longer store IP addresses in a scalar variable and routing will most likely be done in different ways.
Let me know if you have ever heard of a 36-bit datatype in C/C++ or any other language. Im curious.
As a matter of fact, many DEC and IBM computers used 36 bit words. In fact, if I'm not mistaken, the very first version of Unix on the PDP/11 had 36 bit longs. But if you're trying to make some sarcastic point, probably a 40 bit datatype would have worked better (8 * 5 = 40 bits).
Well, first let me say that my crack smoking affected my brain. The time_t type under Unix is a signed type. That said...
Unix data types change size due to platform. So a program you write to compare to 32-bit dates would not work in the future on 64 bit systems as they will use a datatype, called the same, but at a different integal type.
True, but irrelevent. If you are reading a 32 bit unsigned value from a file, it is what it is, 32 bits. If in C I assign that to an "unsigned long", I know I am safe no matter the word size, since C guarantees me that a long is at least 32 bits.
No, just use a larger data type. There is no need for masking in Java as conversion between datatypes is pretty seamless like C.
Nope. If you assign a 32 bit unsigned value that happens to have the high bit set to a 64 bit signed value, it will sign extend the value. You have to mask it off to be safe.
IP addresses consist of four numbers not exceeding 255.
Not true. That's just a convention on how their written. The individual bytes of an IP address have absolutely no intrinsic meaning, although certain fields of bits have meaning depending on whether it's a class A, B or C. An IP address is 32 bits. And incidently, any arbitrary part of those bits can be masked for a subnet; the bits don't even have to be contiguous (well, to be pedantic, any arbitrary part with certain restrictions).
Why would you need a unsigned data type? If the value is too large for the current data type: use one larger. Are you running on a Commodore 64 and worried about the extra 4 bytes?
You mean, move it a larger data type and mask it off. What a pain in the ass, not to mention the performance hit. The question is, why should I have to? And yes, sometimes 4 bytes matters when you have very large arrays.
I actually addressed this the other day, but once again, the obvious case where you need it is in comparing two 32 bits Unix epoch-style dates.
Yes, you can code around the lack of unsigned types. But it's a pain and there's just no excuse for leaving something that critical out of the language, assuming you want to use Java for anything even halfway complex or low level.
Re: Kohler. I have to put in a plug for the low-flush Kohler toilets. I moved into a new house recently, and replaced 4 of the toilets. My old house had a cheap low-flush toilet that I DETESTED. It always plugged up, and never seemed to flush it all the way.
These new Kohlers do a pretty damn good job. I can't say the never plug up, but it's pretty rare and they almost never need a second flush.
I don't recall needing them once in C, and I've written a lot.
You've never compared two 32-bit epoch-formatted dates in Unix? If you used time_t, you used an unsigned type. If you didn't use an unsigned type, your code is broken.
To be honest, if you've written "a lot" of C code, I doubt you've written anything very complex if you've never needed an unsigned type. Communication protocols, compression/decompression algorithms,... hell, I can just remember off-hand that I used an unsigned type in a dynamic resizable hash table algorithm I wrote that used a 16 bit bucket size. If I had to stick with signed integers, it would have a range 32K rather than 64K, which is rather significant.
Man, how about hashing algorithms??? Can you even implement CRC32 without an unsigned data type?
Amusingly, I seem to recall a bug in Sun's 3D libraries where they had to kludge around the lack of an 8 bit unsigned data type, but forgot in a certain instance and the routines failed for high range values. Good thing that "saved" us from the evil, useless unsigned types!
Indeed. There is no doubt that Gosling is a very bright guy, and has a track record of interesting technologies. But all this interview proved is that even smart people can have blind spots and be stupid.
I mean, he can't find ANYTHING positive to say about the technology? How about the multi-language support? And most of even the strident critics of Microsoft who are honest have to admit that there are some interesting ideas in C#.
They might have done something creative around... integrating business logic into the language...
Business logic in the language??? Hey Gosling -- that's the advantage of MULTI LANGUAGE SUPPORT. There's this language called COBOL. Maybe you've heard of it. It has PILES of business logic built into the language.
