That's what I'd like to know: what can you do in Perl that you can't do in Python? My guess is nothing. As for what you can do in Python that you can't do in Perl: Python code is actually maintainable. Perl code generally isn't. And Perl makes it hard as hell to make code that even comes close to readable.
What is it that turns Linux users into crusaders against Windows?
The fact that Perl requires goofy symbols before variables depending on their type pretty well precludes the possibility of writing readable Perl code. That reminds me of that old Perl joke. "I just wrote this Perl script that does all sorts of cool stuff in only 8 characters! Actully, I could've written it in only 6, but this is more readable."
I have yet to see readable Perl code. Today, on those rare occasions when I have to write Perl, I often write it in Python first, and then convert it to Perl. The Perl code looks like the Python code would after being sent over a very noisy serial line with no error correction...
Yes, Python isn't as popular as Perl, but popularity is for Lemmings. I'll take technical superiority any day. Python gets the job done easier than Perl, and if/when bugs need to be fixed or features need to be added, I don't need to pull may hair out trying to understand an unreadable script.
Huh? So how do you read C code that's printed on paper? Do you count the braces, or pay attention to the indentation? I doubt it's the former, and the latter is the same as Python.
I've used both Perl and Python quite a bit. I have to say that Python code tends to be somewhat easier to write, and significantly easier to read.
I at first thought Python's "indentation == block structure" was pretty gross, but it actually makes sense. People pay attention to indentation, so it only makes sense that the computer should too.
But really, anyone who bases their conviction that Perl is better than Python because they think Python's indentation thing is "yucky" obviously hasn't taken a real look at Python, and deserves to be ignored.
I used to use Perl. Then I went and learned Python despite my reservations about it's indentation scheme. Since then, I've converted to Python. Anything Perl is good at, Python is better at. It's easier to write, and easier to read. Even before I learned Python, I found it to be easier to read than Perl. I normally don't "convert" to a language. I know more languages than I have fingers, and quite a few of them I use on a regular basis.
The only "advantage" Perl has is that it's installed on more systems than Python. That's the only reason I use Perl at all actually. When I have to write a script that will run on a machine that has Perl but not Python installed. That is (thankfully) becoming more rare.
My point is that while having C vs Java arguments (for example) are futile -- each languange has its valid uses -- Perl is essentially holding us hostage. There's no reason to use Perl, because Python is strictly better. No reason that is, except for the large installed base of Perl. I won't even mention the obvious parallel to Windows and Linux. D'oh! Looks like I just did.
Someone asked for some Python sample code. I was going to post some, but/. seems to have a bug where even stuff posted as "HTML Formatted" undergoes "entity translation". Of course, that doesn't have anything to do with the Perl code in/. being hard to maintain... right?
That's funny. I seem to recall that the IIe was written "][e", but the IIc was written "//c". I think the earlier ones (plain II and II+) were actually written with I's, but I can't be sure.
_Fountains_of_Paradise by Arthur C. Clarke also had microfilaments. In that book they were very think carbon chains or tubes I believe. Today we'd call the tube form Bucky-Tubes. The ends could turn into some other structure that wouldn't cut through things. You could then have one end fastened to a handle, and then in the other end have a chunk of superconductor. A strong electromagnet in the handle would keep the bucky-tube tight. (the superconductor would be repelled by the electromagnet) Not a light-saber, but a pretty neat high-tech "blade" weapon. The blade would probably be close to invisible too...
Come to think of it, such an idea might've been proposed in a Larry Niven story, though I can't remember what, or even if...
The previous story was about Netscape revamping their search engine to use Google technology. Why does AOL also need a new "Inktomi powered" search engine? They own Netscape...
I think you're quoting out of context. I believe he meant "the first Unix implementation [of TCP/IP] was done in berkeley". He was wrong in thinking that that wasn't the first TCP/IP implementation though, AFAIK.
