“In 10.10, we’ll take a big step forward in the netbook edition, by
moving the menu into the top panel and combining it with the titlebar
for maximised windows.”

Now look at the image on the right.  It’s a screenshot of the proposed ‘Unity UI‘ that is being ‘developed’ by Ubuntu inside an OSX screenshot.  Note the monochrome icons in the system tray, the Apple / Ubuntu logos with identical spacing, the popout options on the dock with the identical black background, curved light borders and association arrow.  There is even a ‘Keep in Launcher’, ‘Keep in Dock’ option.  Ok I suppose the current running program indicator is an arrow instead of a bright spot, but that’s only because they are copying Tiger rather than Snow Leopard.  Add in the recent pointless change of moving the window decorations to the left.  It’s all very suspicious.

Although once you take the above quote into context the only significant difference between them disappears.  That is it is going to get an OSX style universal menu – that is the File / Edit / View menu moved to the top left outside the applications window.  I mean come on, there is being inspired and then there is outright plagarism and there really is far too much similarity here to chalk it up to chance.  As I said before it’s all cargo cult usability – the focus is on copying Apple because they care about usability rather than copying the reasons Apple do what they do.  It’s a significant difference.  It’s all irrelevant anyhow though.  Here’s why:

THE DESKTOP IS A SOLVED PROBLEM!

It’s 2010.  It doesn’t matter any more.  Why are we even talking about this?  I remember years ago, before the Internet was fast or reasonably accessible, playing with the OS for the seemingly sake of it, but nobody ever does this anymore.  Simply put the OS’s only job is to run the important thing – that is the applications.  Nobody tweaks their OS.  The reason you can’t change themes on OSX is because the majority of people don’t want to do that anymore.  Windows 7 lets you change the colours but that’s about it.  Essentially the OS has two jobs, running apps and bothering you as little as possible.  That’s it.

But no, we have Fedora, Ubuntu, Suse, CentOS, Mint, Kubuntu, Edubuntu, Ubuntu Studio, etc, etc, etc.  It’s a constant reinvention and minor tweaks to the overlying thing that doesn’t really matter (the OS) while the thing that really really matters (the apps) languish in obscurity as ‘upstream’.  With Windows (and OSX) there is a vibrant third party community developing all sorts of apps independently of their target platform.  With Linux it’s all about the repository and what apps you get is based on what distro you use.  You get your software updated when the distro decides you should get it updated, not when it is actually updated.  It’s not at all uncommon to have the repo version as several versions behind what is in the wild and if what you want isn’t in the repo you are screwed.  All the talk of ‘choice’ is utterly irrelevant as no matter what distro you ‘choose’ you will still have the exact same selection of software (you know, the things that matters) as every single other distro.

So why are the FOSS communities version of the record industry – the ‘packagers’ – where all the focus lies?  Why does everyone get whipped into a frenzy every few months when some new distro comes out with a fancy new wallpaper and a fancy new theme?  Why are we even still talking about this?  The only reason I can see to ever upgrade your OS is because the new one brings something significant and most people take a largely ‘wait and see’ approach yet with Ubuntu you have the insane BiannualForcedDeathMarch™ and if you don’t upgrade you can forget about ever having newer versions of the software.  Why on earth is there no sane way to install software outside this byzantine system?  LTS support?  Fine if you don’t mind them not bothering to update or add any apps to the repo.  Then they moan about Microsoft not following standards then require software developers to waste days making an individual releases for each version of each distro (see Skype) as they can’t even agree among themselves.  I can get software that was released for Windows 15 years ago and provided the developers didn’t do anything stupid it’ll still work.  Why is doing this so hard?  As I said it’s a solved problem.

Essentially no amount of distro tweaking is going to make Gimp stop sucking, or make OO any less bloated or magically make a decent selection of games appear.  The best Shuttleworth and co are going to end up with at this rate is a copy of OSX without any application support.  What’s the point in that?

If a boat is sinking and when someone does a speech ‘hey, the boat is sinking’ it makes waves* then the issue is not that the boat is sinking, it is instead that nobody has noticed the boat was sinking.  Fixing the boat itself will do nothing.  Suggesting fixes to the boat will do even less.  Your core issue is that nobody actually sees that the boat is sinking.  “Guys, I gave this presentation a year ago and there is still water pouring in the sides, what we need is a team to set up a fund to pay for a crew to board up the holes that you guys keep making.”  Lets face it it’s not really tackling the actual issue – people keep putting holes in the boat.

