"They can already do this through the JCP. Individual membership in the JCP is free"
Yes and IF accepted (unlikely) then your contribution will sit on a shelf for a MINIMUM of 6-12months before being accepted.
"Again I think that's the point I'm making. Sun isn't trying to kill competition. Other companies might if Java was open sourced. That doesn't just hurt sun, that hurts Java and the developer community."
How exactly is it their going to accomplish this I wonder? They can implement a non-standard vm already, hell they can even license sun's from sun and do it. With an open source license like the gpl they would have to gpl their changes which precludes doing anything like what you suggest.
How could it possibly kill java to open source it? I imagine sun would remain the maintainer, so it's not exactly like anything is going to go in they wouldn't approve of.
The whole "millions of forks" argument is pointless. Anyone can look at the specs and implement a Java VM NOW and many DO. In fact there are open source Java VM implementations ALREADY so anyone too lazy to code one from scratch can do it now.
What we have NOW is millions of forks, if java opens and sun actually accepts contributions as well (if they don't, it WILL fork and eventually settle on another major VM, read XFree86) then alot of those will disappear as additional features they offer become part of the new open source Sun java.
Open sourcing java alone is not enough though, two other things need to happen.
1. Sun must actually actively maintain and accept contributions. Or appoint a third party from the community to do so.
2. The spec must become standardized and be taken out of sun's hands so that the community can actually submit extensions, etc.
Basically Java must become open. As far as the VM, yes it's a matter of changing the license.txt file and putting the source out there. Fragmentation won't occur if they don't give it a reason to, if sun is slow or unwilling to accept contributions THEN it will kill java.
Despite what some people are saying, the reason for opening it up (as well as standardizing it) is so that improvements can actually get into real sun vm.
If we can pull that off then there won't be 200,000 different vm's anymore that splinter off from the standard. Forks is what have WITHOUT open source, unification is what we'll have WITH open source... assuming sun actually accepts contributions.
"I do think the entire OS should be called GNU/Linux, instead of just Linux, as most of the packages are GNU stuff. Bash, GCC, GDB, GNU automake..."
This is offtopic and I should be modded into the ground but everytime I see this I just can't let it go.
First point: The ENTIRE OS is the kernel, everything else merely interfaces with the OS and is thus either a library or an application. System utils like those you mentioned above are still just apps. Since the Kernel is the OS, and it's author(s) choose to call it Linux, it is properly refered to as the Linux Operating system or a linux operating system (as the case may be).
Individual applications may be called by their respective names. And distributions again should be called by the name whoever compiles the distro chooses to call it. There's nothing that says Fedora linux, couldn't just be called Fedora for instance (they name the full distribution after the OS which make sense but isn't required).
The GNU applications are often packaged with linux, but there is no particular reason for them to be. The BSD tools for instance would work just fine. And further, most of the GNU utils are pretty trivial stuff to say the least, would you like to see me write half of them using a grand total of less than 100 lines of perl code? They are by no means ALL trivial and some are good stuff, but there's nothing there that couldn't be replaced either...
Most distributions include 100's or even thousands of programs coupled with the Linux Operating system, compared to the total the GNU stuff really isn't all that much I don't know where people get this idea.
In short, there is no actual credible reason to call the Linux OS GNU/Linux. Simply because he's depressed that hurd is pretty thoroughly lame and useless isn't a good enough reason to call it by something other than the name the author and maintainer has chosen for it.
"But Perl was released for the longest time under the Artistic licence, which (IIRC) allows derivitive works, but doesn't allow you to call them Perl. This could keep the one, true source of Java unsullied by broken or incompatible implementations yet gives everybody else the hope that when Sun tanks Java won't."
The gpl will do this as well, sun java remains sun java so long as sun continues to maintain it and passes it off to another to maintain afterward.
Anyone can get the specs and write their own implementation of java now. Java being gpl'd doesn't change this, it simply means that people can write improvements and possibly get them into the main official java instead of having to write a custom vm to get the simplest change implemented.