Fine then, How about counterfeiting money. Is that wrong? I mean, if you go print off $100,000 nobody else is losing any money, are they? You're just gaining.
Now think about what would happen if that was socially acceptable and/or legal. How much would money be worth?
In the same way, if software piracy was legal and socially-acceptable, more people would do it, hence more profit would be lost.
Yes, the "lost profits" reported by software companies do tend to be inflated. It is true that only some of those who pirate would buy the software if they couldn't pirate it. But some of them would.
Trying to justify piracy by saying "it doesn't hurt anyone" will only increase the amount of piracy that goes on, thereby causing many people that currently buy software to instead pirate it.
Software piracy *does* hurt the publishers. Sure, not by as much as (pirated copies)*(retail price), but probably a significant fraction of that. Just because they don't lose *all* of that profit doesn't mean it isn't wrong.
If you can't afford it, tough. If you think the asking price isn't worth it, go elsewhere, or write it yourself. If you think it's okay to pirate because software is "so low quality", then:
1. why are you bothering to pirate it? 2. why don't you write something better?
To further strengthen that argument, that the browser is the unctionality, not the code, look at Windows NT 4.0. NT4 Workstation and Server consist of the same code. The only thing that's different is the functionality, which is set by some registry entries. Remember to O'Reilly hack to turn NT Workstation into NT Server?
At least, IDE's that don't let you integrate whatever editor you want suck. I'd rather use the command line if the IDE won't let me edit with VIM.
And I agree that these things should really be written in Java. Do the C/C++ programmers writing these IDE's know what a Java developer wants in their environment. A lot of the needs are the same, but some are quite different. KLGroup's JProbe Memory debugger is an example of a tool that makes sens for Java, but not something like C or C++. (It lets you find out what objects aren't being collected because they've still got references. C/C++ memory leak tools are quite different.)
Those who benefit from the current system where programs are property offer two arguments in support of their claims to own programs: the emotional argument and the economic argument.
The emotional argument goes like this: ``I put my sweat, my heart, my soul into this program. It comes from me, it's mine!''
This argument does not require serious refutation. The feeling of attachment is one that programmers can cultivate when it suits them; it is not inevitable. Consider, for example, how willingly the same programmers usually sign over all rights to a large corporation for a salary; the emotional attachment mysteriously vanishes. By contrast, consider the great artists and artisans of medieval times, who didn't even sign their names to their work. To them, the name of the artist was not important. What mattered was that the work was done--and the purpose it would serve. This view prevailed for hundreds of years.
So why can't RMS be like "the great artists and artisans of medieval times"? The name GNU implies RMS. He wants it be called GNU/Linux for his own personal gratification. It's as simple as that.
I think the writers of free software do deserve credit. But according to RMS, people don't deserve, nor should they desire, credit for their work. According to him, that's just an excuse for proprietary non-free software. Yet when it comes to his own foundation's name, he demands that people give the FSF, and hence RMS, credit. That seems awfully hypocritical to me.
Re:Sigh... No perfect solution in BSD or GPL.
on
BSD vs GPL
·
· Score: 1
With the BSD license, some big wealthy corporation could hijack some code I spent hours and hours and sweat and blood writing, make some utterly trivial change to it, then distribute it under a proprietary license and make a huge profit and I never see a penny. In addition, if they made VALUABLE changes, they would be closed, not keeping with the "open source" spirit under which the original code was written. Licensing under GPL prevents this.
The GPL only partially protects against this. It does not protect against cases where:
- you write an amazing program - some big corporations takes it, and makea a few insignificant tweaks - big corporation sells "distributions" of your code, while you get nothing
This, to me, seems to be the big problem with both BSD and GPL. They both allow people to gain equal profit regardless of how little work they've done. Because we live in a capitalist society, the end result is that those who do more work get less. That's why open source developers have to get real jobs, and can only work on free projects in our spare time. That "Commercial Open Source" idea is looking better all the time.
