The GPL self-destruct mechanism that is killing Linux

Re: A few more lumps of confusion in the article

Forks may be good for the software, but they are terrible for the customers and Linux devs should think of the people who use their software as a cusomer, not as users. The product may be notionally free, but this doesn't mean that the userbase should expect poor customer services. Forking a product means that customers who have already chosen to use that product are presented with a choice of two or more products which may or may not do all of what they wanted to do in the first place. If you pick the wrong fork you may end up a year down the line using abandonware and face a huge challenge getting back to where you were.

Re: A few more lumps of confusion in the article

"...The advantage of Mono is that if you make a profit Microsoft can change the license and sue you for patent infringement...."

Utter rubbish.

Imagine if the first caveman to invent fire had hidden the secret away and not shared it with anyone and just produced a "fire shop" that you could buy some fire from. I don't think humans would have evolved as quickly as they did. Thus, copying (including academic "copying", which is different to blatant plagiarism) is actually the only real sensible way forward. "Copy, and make better" is the best mantra you can have, and the unstated business plan of many a corporation.

But "better" itself is subjective. A brand-new Honda is almost certainly "better" than my own car in many areas. But if it's "better" for me or not is a much more subtle question - my requirements differ immensely to those of someone designing the perfect car. For a start, I need to afford it, and I'd quite like one I can repair without having to send it back to the manufacturer each time. Of course, a "better" car is likely to cost more and actually be beyond simple repair. But that makes it *not* better for my needs.

BSD vs GPL vs proprietary is a similar argument. Proprietary does things that the others can't (i.e. run the programs I need to use for work and games I want to use at home, although that situation is in flux at the moment, support all the hardware in my PC, etc.). But equally the open-source offerings provide me with advantages that I can't get from proprietary software, like being able to hack into the source and change stuff (and, although I'm one of the few that can, this has saved my employers lots of money several times already on everything from fax systems to access-control systems to simple things like making use of old hardware), and being able to deploy as many units as I like without counting licenses.

And even, when it comes to it, making up for some proprietary shortfall with some knocked-together solution. A case in point? How about an expensive proprietary access control system that stores all its data in a Firebird database on the controller machine - think SQLite, the whole database saved in a folder and run locally but you can still do "database-y" things to it? We wanted a list of people who are on the premises when we click a certain button. Proprietary offering is another £x plus some more on top of software we already had to pay for and doesn't really do much. We can't hack into the program or fake a key to give us that feature without breaking the license agreement. But I can load a Firebird-compatible DB layer onto a Linux machine, probe the database remotely over Samba, throw in a couple of SQL statements and viola, my results - Samba is GPL, Firebird is MPL, my code was "who cares, only I get to see it"PL, and I get my solution. No doubt the proprietary software is "better" and would do a better, more accurate, more coherent job of that task. But in terms of the end result, my hacked-together script is "better" for my workplace (so much so, I had their reseller's field engineers ask me for it so they could use it themselves).

Open source isn't "better" generally because nothing is. And if someone doesn't understand that, then I doubt they understand how to argue their own system is better anyway.

The alternative argument is that if you can't just copy something you waste endless man hours reinventing the wheel.

Citation needed

Can you give me an example of elimination of competition due to open source?

@Pete : Re: Divide and conquer

'Instead of being usable by "the average person" it requires enormous, duplicated effort by each individual who wishes to install some of its parts - usually requiring a whole load of other dependent parts, too.'

Thank you for repeating the same, old, tired bollocks - I was wondering when that was going to make an appearance and was worried i'd missed it.

Please pray tell, exactly what things are required by the "average user" on your average Linux-based OS that require "enormous, duplicated effort by each individual who wishes to install" it ?

Re: Divide and conquer

A common misconception is that EVERYONE needs to contribute back. It's just not true. I have code in open-source projects out there, but it's pretty minimal to say the least, and I have several private patches that I wouldn't dare to pollute a foreign codebase with. But most of the people I know who are the largest users of open-source don't contribute anything at all, and yet it still thrives and grows every year (use Firefox? You're one of them).

The reality is that *I* benefit from other people being able to see the code and play with it. The usual argument is security, but it goes far, far beyond that. A single patch, approved, tested and posted to the Linux kernel will end up on MILLIONS of machines within hours. I benefit from that. I benefit from moaning about it not working too. I benefit from other people looking at code and saying "I can't understand that, it looks like complete nonsense" on any of the projects that I use. I even benefit when projects are forked or abandoned because of that (because otherwise *I* would have to fork myself, be left without support, have to start up a rival project from scratch, or go seek out alternatives on my own).

99.9999% of GPL-licensed software users push back exactly zilch. That's not a problem at all. Nobody really cares, but it's more than, say, the number of Microsoft users whose code ends up in Windows (good luck with that!). You're confusing users with developers, though.

As the barest of bare amateur developers, I have tested, patched and hacked the code I have available to make it work the way I need to. Everything from patching the rt2500 drivers on my private systems with a patch that I had to craft to get it to compile on a new kernel (pre kernel inclusion, something to do with the way interrupts were handled changing in that kernel, if I remember), to the patches I have to make to my own copy of Hylafax to get it to run the numerous fax lines in my workplace without forcing me to upgrade to the next version to get feature X (risking a hefty mid-cycle upgrade), to fixing TuxPaint (I work in schools, it's one of their most used programs) to juggle the menu items to make them easier for little kids, to providing a routine to OpenTTD that reduced the amount of bug reports they got where people were using hacked/unofficial datafiles (which has since stopped a lot of spam on their bug monitoring and provided several people with the knowledge that their datafile were unintentionally corrupted - who knows, maybe I found out for someone that their hard disk was corrupt by that patch!).

Those little changes are the freedom I pay for. If notepad doesn't want to open a particular file, there is bog-all I can do about it. But if *metapad*, the program I use in preference because it lets you do lots more, does it then I can work out why and change its behaviour, or lobby to have it changed. I don't expect a user to, but like some of the things mentioned above, after some time they may be able to do exactly what I've done without having to know how (the first person to invent the wheel was a genius, the people who followed after benefited from his genius, and now we all take wheels for granted).

Similarly, I just hacked Classic Shell because that open-source project refused to allow in a feature that I think it needs (an option so right-clicking the toolbar provides the "old" Windows context menu by default and the "new" Windows context menu if you're holding Shift, instead of vice-versa). So I hacked the code myself, added the feature, and *I'm* better off even if no-one else is. And I didn't require their co-operation at all (and received just the opposite). But how many people can actually go and tweak their *Microsoft* Start bar or put it back into Windows no matter how nice they are to MS.

The GPL, especially, gives users and developers the freedom to benefit from each other. It's arguable that the average end user ever really benefits from MS developer's wild ideas. But certainly, though your granny can't hack into the KDE source code, she benefits from it being available to others.

Think about the movie I, Robot:

Man (sarcastically): "Can a robot create a masterpiece?"

Robot: "Can you?"

Arguing that users don't benefit is like saying that air passengers don't benefit from someone looking into aircraft designs, their safety, effectiveness and room for improvement, from outside the industry, or even the general public provoking outcry when a particular type of plane keeps having problems. Of course they do. They just can't necessarily do it themselves directly.

The GPL's epitaph? "Someone better came along and replaced us. Mission accomplished."

Re: towards the organised and fatally disciplined ranks of professional software developers

So you're one of those people who think that only amateur hobbyists are Linux developers? Or are you just a FUD-spreading shill?