Under the gpl we won't see a reappearance of the Microsoft Java VM either... or if we do it will have to be gpl'd.
A few things that can be done to improve on java. Make the standard a standard housed with an officially recognized standards body which does not allow patented features to be accepted into standards.
Right now java is in sun's hands and nobody has a say in the standard but sun.
And yes, open source sun's implementation, it's the only one anybody codes for! It does little good to split off a non-sun java vm. If they gpl it we don't have to worry about Microsoft doing this again either.
And yes sun java is still slow as hell, whatever you say, you can't change that. With things opened up the interpreter will be drastically improved... it's not like sun will stop working on it, and it's not like sun alone could ever compare to what the community can do with it. Just look at Open Office 1.1 if you want an example.
Java being in demand is all the more reason to open it up, why do you think people are clammoring for this? Sun needs to learned that closed standards and closed sourcecode are not where the money is anymore and start gpling all their software (it's not like software is sun's money maker anyway). If they want to be friends with this community they have to contribute to it and work in an open to all fashion (not merely visible to all).
Piracy generally isn't that big a problem (although it can hurt small commercial projects if someone is really low and won't pay the $5 or $10), for the most part it's actually advertising. It doesn't cost the software house a dime. It's not like stealing a box off the shelf, in that there is something gone which cost money to produce.
In the case of software if you borrow my cd and copy it, guess what, my cd isn't gone, I haven't lost anything. Most of those who pirate a piece of software would never have bought the program or WILL buy the program after trying out a fully functional non-time limited demo. It's no different than music, where cd sales have dropped by the bottom line is actually ahead since filesharing has begun.
Now for the second side of the coin, your tools for rapid development. People are always looking for ways to develope software faster... lots of people at given second of every given day and improvements are being made.
But the truth is the kind of tools you speak of already exist, most are built right into programming languages their called loops, functions, objects and variables. Without those development would take 100 fold as long. There are plenty of other tools as well which speed up coding time, and those have reduced things 10 fold from the reduction you get with these constructs in the language.
While new tools are being made that work, and work well there is no magic tool which makes it that much faster. The reason is simple, speeding it up drastically further than what we have now requires simplifying programming, and simplifying programming means giving up flexibility, which in turn means not being able write any non-trivial program.
It also means a lack of efficiency. If I hand write code I can choose to do it in the most efficient manner possible, in alot of cases this is critical because otherwise the program will be dog slow. If I use cookie cutter functions to make things easier I end up with lots of generic redundant code which is aimed to be as generic and reusable as possible so it's the need of everyone who might use that cookie cutter. It's not tailored for MY use.
A good example of this is VB, VB is about as dumbed down as you can get (in the typically unintuitive microsoft fashion) and with VB you can get a graphical application up in a matter of minutes. However even the most trivial vb application would be faster if it was written in C or C++, it would use less space on the hard drive, use less memory and use less processor. And by most trivial I mean something that simply pops up a diaglog which says "Hello, world!" with an ok button.
"While I agree that the brains behind OSS are largely individuals, the developers themselves, Linux simply would not be what it is today if it weren't for companies pumping money into Linux."
I'm not sure how you arrive at this conclusion but I have a suspicion. When companies (such as IBM for instance) started putting money into linux it was when the Platform as a whole was starting to mature.
It's been since then that alot of the big visible improvements have occured... but 99% of them didn't come from the corporations! They came from the same place the rest did, they came from the community. Most of the corporate money that has gone into linux has gone toward kernel features that are useful to a very select group of people (uber number of a processors, numa, etc). Sun contributed open office and that was significant, and redhat has contributed quite a bit... other than that I know of very little corporate money churning new features into open source.
Now by no means am I saying it's not a nice thing to see a company contribute, particularly by hiring an open source developer and paying him to work full time on a project. I'm just saying that presently and over the past 2 or 3yrs when all the great developments have happened that IS NOT what has been the driving force. It's simply been a matter of alot of things ripening in a given timeframe and more people (thus more open source developers) using the platform.