And, I mean, the fact that the syntax [of C#] is so much -- is like exactly the same, or just about exactly the same [as that of Java].
That's such bullshit. Yeah, and I could also say that "... the syntax [of Java] is [...] just about exactly the same as that of C++". Their both C++ derived languages. Of course they're going to look similar.
The difference is that C# has fixed some of Java's brain damage, one of which is the lack of an unsigned data type which is just unforgivable.
All that proved to me is that Sun is really, really frightened about the potential of.NET. Java is an interesting platform, and an interesting language. But there's a huge opportunity for someone to come in with better solutions, and Sun knows it.
Re:This is what'll screw us all in the end
on
The Root of All E-Mail
·
· Score: 5, Insightful
Security through obscurity will never solve anything when used as the first line of defense.
Dude, it's the first line of defense, not the ONLY line of defense. Read the article.
There is nothing wrong with security through obscurity as one facet of security. It's when it's the only security that it's a problem.
Especially interesting in light of the recent security lockdowns throughout much of the Western world. The havoc of losing the A root server would be bad, like Staypuft Marshmallow Man bad.
Absolute proof that the Slashdot editors don't even bother to read the articles, and just depend on their wrong understanding of things.
From the article...
"The DNS is built so that eight or more of the world's 13 master root servers would have to fail before ordinary Internet users started to see slowdowns, according to John Crain, manager of technical operations for the Internet Corporation for Assigned Names and Numbers (ICANN).
ICANN manages the DNS and sets policies for registry operators and domain name retailers.
"Theoretically, if 'A' were to disappear, we could pick it up from one of the other servers," Crain said. "Moving the place where the zone is picked up is very simple."
In other words, don't panic. The A server is just the highest profile target.
Now you say that security is not an issue...which is it?
Sheesh, I don't know why I'm wasting my time with this, but the issue is that the bugs in Mozilla will not compromise the security of the site. However, that doesn't mean that proper functioning of the site is not important.
I was going to ignore this post since it was so off the deep end, but since you got moderated to +5 apparently someone thinks that your post is relevant to something.
Bottom line, all I'm talking about is the application not functioning correctly because of bugs in Mozilla (which may have since been fixed according to another poster). Security is not an issue in anything.
OK, now imagine a car that explodes when you try and use a standard gas pump. It's not a stretch to imagine that it's better to block those cars and give them a warning rather than let them go on.
I'm not advocating this policy for every site, but for my particular case (where we are handling extremely sensitive data, namely credit reports), it does make sense. It's a vertical market application. Our users are interested in the web site working, and working correctly.
But, your attitude is really poor, and it's attitudes like yours that have made the web as lousy as it is today.
First of all, the web site I work on is not a public web site. It's a vertical market application, so it's not a big deal if we tell them IE only. That's not to say that I wouldn't like to support other browsers, but how much time do I have to waste on Netscape/Mozilla?
Quite frankly, I AM F***ING TIRED OF WORKING AROUND BUGS IN NETSCAPE/MOZILLA.
I'm certain that anyone who has been down this road would sympathize with me. Trying to get Netscape 4.7 to work is out of the question, but I was half-willing to spend some time getting Netscape 6 to work. Then it became the same old story -- I try and make something work, and it just doesn't work. How many hours am I supposed to waste on it?
Trust me, I have way better things to do than code around Netscape/Mozilla bugs.
Slashdot Mirror
I'm no fan of Michael, but who's the idiot who moderated this as insightful? Just for the record since some people apparently think this guy has a point, this post is complete bullshit and most likely a troll.
Is anyone using any critical thinking? How can a particular combination of bits on a CD crash your computer, much less "cause damage to your computer"?
If your computer crashes based on a bad CD, then get a new CD-ROM drive because it's a piece of crap.
Assuming Sony is not doing anything physically wrong to the disk (like making it too thick or something absurd), there is no story here.
I don't really keep track of RMS, so I don't have the articles where I've seen him say it that baldly. But read his own writings, and you'll see that it's not hard to see him believing it.
The biggest problem with RMS is that he has said several times that not only does he think all software should be free, but you should be required by law to make your software free. There is no room in his philosophy for people to choose what type of software they want to use.
you must replace your propriety software with free software, software that lets you have those freedoms.