Q: Wouldn't the web be much better if it was all LaTeX based?
A: No, it would suck even worse.
Last time I checked, TeX was Turing complete. That leads to all sorts of problems when you try to extract data from TeX documents.
HTML has its problems, but TeX and LaTeX are exactly the wrong direction to change things. Since you've only "had to write HTML once" I guess that explains why you're looking at it in the wrong way.
BTW, HTML is not a text formatting language. It's a markup language, which is quite a different thing.
Agreed, the killing of female children is horrible. But the problem isn't the one child policy. The problem is with those people ignorant enough to kill their female child for want of a male, and with the Chinese government for letting them get away with it.
If someone has a female child and then kills her, not only should they be held guilty of murder, but that should also count as "using up their quota". The problem isn't with the one child system, the problem is that people are able to get away with killing baby girls.
Also, imagine what those people would do without the one child policy. They'd keep having kids until they get a boy. Then maybe they'd want a second boy. So you end up with a country of several billion, all starving.
In the urban parts of China people tend to care about their only child a lot, regardless of sex. The old belief that a girl wasn't really part of the family, but part of her future husband's family, is going away.
What is it that makes the US constitution sacred? I have nothing wrong with people being allowed to carry guns, but perhaps some intelligence tests should be required. The problem right now is that 99% of the people with guns are idiots, and 99% of intelligent people don't feel the need to carry a gun.
I agree that breeding should be regulated as well. Of course, most Americans won't agree with that. Heck, most Americans think that it's evil that China has a one child policy, when they clearly *need* it. Without its one child policy, China's population would quickly reach the point where there'd be mass starvation, yet I always hear Americans saying things like "That's so terrible! I couldn't imagine not growing up with my six brothers and sisters!"
And switch to metric, dammit! What's wrong with you people?
You seem to be making the assumption that being a hacker is mutually exclusive with being a software engineer. You're right, each approach has its strengths and weaknesses.
If I need something done quick, and only once, I'll make a quick hack. If I find that I'm using the same hack frequently, I'll switch hats, and redesign/rewrite the thing to be more generalized and robust.
Conservative garbage collectors are inherently broken. Some JVMs actually use conservative garbage collectors, and there are serious problems with them. They're particularly bad for applications that deal with very large amounts of almost random data. The collector believes that almost everything that was allocated has a reference, and doesn't free anything.
And what's this about Java programs needing to copy things unnecessarily? A "well-crafted" Java program doesn't need to do that either.
Finally, while Java programs always use the "new" syntax to allocate objects, it's really up to your virtual machine to decide where objects get allocated. A JVM could, conceivably, allocate short-lived object on the stack. And if you have a good heap allocator, that doesn't matter too much anyways. (HotSpot apparently has one that's significantly faster than any known C++ heap allocator, and some of its speed comes from GC-related optimizations)
Since when has telnet been built into Netscape? Never. Yes, it supports telnet: URL's, but it does that using a separate application. This would also make sense for mail, news, irc, and many other protocols.
Like I said before, if this chat thing requires *no changes* to the Mozilla base code at all, then it's fine with me. But as soon as Mozilla needs even a special set of hooks for "chat plugins" then they've gone too far. Leave it out.
I think it would be nice if I could click on a link to "netrek://mit.netrek.org" and have it bring up Netrek. I don't think they should add Netrek, or even "networked space game" support to Mozilla though.
Mozilla should have support for adding new protocol identifiers in an easy way, and having several ways of handling them: seperate application, special type of plug-in, etc. Much like the existing mime-type stuff Netscape has had all along. This should be totally generic though. Explicit support for particular things like chatting or text editting or whatever just leads to bloat and bad design. A good metric would be this: if chatting is mentioned anywhere in the Mozilla source, then they're doing it wrong.
JavaScript was originally called "liveScript". "liveConnect" is something else... I think it's the think that lets Java and JavaScript (and possibly some other things) talk to eachother. One use for it is so Java applets can access cookies. (JavaScript can access cookies, but Java applets cannot. But if a Java applet can communicate with JavaScript on a page...)