"If we had a system in which people didn't have to work, then the corporate contributions would be more or less irrelevant because people would have plenty of time to devote to free software development."
Now that sir I absolutely agree with. Good points though. It's always a breath of fresh air to debate with someone who's not a religious zealot of some sort or other;)
An OS X interface would just be a perk for OS X users. Gimp only runs on OS X become some OS X users would like it to.
It's main userbase runs it on linux which IMHO is what everyone should be running on their desktop (whether it be mac, pc, handheld, or IBM Bluegene).
The interface already integrates quite nicely in X11 and windows for that matter due to people interested in having it there. If you OS X users want it to integrate well with MacOS X I'd suggest you start coding, I'm sure gimp will accept your code if well written.
"cable "premium" channels which cost about 30 pounds per month"
This is what I was refering to, if you mean the BBC I believe that work out to more like $210/year but I was talking about the other premium channels in comparison to ours.
"Plus, I don't know where you're getting "150Mb+" from; my Windows Task Manager is right now reporting Photoshop CS as having a memory footprint of 34,076k."
It used to have a Hard disk footprint of 24mb. I was talking about size on disk not size in memory. At the time Photoshop 5 came out disk space was till an issue (actually it still is IMHO, but many would disagree... I tend to think plentiful resources aren't a valid excuse for bloat).
"No, which is why you can still do this - only now the type actually appears in your composition, rather than in a separate requester, which I think is a major improvement."
Actually I believe in 5 it gave a sample set of characters in the font. In later versions I know it shows it on the image, which I personally don't find an improvement over 4 but prefer to the behavior of 5.
"an OSS programmer working from free for no profit isn't likely to produce anything for you unless they don't have bills to pay and can afford to write softwre for you that will make you money but they themselves don't care about being paid. Very unlikely."
Odd, I was under the impression these are exactly the circumstances which MOST open source programmers write code (aside from the no bills thing). A few get paid, most don't, most write software in their free time. Open source development is typically cheaper if the program is useful to others anyway, and if you'll sponsor a project to get what you want it's the cheapest route to go.
"If you want good quality software with least defects that suits your customer's needs (implying customization)"
Good quality software with the least defects means using an open source codebase, if customization is needed it is done to that codebase. No private firm is going to debug nearly so well as 10,000 geeks.
Custom software is something that should be avoided unless there is absolutely no other option and situations which require it are rare.
Since there is likely an open source project at least similar to what is needed, why on earth would we start from scratch when we could adapt it and best yet, get the changes maintained for free if it's something that can be contributed back!
Right about now, your scratching your head and trying to figure out which role I'm coming from, management, developer, etc.
The truth is I'm a network technician/administrator/programmer, the company I work administrates the networks of hundreds of small buisnesses. If the company I work for, or one our customers needs software, then I may write it, may buy it, or may have it written depending on the case. If I write the software the customer (or my company) will be licensed the software under the gpl. Or I may modify an open source program to suit (probably contributing back changes if relevant).
Or we may hire an outside firm to adapt an open source program and license to us under a BSD style license with source... so we can release to the customer under the gpl.
I'd go further back than that, I'd say 4.1. Ok 5 did add some useful tools (something they haven't done since) but it added shitloads of bloat. Photoshop went from 20+ mb to 150+ mb in a single version and screwed up simple things, getting rid of a critical tool on the toolbar for one.
They screwed up type preview (is the ability to arrow down through the fonts and have them update on the type you've put in the box and are actually putting in the graphic really something you want to take away?
So yes they added some good functionality, but nothing critical, and they took away some I consider critical for productivity... especially in web graphics where half of them need text.
"both companies which make money on Linux, and companies which make the software"
Neither barney the purple webdesign website nor slashdot are the right place to reach either of these groups. (sorry couldn't resist the Barney thing, that website looks like it should be seeing feminine hygene products not giving advice on web design)
"The latter group, including companies like Adobe, are interested in what Linux users want for two reasons."
A little offtopic wouldn't you think, since the parent wasn't talking about linux users, he was talking about non-linux users saying they'd use linux if it weren't for this or that.