When he says "must", he means it as in, "you will be required to use free software."
Is there anything worse than a zealot who requires everyone to conform to his beliefs in the name of "freedom"?
I don't have a problem with RMS living his life the way he wants to live it. I have a big problem with his shoving his version of "freedom" down my throat. If I want to use closed source, proprietary software, then dammit RMS stay the hell out of my machine.
These are my last posts in this thread. I grow weary of debating this. I'm not even sure why I'm bothering.
No you werent right.
Sorry, but you are totally wrong. Unfortunately, I can't find my copy of the C standard to quote directly from it. However, at this link, I quote:
And by the way, C++ has the same limits. I quote from Bjarne Stroustrap, The C++ Programming Langauge, Second Edition, Page 50: "In addition, it is guaranteed that a char has at least 8 bits, a short at least 16 bits, and a long at least 32 bits".
There, is that good enough for you?
Dont use 64 bit arithmetic. Use bitwise operators. Thats pretty simple.
It's pretty hard to do a compare operation with bitwise operators. And even if one could, I shouldn't have to.
Now ask yourself this: why define these if they are standard? Could it be, maybe, someone decided that they could be variable depending on the platform you are using? This would explain why a C integer's on a MIPS processor are smaller than those on a i386 processor?
What part of "at least" don't you understand? Standard C guarantees that a byte is at least 8 bits, a short is at least 16 bits, and a long is at least 32 bits. If you don't believe me, please go read the C standard or post a reference. I am right, you are wrong about this.
This is nice for C making small footprints but bad for crossplatform capability.
No, this is nice for performance. It makes it more difficult to achieve cross-platform compatibility, but it's defined that way specifically for cross-platform compatibility. I can write programs in C that are completely portable, yet are still efficient across a large majority of platforms.
It's not even that difficult; it mostly takes experience to know how to do things so that they'll be portable.
Okay, take a 32 bit signed variable in C and go past its maximum range. You will get a negative number. Wow, it works just like Java! Go past the bounds on an unsigned integer and what do your get? A mess.
The point is that I did NOT go past the limits of an unsigned 32 bit variable. I think you don't understand what sign extension means. What that means is that assigning a 32 bit integer of, say, 0xffffffff to a 64 bit integer will get you 0xffffffffffffffff instead of 0x00000000ffffffff. That's why you have to mask it off if you want to simulate unsigned variables.
If you dont want it to become a negative number, use a larger datatype.
Which is why Java is brain damaged as a general purpose language. Yes, you can kludge around the lack of unsigned data types, but I shouldn't have to do very slow 64 bit arithmetic just to be able to deal with a 32 bit unsigned value.
But if you are expecting to read in, say a unsigned int from a fopen, what size is the uint datatype on the different platforms running Unix?
It doesn't matter. If I read in 32 bits, the holder variables just needs to be at least 32 bits. Where you get into trouble is reading an unsigned number and assigning it to a signed variable longer than 32 bits.
Not true. What is the word datatype size on a Palm or that on a Mainframe?
Sheesh, man, please go read the C standard before spouting BS. A long variable is at least 32 bits (actually, they specify a range of values). A short is at least 16 bits. A char is at least 8 bits.
What? There is no unsigned values in Java, so according to your own words masking is STILL not needed.
No. Read what I wrote. Better yet, try an experiment. Assign a long variable (32 bit) to 4,000,000,000. Assign that to a 64 bit variable. Watch it get butchered into a negative value, because Java treats it as a signed value, which is then signed extended.
If you want to "pretend" a 32 bit value is unsigned and assigned it to a 64 bit value, you have to do "a64 = b32 & 0xffffffff".
You are most likely thinking IPs are 32 bits rather than the sum of four bytes with each byte representing an address class of A,B,C, and D.
Sorry, you are wrong. The address classes happen to mask to single bytes, but the subnet masking does not.
The problem occurs in that if they ever run out of addresses and make more, they would add another byte.
What?? What part of your anatomy did you pull that out of? IPv4 is, by definition and unchanging, 32 bits. The next version of TCP/IP is IPv6, which uses 128 bit addresses. Of course, at that point we no longer store IP addresses in a scalar variable and routing will most likely be done in different ways.