Even saying something as fundamentally obvious as ‘Choice in software is good, choice in standards is bad’ creates controversy and argument.  Essentially what has happened is Linux has become a movement, rather than a bunch of lines of code.  Tenets of that movement such as ‘choice, freedom, superior development model’ are viewed as infallible, thus anything which casts doubt on this is attacked.  Saying ‘standardize on a package format’ goes against the ‘choice’ tenet and as such you’ll get slammed for it.  Saying Gimp isn’t very good – especially compared to Photoshop – goes against ‘superior development model’ and again, you’ll get slammed for it.  Many people actually believe that Gimp is equal or better than Photoshop – which is incredible.

It seems to me that the problem is a culture of defensiveness and denial is at the heart of the issue.  Linux uptake is staggeringly slow (~1% marketshare and, what, 50% increase in the last 5 years?) yet the attitude among the supporters when faced with unhappy users is “If you don’t like it piss off, it’s free”.  The community consists entirely of people who like it more than the competition – people who don’t like it leave pretty quickly.  If the Leopard Print Couch Company wasn’t doing so well, do you think asking a cross section of the people who own Leopard Print couches would provide valuable information into why they were not selling much stock?

I suppose what I am trying to say is that the Linux community reminds me of the Leopard Print Couch Company, who want a couch in every house but can’t figure out why it’s not happening.  Someone walks into the shop and says ‘I don’t like Leopard print’ and they’ll get told ‘everyone else here likes it’ and tell the customer to leave and then continue on discussing changes to the pattern “If only we could add more spots, then we’d succeed!”

The feeling I took away from this presentation was that which I get from virtually all these ‘lets be humble’ sessions – “we are nearly there, we just need to sort out a few little issues and then it’ll really take off” – when from where I am standing there is a massive, and widening, gulf between where Linux is and where it needs to be.

The true issue is that the people responsible for the holes in the side of the boat don’t even realise that they are creating them.  Pointing at the holes is pointless, getting the people making them to stop doing it is the actual solution.  Good luck with that though, as the community largely likes the holes**.

* Sorry!

** Just as the Leopard Print Couch Company wouldn’t change pattern – everyone likes Leopard print after all, just ask our customers!

What am I referring to?  Comments such as this (or Google your own) …

Maybe if it wasn’t so easy to change it would matter a little but now it’s starting to sound like a bunch of spoiled babies crying because mommy didn’t cut the crust off of their sandwich. Get over it people…

In the above example the zealot is saying how you shouldn’t complain about a usability regression because it’s so easy to change back.  Here’s what people are referring to as easy:

It is easy to change back to the traditional top RHS.
Alt +F2 to open launcher
gconf-editor
select: apps/metacity/general/button_layout
Place ‘spacer, after ‘menu:’
menu:spacer,maximize,minimize,close

It’s almost as if you say something is easy enough times it’ll suddenly be easier – such as ‘updating is easy, just type sudo apt-get dist-upgrade’ – ignoring the fact that words have actual meanings, and changing the meaning to make something seem better than it is is dishonest.  Here’s the dictionary definition of ‘easy’:

Easy: achieved without great effort; presenting few difficulties : an easy way of

Of course when you actually bring up the fact that it’s not easy you’ll get this fun little qualifier added ‘It’s easy when you know how‘.  But as far as all the dictionaries I have checked none have ‘when you know how’ anywhere in the description.  And why not?  Because ‘when you know’ negates the whole meaning of ‘easy’.  Here’s a bunch of other things that are ‘easy when you know how’.

• C Pointers
• Regular Expressions
• Mod Rewrite
• Juggling
• Professional Snooker
• Tightrope Walking
• Multidimensional Arrays
• Assembly Language
• Windows Arabic Edition (it’s easy once you learn Arabic)

You see where I am going with this.  In fact I can’t think of a single thing that isn’t ‘easy once you know how’.  Even the deliberately obfuscated languages such as Brainfuck are probably fairly simple once you get to know it well enough.

CLI vs GUI

Think of a GUI as a complex network of roads and paths.  You have all the paths in front of you and can see the main highways (start button) and little hidden side-roads (control panel) but although they are sometimes twisty and poorly signposted it is possible to get from A to B without a map.  Not that a map wouldn’t help, and not that the signs can’t be better, but getting from A to B does not require one.