Also because this article was written by a MacOSX user, NOT a linux user. 99% of the things he complained about were really one thing, it didn't have an OS X interface, it's designed for X11.
"So, not only is it a waste of your time to complain about people complaining about the one last thing missing from Linux, but your will actually runs counter to the companies pumping money into Linux, or put another way, the hand that's feeding you."
That hands that feed the linux community are open source developers. While some of those are being paid by corporations, for the most part they are not. Corporations having money only means so much in this crowd, after all while there are exceptions, they've managed to produce mostly crap. Look at Microsoft for instance, they've managed to produce ONLY crap and they've got more money than any of the others.
"Photoshop is clearly one of the killer apps of all time. It was even ported to one or two kinds of Unix (likewise, FrameMaker) due to the fact that there was no (and still is no) superior product, and at the time Unix workstations were more powerful than PCs. Those days are gone, and so there is little to no motivation to develop new versions of Photoshop for Unix systems. On the other hand, Linux market share is growing, and people are starting to demonstrate a willingness to pay for software which runs on Linux. Not just corporations mind you, but also individuals and small businesses."
A piece of killer app proprietary software is great for linux adoption which ultimately ends up being good for the community, since it helps yield more open source developers to work on things like gimp so it eventually can get that proprietary app out of our otherwise clean open system. A better solution of course would be for Adobe to open the photoshop sourcecode.
Since I don't see that happening, opening the Acrobat reader source would be nice. The linux version of Acrobat reader is pretty shoddy to say the least, it's ugly, unpolished, and bugridden. If Acrobat reader were opened then we could either fix it (if it was worth fixing) or get a look at it's pdf handling, since actually opening and rendering the pdf is the only thing it does in a superior way to the other dozen pdf reading linux apps. AR is faster and has better print options. I think it could be salvaged myself, open source would quickly eradicate the bugs and fix the UI.
But did I miss something, or is this nothing but a Mac user whining that an X11 program doesn't have a native OS X interface and doesn't work just like photoshop.
That aside, are we really supposed to be taking advice from a guy who puts up a website which indicates it gives website design advice and the first thing you see is a banner add and a giant purple page...
Worth noting that even on win32 there is not only a setup.exe that works perfectly but the interface fits quite nicely (it's not like windows users REALLY expect standardized interfaces anyway).
I was just wondering why you skipped the first response which asked about using an Ogg container format asking how proprietary the container format you are using is.
"There's nothing about Dirac, BTW, that requires C++, or even object-oriented programming."
I'd tend to think that's a given since there is nothing that requires C++ or object-oriented programming.
Not to say there aren't some things it's faster and easier to implement with OO, there's nothing that requires it. And there's certainly nothing requiring it be C++ even then;)
"On the rare occasions when I can't figure out what some chunk of code does, I'll just comment it out and run the tests. If nothing breaks, then I delete it and move on. And if something does break, then I delete it and write something clearer in its place. It's very freeing!"
This statement says nothing about them not writting a test, it says if you see code you don't understand. And specifically says if something does break when you run the tests you rewrite the code it's clearer. And for that I refer you back to the post you replied to. It may have been something that was rewritten to correct a problem you didn't understand to begin with, and that's why it wasn't clear to you.
In the case of a rewrite that fixes a security hole, it's possible but unlikely the test would be an exploit, more likely the test would be the same as for the original code making sure it functions correctly. Looking at the test would only confirm to yourself that you were "cleaning up" the code while you were actually putting a security hole right back in.
Slashdot Mirror
Except changing the specification is not open, it rests with sun instead of a standards body.
"They can already do this through the JCP. Individual membership in the JCP is free"
Yes and IF accepted (unlikely) then your contribution will sit on a shelf for a MINIMUM of 6-12months before being accepted.
"Again I think that's the point I'm making. Sun isn't trying to kill competition. Other companies might if Java was open sourced. That doesn't just hurt sun, that hurts Java and the developer community."