Even if it'sgoing to be a "modular" piece, it still doesn't make sense. Maybe they should also make Netrek, wordprocessor, CAD, and StarCraft modules? After all, if you don't want 'em, you don't need to download 'em, right?
It's great that Mozilla is finally becoming modular. But it should be somewhat restrictive about what kinds of modules it supports. A unified interface makes sense for similar tasks. Chatting is in no way similar to web browsing though, so I can't see a good reason for Mozilla to have any support for chatting whatsoever.
This chat module thing: is it going to require *any* code changes to Mozilla itself, or are they just going to use interfaces Mozilla already had? If it's the latter, then I don't see an real harm, but then it may as well be a separate project. If it's the former however, they're just going to delay the release of Mozilla, and overcomplicate it for a feature that shouldn't be there in the first place.
Mozilla should be made as simple as possible. Essentially just a container with a few programming interfaces for hooking up things like protocols and rendering engines. An interface for other applications to be able to communicate with Mozilla would be good too. Much more than that is pushing it.
It doesn't matter how much time went into it? So even if it takes hundreds of people to develop some very complex software for a relatively small market, $500 is too much? Then why would they ever develop the software? Most free software was written to "scratch the developer's own itch". How many developers are just itching for some inventory control software, or point of sale software?
That isn't a reasonable scenario for developers because of this:
Lets suppose a group of developers write some application, make it GPL, and try to sell it for $20. Somone else gets the source and then sells their own "distribution" for $15. They can afford to sell it cheaper because they have only distribution costs, while the developers also have development costs. What they (the distributors) are doing is also perfectly legal, since the code is GPL, remember.
Grandma then has three choices:
1. get it free, but build it herself 2. get it from the developers for $20 3. get it from the distributor for $15
Which choice will grandma make? Probably #3. Especially if the distributor is a relatively well known name like RedHat, while the developers are relatively unknown. So the developers get little or nothing.
You might argue that "the developers could provide better support", but really, how many developers do you know that also work in tech support? No good developers I'm sure.
The people who make money from free software are in tech support and distribution. Developers get nothing, even though they're the ones producing the actual product.
Slashdot Mirror
That's what I'd like to know: what can you do in Perl that you can't do in Python? My guess is nothing. As for what you can do in Python that you can't do in Perl: Python code is actually maintainable. Perl code generally isn't. And Perl makes it hard as hell to make code that even comes close to readable.
What is it that turns Linux users into crusaders against Windows?
The fact that Perl requires goofy symbols before variables depending on their type pretty well precludes the possibility of writing readable Perl code. That reminds me of that old Perl joke. "I just wrote this Perl script that does all sorts of cool stuff in only 8 characters! Actully, I could've written it in only 6, but this is more readable."
I have yet to see readable Perl code. Today, on those rare occasions when I have to write Perl, I often write it in Python first, and then convert it to Perl. The Perl code looks like the Python code would after being sent over a very noisy serial line with no error correction...
Yes, Python isn't as popular as Perl, but popularity is for Lemmings. I'll take technical superiority any day. Python gets the job done easier than Perl, and if/when bugs need to be fixed or features need to be added, I don't need to pull may hair out trying to understand an unreadable script.
Huh? So how do you read C code that's printed on paper? Do you count the braces, or pay attention to the indentation? I doubt it's the former, and the latter is the same as Python.
I've used both Perl and Python quite a bit. I have to say that Python code tends to be somewhat easier to write, and significantly easier to read.
I at first thought Python's "indentation == block structure" was pretty gross, but it actually makes sense. People pay attention to indentation, so it only makes sense that the computer should too.
But really, anyone who bases their conviction that Perl is better than Python because they think Python's indentation thing is "yucky" obviously hasn't taken a real look at Python, and deserves to be ignored.