Let me know if you have ever heard of a 36-bit datatype in C/C++ or any other language. Im curious.
As a matter of fact, many DEC and IBM computers used 36 bit words. In fact, if I'm not mistaken, the very first version of Unix on the PDP/11 had 36 bit longs. But if you're trying to make some sarcastic point, probably a 40 bit datatype would have worked better (8 * 5 = 40 bits).
I made no point about needing unsigned data types for IP addresses, only that your understanding of the structure of IP addresses is wrong.
Of course, one could argue it's pretty lame to be required to store an unsigned value in a signed variable and just "pretend" it's unsigned.
Well, first let me say that my crack smoking affected my brain. The time_t type under Unix is a signed type. That said...
Unix data types change size due to platform. So a program you write to compare to 32-bit dates would not work in the future on 64 bit systems as they will use a datatype, called the same, but at a different integal type.
True, but irrelevent. If you are reading a 32 bit unsigned value from a file, it is what it is, 32 bits. If in C I assign that to an "unsigned long", I know I am safe no matter the word size, since C guarantees me that a long is at least 32 bits.
No, just use a larger data type. There is no need for masking in Java as conversion between datatypes is pretty seamless like C.
Nope. If you assign a 32 bit unsigned value that happens to have the high bit set to a 64 bit signed value, it will sign extend the value. You have to mask it off to be safe.
IP addresses consist of four numbers not exceeding 255.
Not true. That's just a convention on how their written. The individual bytes of an IP address have absolutely no intrinsic meaning, although certain fields of bits have meaning depending on whether it's a class A, B or C. An IP address is 32 bits. And incidently, any arbitrary part of those bits can be masked for a subnet; the bits don't even have to be contiguous (well, to be pedantic, any arbitrary part with certain restrictions).
Why would you need a unsigned data type? If the value is too large for the current data type: use one larger. Are you running on a Commodore 64 and worried about the extra 4 bytes?
You mean, move it a larger data type and mask it off. What a pain in the ass, not to mention the performance hit. The question is, why should I have to? And yes, sometimes 4 bytes matters when you have very large arrays.
I actually addressed this the other day, but once again, the obvious case where you need it is in comparing two 32 bits Unix epoch-style dates.
Yes, you can code around the lack of unsigned types. But it's a pain and there's just no excuse for leaving something that critical out of the language, assuming you want to use Java for anything even halfway complex or low level.
Who uses java?
People who don't need an unsigned data type.
Re: Kohler. I have to put in a plug for the low-flush Kohler toilets. I moved into a new house recently, and replaced 4 of the toilets. My old house had a cheap low-flush toilet that I DETESTED. It always plugged up, and never seemed to flush it all the way.
These new Kohlers do a pretty damn good job. I can't say the never plug up, but it's pretty rare and they almost never need a second flush.
I don't recall needing them once in C, and I've written a lot.
You've never compared two 32-bit epoch-formatted dates in Unix? If you used time_t, you used an unsigned type. If you didn't use an unsigned type, your code is broken.
To be honest, if you've written "a lot" of C code, I doubt you've written anything very complex if you've never needed an unsigned type. Communication protocols, compression/decompression algorithms, ... hell, I can just remember off-hand that I used an unsigned type in a dynamic resizable hash table algorithm I wrote that used a 16 bit bucket size. If I had to stick with signed integers, it would have a range 32K rather than 64K, which is rather significant.
Man, how about hashing algorithms??? Can you even implement CRC32 without an unsigned data type?
Amusingly, I seem to recall a bug in Sun's 3D libraries where they had to kludge around the lack of an 8 bit unsigned data type, but forgot in a certain instance and the routines failed for high range values. Good thing that "saved" us from the evil, useless unsigned types!
I did a search on RM101's real name. Teoma returns 45 results. Google returns 173 results.
Therefore, Google is clearly the superior search engine, since it returns far more important results. :)
Indeed. There is no doubt that Gosling is a very bright guy, and has a track record of interesting technologies. But all this interview proved is that even smart people can have blind spots and be stupid.