How exactly is it their going to accomplish this I wonder? They can implement a non-standard vm already, hell they can even license sun's from sun and do it. With an open source license like the gpl they would have to gpl their changes which precludes doing anything like what you suggest.
How could it possibly kill java to open source it? I imagine sun would remain the maintainer, so it's not exactly like anything is going to go in they wouldn't approve of.
The whole "millions of forks" argument is pointless. Anyone can look at the specs and implement a Java VM NOW and many DO. In fact there are open source Java VM implementations ALREADY so anyone too lazy to code one from scratch can do it now.
What we have NOW is millions of forks, if java opens and sun actually accepts contributions as well (if they don't, it WILL fork and eventually settle on another major VM, read XFree86) then alot of those will disappear as additional features they offer become part of the new open source Sun java.
Open sourcing java alone is not enough though, two other things need to happen.
1. Sun must actually actively maintain and accept contributions. Or appoint a third party from the community to do so.
2. The spec must become standardized and be taken out of sun's hands so that the community can actually submit extensions, etc.
Basically Java must become open. As far as the VM, yes it's a matter of changing the license.txt file and putting the source out there. Fragmentation won't occur if they don't give it a reason to, if sun is slow or unwilling to accept contributions THEN it will kill java.
Despite what some people are saying, the reason for opening it up (as well as standardizing it) is so that improvements can actually get into real sun vm.
If we can pull that off then there won't be 200,000 different vm's anymore that splinter off from the standard. Forks is what have WITHOUT open source, unification is what we'll have WITH open source... assuming sun actually accepts contributions.
"I do think the entire OS should be called GNU/Linux, instead of just Linux, as most of the packages are GNU stuff. Bash, GCC, GDB, GNU automake..."
This is offtopic and I should be modded into the ground but everytime I see this I just can't let it go.
First point: The ENTIRE OS is the kernel, everything else merely interfaces with the OS and is thus either a library or an application. System utils like those you mentioned above are still just apps. Since the Kernel is the OS, and it's author(s) choose to call it Linux, it is properly refered to as the Linux Operating system or a linux operating system (as the case may be).
Individual applications may be called by their respective names. And distributions again should be called by the name whoever compiles the distro chooses to call it. There's nothing that says Fedora linux, couldn't just be called Fedora for instance (they name the full distribution after the OS which make sense but isn't required).
The GNU applications are often packaged with linux, but there is no particular reason for them to be. The BSD tools for instance would work just fine. And further, most of the GNU utils are pretty trivial stuff to say the least, would you like to see me write half of them using a grand total of less than 100 lines of perl code? They are by no means ALL trivial and some are good stuff, but there's nothing there that couldn't be replaced either...
Most distributions include 100's or even thousands of programs coupled with the Linux Operating system, compared to the total the GNU stuff really isn't all that much I don't know where people get this idea.
In short, there is no actual credible reason to call the Linux OS GNU/Linux. Simply because he's depressed that hurd is pretty thoroughly lame and useless isn't a good enough reason to call it by something other than the name the author and maintainer has chosen for it.
"But Perl was released for the longest time under the Artistic licence, which (IIRC) allows derivitive works, but doesn't allow you to call them Perl. This could keep the one, true source of Java unsullied by broken or incompatible implementations yet gives everybody else the hope that when Sun tanks Java won't."
The gpl will do this as well, sun java remains sun java so long as sun continues to maintain it and passes it off to another to maintain afterward.
Anyone can get the specs and write their own implementation of java now. Java being gpl'd doesn't change this, it simply means that people can write improvements and possibly get them into the main official java instead of having to write a custom vm to get the simplest change implemented.
Under the gpl we won't see a reappearance of the Microsoft Java VM either... or if we do it will have to be gpl'd.
A few things that can be done to improve on java. Make the standard a standard housed with an officially recognized standards body which does not allow patented features to be accepted into standards.
Right now java is in sun's hands and nobody has a say in the standard but sun.
And yes, open source sun's implementation, it's the only one anybody codes for! It does little good to split off a non-sun java vm. If they gpl it we don't have to worry about Microsoft doing this again either.