I used to use Perl. Then I went and learned Python despite my reservations about it's indentation scheme. Since then, I've converted to Python. Anything Perl is good at, Python is better at. It's easier to write, and easier to read. Even before I learned Python, I found it to be easier to read than Perl. I normally don't "convert" to a language. I know more languages than I have fingers, and quite a few of them I use on a regular basis.
The only "advantage" Perl has is that it's installed on more systems than Python. That's the only reason I use Perl at all actually. When I have to write a script that will run on a machine that has Perl but not Python installed. That is (thankfully) becoming more rare.
My point is that while having C vs Java arguments (for example) are futile -- each languange has its valid uses -- Perl is essentially holding us hostage. There's no reason to use Perl, because Python is strictly better. No reason that is, except for the large installed base of Perl. I won't even mention the obvious parallel to Windows and Linux. D'oh! Looks like I just did.
Someone asked for some Python sample code. I was going to post some, but /. seems to have a bug where even stuff posted as "HTML Formatted" undergoes "entity translation". Of course, that doesn't have anything to do with the Perl code in /. being hard to maintain... right?
That's funny. I seem to recall that the IIe was written "][e", but the IIc was written "//c". I think the earlier ones (plain II and II+) were actually written with I's, but I can't be sure.
_Fountains_of_Paradise by Arthur C. Clarke also had microfilaments. In that book they were very think carbon chains or tubes I believe. Today we'd call the tube form Bucky-Tubes. The ends could turn into some other structure that wouldn't cut through things. You could then have one end fastened to a handle, and then in the other end have a chunk of superconductor. A strong electromagnet in the handle would keep the bucky-tube tight. (the superconductor would be repelled by the electromagnet) Not a light-saber, but a pretty neat high-tech "blade" weapon. The blade would probably be close to invisible too...
Come to think of it, such an idea might've been proposed in a Larry Niven story, though I can't remember what, or even if...
The previous story was about Netscape revamping their search engine to use Google technology. Why does AOL also need a new "Inktomi powered" search engine? They own Netscape...
I think you're quoting out of context. I believe he meant "the first Unix implementation [of TCP/IP] was done in berkeley". He was wrong in thinking that that wasn't the first TCP/IP implementation though, AFAIK.
It's here.
Fine then, How about counterfeiting money. Is that wrong? I mean, if you go print off $100,000 nobody else is losing any money, are they? You're just gaining.
Now think about what would happen if that was socially acceptable and/or legal. How much would money be worth?
In the same way, if software piracy was legal and socially-acceptable, more people would do it, hence more profit would be lost.
Yes, the "lost profits" reported by software companies do tend to be inflated. It is true that only some of those who pirate would buy the software if they couldn't pirate it. But some of them would.
Trying to justify piracy by saying "it doesn't hurt anyone" will only increase the amount of piracy that goes on, thereby causing many people that currently buy software to instead pirate it.
Software piracy *does* hurt the publishers. Sure, not by as much as (pirated copies)*(retail price), but probably a significant fraction of that. Just because they don't lose *all* of that profit doesn't mean it isn't wrong.
If you can't afford it, tough. If you think the asking price isn't worth it, go elsewhere, or write it yourself. If you think it's okay to pirate because software is "so low quality", then:
1. why are you bothering to pirate it?
2. why don't you write something better?
To further strengthen that argument, that the browser is the unctionality, not the code, look at Windows NT 4.0. NT4 Workstation and Server consist of the same code. The only thing that's different is the functionality, which is set by some registry entries. Remember to O'Reilly hack to turn NT Workstation into NT Server?
At least, IDE's that don't let you integrate whatever editor you want suck. I'd rather use the command line if the IDE won't let me edit with VIM.
And I agree that these things should really be written in Java. Do the C/C++ programmers writing these IDE's know what a Java developer wants in their environment. A lot of the needs are the same, but some are quite different. KLGroup's JProbe Memory debugger is an example of a tool that makes sens for Java, but not something like C or C++. (It lets you find out what objects aren't being collected because they've still got references. C/C++ memory leak tools are quite different.)