I mean, he can't find ANYTHING positive to say about the technology? How about the multi-language support? And most of even the strident critics of Microsoft who are honest have to admit that there are some interesting ideas in C#.
They might have done something creative around ... integrating business logic into the language ...
Business logic in the language??? Hey Gosling -- that's the advantage of MULTI LANGUAGE SUPPORT. There's this language called COBOL. Maybe you've heard of it. It has PILES of business logic built into the language.
And, I mean, the fact that the syntax [of C#] is so much -- is like exactly the same, or just about exactly the same [as that of Java].
That's such bullshit. Yeah, and I could also say that "... the syntax [of Java] is [...] just about exactly the same as that of C++". Their both C++ derived languages. Of course they're going to look similar.
The difference is that C# has fixed some of Java's brain damage, one of which is the lack of an unsigned data type which is just unforgivable.
All that proved to me is that Sun is really, really frightened about the potential of .NET. Java is an interesting platform, and an interesting language. But there's a huge opportunity for someone to come in with better solutions, and Sun knows it.
Security through obscurity will never solve anything when used as the first line of defense.
Dude, it's the first line of defense, not the ONLY line of defense. Read the article.
There is nothing wrong with security through obscurity as one facet of security. It's when it's the only security that it's a problem.
Hemos said...
Especially interesting in light of the recent security lockdowns throughout much of the Western world. The havoc of losing the A root server would be bad, like Staypuft Marshmallow Man bad.
Absolute proof that the Slashdot editors don't even bother to read the articles, and just depend on their wrong understanding of things.
From the article...
"The DNS is built so that eight or more of the world's 13 master root servers would have to fail before ordinary Internet users started to see slowdowns, according to John Crain, manager of technical operations for the Internet Corporation for Assigned Names and Numbers (ICANN).
ICANN manages the DNS and sets policies for registry operators and domain name retailers.
"Theoretically, if 'A' were to disappear, we could pick it up from one of the other servers," Crain said. "Moving the place where the zone is picked up is very simple."
In other words, don't panic. The A server is just the highest profile target.
Now you say that security is not an issue...which is it?
Sheesh, I don't know why I'm wasting my time with this, but the issue is that the bugs in Mozilla will not compromise the security of the site. However, that doesn't mean that proper functioning of the site is not important.
Indeed. This is like an article saying, "Scientists today captured a GIANT ANT. The feisty Formicidae measured over ONE QUARTER OF AN INCH LONG."
Now, that would be pretty giant by ant standards, but when I think of giant ants, I want cars being crushed and houses being levelled.
It's hard to believe that this made CNN.
I was going to ignore this post since it was so off the deep end, but since you got moderated to +5 apparently someone thinks that your post is relevant to something.
Bottom line, all I'm talking about is the application not functioning correctly because of bugs in Mozilla (which may have since been fixed according to another poster). Security is not an issue in anything.
Imagine a gas station that blocked all Fords.
OK, now imagine a car that explodes when you try and use a standard gas pump. It's not a stretch to imagine that it's better to block those cars and give them a warning rather than let them go on.
I'm not advocating this policy for every site, but for my particular case (where we are handling extremely sensitive data, namely credit reports), it does make sense. It's a vertical market application. Our users are interested in the web site working, and working correctly.
Unfortunately, my actual problem was wanting to set 'display' to 'none', so your trick wouldn't work for that.
But, your attitude is really poor, and it's attitudes like yours that have made the web as lousy as it is today.
First of all, the web site I work on is not a public web site. It's a vertical market application, so it's not a big deal if we tell them IE only. That's not to say that I wouldn't like to support other browsers, but how much time do I have to waste on Netscape/Mozilla?
Quite frankly, I AM F***ING TIRED OF WORKING AROUND BUGS IN NETSCAPE/MOZILLA.
I'm certain that anyone who has been down this road would sympathize with me. Trying to get Netscape 4.7 to work is out of the question, but I was half-willing to spend some time getting Netscape 6 to work. Then it became the same old story -- I try and make something work, and it just doesn't work. How many hours am I supposed to waste on it?
Trust me, I have way better things to do than code around Netscape/Mozilla bugs.