And yes sun java is still slow as hell, whatever you say, you can't change that. With things opened up the interpreter will be drastically improved... it's not like sun will stop working on it, and it's not like sun alone could ever compare to what the community can do with it. Just look at Open Office 1.1 if you want an example.
Java being in demand is all the more reason to open it up, why do you think people are clammoring for this? Sun needs to learned that closed standards and closed sourcecode are not where the money is anymore and start gpling all their software (it's not like software is sun's money maker anyway). If they want to be friends with this community they have to contribute to it and work in an open to all fashion (not merely visible to all).
A couple problems with this:
Piracy generally isn't that big a problem (although it can hurt small commercial projects if someone is really low and won't pay the $5 or $10), for the most part it's actually advertising. It doesn't cost the software house a dime. It's not like stealing a box off the shelf, in that there is something gone which cost money to produce.
In the case of software if you borrow my cd and copy it, guess what, my cd isn't gone, I haven't lost anything. Most of those who pirate a piece of software would never have bought the program or WILL buy the program after trying out a fully functional non-time limited demo. It's no different than music, where cd sales have dropped by the bottom line is actually ahead since filesharing has begun.
Now for the second side of the coin, your tools for rapid development. People are always looking for ways to develope software faster... lots of people at given second of every given day and improvements are being made.
But the truth is the kind of tools you speak of already exist, most are built right into programming languages their called loops, functions, objects and variables. Without those development would take 100 fold as long. There are plenty of other tools as well which speed up coding time, and those have reduced things 10 fold from the reduction you get with these constructs in the language.
While new tools are being made that work, and work well there is no magic tool which makes it that much faster. The reason is simple, speeding it up drastically further than what we have now requires simplifying programming, and simplifying programming means giving up flexibility, which in turn means not being able write any non-trivial program.
It also means a lack of efficiency. If I hand write code I can choose to do it in the most efficient manner possible, in alot of cases this is critical because otherwise the program will be dog slow. If I use cookie cutter functions to make things easier I end up with lots of generic redundant code which is aimed to be as generic and reusable as possible so it's the need of everyone who might use that cookie cutter. It's not tailored for MY use.
A good example of this is VB, VB is about as dumbed down as you can get (in the typically unintuitive microsoft fashion) and with VB you can get a graphical application up in a matter of minutes. However even the most trivial vb application would be faster if it was written in C or C++, it would use less space on the hard drive, use less memory and use less processor. And by most trivial I mean something that simply pops up a diaglog which says "Hello, world!" with an ok button.
"While I agree that the brains behind OSS are largely individuals, the developers themselves, Linux simply would not be what it is today if it weren't for companies pumping money into Linux."
;)
I'm not sure how you arrive at this conclusion but I have a suspicion. When companies (such as IBM for instance) started putting money into linux it was when the Platform as a whole was starting to mature.
It's been since then that alot of the big visible improvements have occured... but 99% of them didn't come from the corporations! They came from the same place the rest did, they came from the community. Most of the corporate money that has gone into linux has gone toward kernel features that are useful to a very select group of people (uber number of a processors, numa, etc). Sun contributed open office and that was significant, and redhat has contributed quite a bit... other than that I know of very little corporate money churning new features into open source.
Now by no means am I saying it's not a nice thing to see a company contribute, particularly by hiring an open source developer and paying him to work full time on a project. I'm just saying that presently and over the past 2 or 3yrs when all the great developments have happened that IS NOT what has been the driving force. It's simply been a matter of alot of things ripening in a given timeframe and more people (thus more open source developers) using the platform.
"If we had a system in which people didn't have to work, then the corporate contributions would be more or less irrelevant because people would have plenty of time to devote to free software development."
Now that sir I absolutely agree with. Good points though. It's always a breath of fresh air to debate with someone who's not a religious zealot of some sort or other
An OS X interface would just be a perk for OS X users. Gimp only runs on OS X become some OS X users would like it to.