So why can't RMS be like "the great artists and artisans of medieval times"? The name GNU implies RMS. He wants it be called GNU/Linux for his own personal gratification. It's as simple as that.
I think the writers of free software do deserve credit. But according to RMS, people don't deserve, nor should they desire, credit for their work. According to him, that's just an excuse for proprietary non-free software. Yet when it comes to his own foundation's name, he demands that people give the FSF, and hence RMS, credit. That seems awfully hypocritical to me.
With the BSD license, some big wealthy corporation could hijack some code I spent hours and hours and sweat and
blood writing, make some utterly trivial change to it, then distribute it under a proprietary license and make a huge
profit and I never see a penny. In addition, if they made VALUABLE changes, they would be closed, not keeping with
the "open source" spirit under which the original code was written. Licensing under GPL prevents this.
The GPL only partially protects against this. It does not protect against cases where:
- you write an amazing program
- some big corporations takes it, and makea a few insignificant tweaks
- big corporation sells "distributions" of your code, while you get nothing
This, to me, seems to be the big problem with both BSD and GPL. They both allow people to gain equal profit regardless of how little work they've done. Because we live in a capitalist society, the end result is that those who do more work get less.
That's why open source developers have to get real jobs, and can only work on free projects in our spare time. That "Commercial Open Source" idea is looking better all the time.
Not if it was Canadian Mountain Dew.
Q: Wouldn't the web be much better if it was all LaTeX based?
A: No, it would suck even worse.
Last time I checked, TeX was Turing complete. That leads to all sorts of problems when you try to extract data from TeX documents.
HTML has its problems, but TeX and LaTeX are exactly the wrong direction to change things. Since you've only "had to write HTML once" I guess that explains why you're looking at it in the wrong way.
BTW, HTML is not a text formatting language. It's a markup language, which is quite a different thing.
Agreed, the killing of female children is horrible. But the problem isn't the one child policy. The problem is with those people ignorant enough to kill their female child for want of a male, and with the Chinese government for letting them get away with it.
If someone has a female child and then kills her, not only should they be held guilty of murder, but that should also count as "using up their quota". The problem isn't with the one child system, the problem is that people are able to get away with killing baby girls.
Also, imagine what those people would do without the one child policy. They'd keep having kids until they get a boy. Then maybe they'd want a second boy. So you end up with a country of several billion, all starving.
In the urban parts of China people tend to care about their only child a lot, regardless of sex. The old belief that a girl wasn't really part of the family, but part of her future husband's family, is going away.
Unforutnately, we can't change em all. I guess it's going to end up like the metric system.
You mean like the way every civilized nation uses metric?
What is it that makes the US constitution sacred? I have nothing wrong with people being allowed to carry guns, but perhaps some intelligence tests should be required. The problem right now is that 99% of the people with guns are idiots, and 99% of intelligent people don't feel the need to carry a gun.
I agree that breeding should be regulated as well. Of course, most Americans won't agree with that. Heck, most Americans think that it's evil that China has a one child policy, when they clearly *need* it. Without its one child policy, China's population would quickly reach the point where there'd be mass starvation, yet I always hear Americans saying things like "That's so terrible! I couldn't imagine not growing up with my six brothers and sisters!"
And switch to metric, dammit! What's wrong with you people?
You seem to be making the assumption that being a hacker is mutually exclusive with being a software engineer. You're right, each approach has its strengths and weaknesses.
If I need something done quick, and only once, I'll make a quick hack. If I find that I'm using the same hack frequently, I'll switch hats, and redesign/rewrite the thing to be more generalized and robust.
Conservative garbage collectors are inherently broken. Some JVMs actually use conservative garbage collectors, and there are serious problems with them. They're particularly bad for applications that deal with very large amounts of almost random data. The collector believes that almost everything that was allocated has a reference, and doesn't free anything.