It's main userbase runs it on linux which IMHO is what everyone should be running on their desktop (whether it be mac, pc, handheld, or IBM Bluegene).
The interface already integrates quite nicely in X11 and windows for that matter due to people interested in having it there. If you OS X users want it to integrate well with MacOS X I'd suggest you start coding, I'm sure gimp will accept your code if well written.
"cable "premium" channels which cost about 30 pounds per month"
This is what I was refering to, if you mean the BBC I believe that work out to more like $210/year but I was talking about the other premium channels in comparison to ours.
ah well thanks for making the effort :)
As someone else pointed out, with the codec itself open source it should be relatively trivial to pack it into a different container anyway.
"Such a choice is not my decision, and, for the moment, our concerns lie elsewhere."
Does this mean the issue isn't closed but won't be changing anytime right away (if it does)?
"Plus, I don't know where you're getting "150Mb+" from; my Windows Task Manager is right now reporting Photoshop CS as having a memory footprint of 34,076k."
It used to have a Hard disk footprint of 24mb. I was talking about size on disk not size in memory. At the time Photoshop 5 came out disk space was till an issue (actually it still is IMHO, but many would disagree... I tend to think plentiful resources aren't a valid excuse for bloat).
"No, which is why you can still do this - only now the type actually appears in your composition, rather than in a separate requester, which I think is a major improvement."
Actually I believe in 5 it gave a sample set of characters in the font. In later versions I know it shows it on the image, which I personally don't find an improvement over 4 but prefer to the behavior of 5.
"A comparison with this cost is the cost of satellite or cable "premium" channels which cost about 30 pounds per month"
I should have been more clear about what I was referring to.
"an OSS programmer working from free for no profit isn't likely to produce anything for you unless they don't have bills to pay and can afford to write softwre for you that will make you money but they themselves don't care about being paid. Very unlikely."
Odd, I was under the impression these are exactly the circumstances which MOST open source programmers write code (aside from the no bills thing). A few get paid, most don't, most write software in their free time. Open source development is typically cheaper if the program is useful to others anyway, and if you'll sponsor a project to get what you want it's the cheapest route to go.
"If you want good quality software with least defects that suits your customer's needs (implying customization)"
Good quality software with the least defects means using an open source codebase, if customization is needed it is done to that codebase. No private firm is going to debug nearly so well as 10,000 geeks.
Custom software is something that should be avoided unless there is absolutely no other option and situations which require it are rare.
Since there is likely an open source project at least similar to what is needed, why on earth would we start from scratch when we could adapt it and best yet, get the changes maintained for free if it's something that can be contributed back!
Right about now, your scratching your head and trying to figure out which role I'm coming from, management, developer, etc.
The truth is I'm a network technician/administrator/programmer, the company I work administrates the networks of hundreds of small buisnesses. If the company I work for, or one our customers needs software, then I may write it, may buy it, or may have it written depending on the case. If I write the software the customer (or my company) will be licensed the software under the gpl. Or I may modify an open source program to suit (probably contributing back changes if relevant).
Or we may hire an outside firm to adapt an open source program and license to us under a BSD style license with source... so we can release to the customer under the gpl.
I'd go further back than that, I'd say 4.1. Ok 5 did add some useful tools (something they haven't done since) but it added shitloads of bloat. Photoshop went from 20+ mb to 150+ mb in a single version and screwed up simple things, getting rid of a critical tool on the toolbar for one.
They screwed up type preview (is the ability to arrow down through the fonts and have them update on the type you've put in the box and are actually putting in the graphic really something you want to take away?
So yes they added some good functionality, but nothing critical, and they took away some I consider critical for productivity... especially in web graphics where half of them need text.
"both companies which make money on Linux, and companies which make the software"
Neither barney the purple webdesign website nor slashdot are the right place to reach either of these groups. (sorry couldn't resist the Barney thing, that website looks like it should be seeing feminine hygene products not giving advice on web design)
"The latter group, including companies like Adobe, are interested in what Linux users want for two reasons."
A little offtopic wouldn't you think, since the parent wasn't talking about linux users, he was talking about non-linux users saying they'd use linux if it weren't for this or that.
Also because this article was written by a MacOSX user, NOT a linux user. 99% of the things he complained about were really one thing, it didn't have an OS X interface, it's designed for X11.
"So, not only is it a waste of your time to complain about people complaining about the one last thing missing from Linux, but your will actually runs counter to the companies pumping money into Linux, or put another way, the hand that's feeding you."
That hands that feed the linux community are open source developers. While some of those are being paid by corporations, for the most part they are not. Corporations having money only means so much in this crowd, after all while there are exceptions, they've managed to produce mostly crap. Look at Microsoft for instance, they've managed to produce ONLY crap and they've got more money than any of the others.
"Photoshop is clearly one of the killer apps of all time. It was even ported to one or two kinds of Unix (likewise, FrameMaker) due to the fact that there was no (and still is no) superior product, and at the time Unix workstations were more powerful than PCs. Those days are gone, and so there is little to no motivation to develop new versions of Photoshop for Unix systems. On the other hand, Linux market share is growing, and people are starting to demonstrate a willingness to pay for software which runs on Linux. Not just corporations mind you, but also individuals and small businesses."
A piece of killer app proprietary software is great for linux adoption which ultimately ends up being good for the community, since it helps yield more open source developers to work on things like gimp so it eventually can get that proprietary app out of our otherwise clean open system. A better solution of course would be for Adobe to open the photoshop sourcecode.
Since I don't see that happening, opening the Acrobat reader source would be nice. The linux version of Acrobat reader is pretty shoddy to say the least, it's ugly, unpolished, and bugridden. If Acrobat reader were opened then we could either fix it (if it was worth fixing) or get a look at it's pdf handling, since actually opening and rendering the pdf is the only thing it does in a superior way to the other dozen pdf reading linux apps. AR is faster and has better print options. I think it could be salvaged myself, open source would quickly eradicate the bugs and fix the UI.
But did I miss something, or is this nothing but a Mac user whining that an X11 program doesn't have a native OS X interface and doesn't work just like photoshop.
That aside, are we really supposed to be taking advice from a guy who puts up a website which indicates it gives website design advice and the first thing you see is a banner add and a giant purple page...
Worth noting that even on win32 there is not only a setup.exe that works perfectly but the interface fits quite nicely (it's not like windows users REALLY expect standardized interfaces anyway).
Gee and I consider the cost here in the US for a package of premium channels at $50/month high...
You guys pay the equivelent of $53.xx a month for a single channel?
I was just wondering why you skipped the first response which asked about using an Ogg container format asking how proprietary the container format you are using is.
"There's nothing about Dirac, BTW, that requires C++, or even object-oriented programming."
;)
I'd tend to think that's a given since there is nothing that requires C++ or object-oriented programming.
Not to say there aren't some things it's faster and easier to implement with OO, there's nothing that requires it. And there's certainly nothing requiring it be C++ even then
I think your reading that backwards, I believe the parent was saying this is their response to those feelings of unfairness, not the reason for them.
;)
Then again I could be the one reading it backwards
Have you considered submitting a slashdot story rather than hitting random threads?
It's a great thing by the way, just think a story would amount to a bit more exposure.
"On the rare occasions when I can't figure out what some chunk of code does, I'll just comment it out and run the tests. If nothing breaks, then I delete it and move on. And if something does break, then I delete it and write something clearer in its place. It's very freeing!"
This statement says nothing about them not writting a test, it says if you see code you don't understand. And specifically says if something does break when you run the tests you rewrite the code it's clearer. And for that I refer you back to the post you replied to. It may have been something that was rewritten to correct a problem you didn't understand to begin with, and that's why it wasn't clear to you.
In the case of a rewrite that fixes a security hole, it's possible but unlikely the test would be an exploit, more likely the test would be the same as for the original code making sure it functions correctly. Looking at the test would only confirm to yourself that you were "cleaning up" the code while you were actually putting a security hole right back in.