And what's this about Java programs needing to copy things unnecessarily? A "well-crafted" Java program doesn't need to do that either.
Finally, while Java programs always use the "new" syntax to allocate objects, it's really up to your virtual machine to decide where objects get allocated. A JVM could, conceivably, allocate short-lived object on the stack. And if you have a good heap allocator, that doesn't matter too much anyways. (HotSpot apparently has one that's significantly faster than any known C++ heap allocator, and some of its speed comes from GC-related optimizations)
Since when has telnet been built into Netscape? Never. Yes, it supports telnet: URL's, but it does that using a separate application. This would also make sense for mail, news, irc, and many other protocols.
Like I said before, if this chat thing requires *no changes* to the Mozilla base code at all, then it's fine with me. But as soon as Mozilla needs even a special set of hooks for "chat plugins" then they've gone too far. Leave it out.
I think it would be nice if I could click on a link to "netrek://mit.netrek.org" and have it bring up Netrek. I don't think they should add Netrek, or even "networked space game" support to Mozilla though.
Mozilla should have support for adding new protocol identifiers in an easy way, and having several ways of handling them: seperate application, special type of plug-in, etc. Much like the existing mime-type stuff Netscape has had all along. This should be totally generic though. Explicit support for particular things like chatting or text editting or whatever just leads to bloat and bad design. A good metric would be this: if chatting is mentioned anywhere in the Mozilla source, then they're doing it wrong.
JavaScript was originally called "liveScript". "liveConnect" is something else... I think it's the think that lets Java and JavaScript (and possibly some other things) talk to eachother. One use for it is so Java applets can access cookies. (JavaScript can access cookies, but Java applets cannot. But if a Java applet can communicate with JavaScript on a page...)
Even if it'sgoing to be a "modular" piece, it still doesn't make sense. Maybe they should also make Netrek, wordprocessor, CAD, and StarCraft modules? After all, if you don't want 'em, you don't need to download 'em, right?
It's great that Mozilla is finally becoming modular. But it should be somewhat restrictive about what kinds of modules it supports. A unified interface makes sense for similar tasks. Chatting is in no way similar to web browsing though, so I can't see a good reason for Mozilla to have any support for chatting whatsoever.
This chat module thing: is it going to require *any* code changes to Mozilla itself, or are they just going to use interfaces Mozilla already had? If it's the latter, then I don't see an real harm, but then it may as well be a separate project. If it's the former however, they're just going to delay the release of Mozilla, and overcomplicate it for a feature that shouldn't be there in the first place.
Mozilla should be made as simple as possible. Essentially just a container with a few programming interfaces for hooking up things like protocols and rendering engines. An interface for other applications to be able to communicate with Mozilla would be good too. Much more than that is pushing it.
It doesn't matter how much time went into it? So even if it takes hundreds of people to develop some very complex software for a relatively small market, $500 is too much? Then why would they ever develop the software? Most free software was written to "scratch the developer's own itch". How many developers are just itching for some inventory control software, or point of sale software?
That isn't a reasonable scenario for developers because of this:
Lets suppose a group of developers write some application, make it GPL, and try to sell it for $20. Somone else gets the source and then sells their own "distribution" for $15. They can afford to sell it cheaper because they have only distribution costs, while the developers also have development costs. What they (the distributors) are doing is also perfectly legal, since the code is GPL, remember.
Grandma then has three choices:
1. get it free, but build it herself
2. get it from the developers for $20
3. get it from the distributor for $15
Which choice will grandma make? Probably #3. Especially if the distributor is a relatively well known name like RedHat, while the developers are relatively unknown. So the developers get little or nothing.
You might argue that "the developers could provide better support", but really, how many developers do you know that also work in tech support? No good developers I'm sure.
The people who make money from free software are in tech support and distribution. Developers get nothing, even though they're the ones producing the actual product.