The CLI on the other hand is like the same network of roads, but instead of having signposts you have a blindfold.  You need to know exactly what road you need, where it is and how to get down it.  One mistake and you are lying in the ditch.  It’s the whole point of a GUI and why they have been so wildly successful – they make things easy.

I remember telling a veteran Linux user about tail -f to monitor updates on a log file.  It’s not because he was stupid, it’s because he had never ever read the particular bit of documentation that described this feature.  That is, to effectively use a CLI based interface you must have already read and remembered everything about it.

If we take the above example as a GUI then the program will often pop up a message ‘changes detected, do you want to update to the file stored on disk’, with Yes, No and Always as an option.  There’s no way you could not know about it.  Same thing with mounting volumes, I’ve used various GUI partition tools for the last 20 years and the only thing I need to know going in was what I wanted.  I used the Linux mount command for the first time and wasted the best part of an hour due to not knowing to add ‘-umask=0666′ to the command to make it user-readable.  If that was a GUI there would have been an ‘allow user access’ checkbox and the problem would not have existed.

If something requires a mass amount of prerequisite knowledge then it is not easy!  If there is no way of knowing that something is even possible, let possible to figure out yourself then it is not easy!  Just because once you have invested the months of practice into learning whatever you happen to be using (Bash, VI, Emacs, Perl, etc) you can do something in 5 seconds does not make it easy!  If you have to copy line by line from the internet into a terminal window with no idea what you are typing means, it is not easy!

Hell, I am at the bottom of this article and I simply can’t remember what the steps were that were outlined as ‘easy’ – so if something can be written and read multiple times and you still can’t remember how to do it, it’s not easy.

Update: That’s not to say you don’t get some terrible GUI apps – you do.  But the natural state of a well made GUI app is to be intuitive, obvious and easy.  A CLI apps natural state is to have a mandatory RTFM requirement.

The point of this is to try to keep your customers happy, and to listen to their concerns.  If a customer is not happy with something about your service or product no decent business would just tell them to ‘piss off’, instead they would try their best to address the customers concern and to try to prevent such occurances in the future.  For example if you sold a spade which the handle kept coming off you’d either replace the item or refund the customers money and, but more importantly, you’d try to identify the cause of the problem.

While often customer complaints are a one-off and usually end there, what is important is to keep a lookout for repeated complaints as they indicate a systemic problem.  For example in a restaurant if a certain dish gets sent back regularly you wouldn’t say that the customers have no taste, instead you’d look into what is in it and how it is made in an attempt to solve the issue entirely.

Customers vs FOSS

Take this blog from Preston Gralla for example.  He says that installing software that is not in the repo’s is too difficult and is holding back Linux adoption.  Not a new statement in the slightest, and one that I both agree with and am sure I have even said before.

It’s not even like he’s being an angry shouter like a lot of us embittered haters have become, he’s clearly following the ‘make every second paragraph praise’ approach which is required (as an offering to the Holy GNU) when writing any article that dares be critical of Linux.  Not that it helped him at all anyway.

He is a ‘customer’ of Ubuntu, just as I am and just as thousands of other people who no doubt share the same view are.  And while many people are about to say something pithy like “Linx doesn’t want you”, may I remind you of Ubuntu Bug #1 – Microsoft has majority marketshare – implying that Ubuntu actually wants customers.

Yet this guy says something that, to him, is an impediment to him using Ubuntu and immediately gets his head bitten off by a horde of angry Linux users who then post massive amounts of comments saying how he is wrong, calling him a shill, claiming he’s getting paid by MS, claiming that the site is fundamentally biased (despite it being home to the infamous SJVN) and generally denying that the problem he outlines could possibly exist, and it’s only because he’s either stupid, or being paid that he can come to such a conclusion.

The following two questions then get raised: Will this torrent of abuse somehow make him change his mind about his claim and realise that the problem he has suddenly isn’t a problem anymore?  And more importantly will he be more disposed to trying and promoting Linux after recieving those responses than before?

Or consider if you went into a bar and ordered a drink, and the drink tasted like cleaning fluid (or otherwise nasty) and you pointed this out to the bartender.  If he (and the other patrons) proceeded to call you an idiot moron you would simply not ever go back and tell all your friends about the bad experience (welcome to my blog!) rather than go ‘you’re right I am a moron’.

The customers complaints are always valid.  Calling them a moron and denying they have a complaint does solve the problem – you no longer have them as a customer.

Installing Software In Linux Sucks

Besides, he’s right.  It’s amusing that the community that touts ‘choice’ as it’s primary selling point presents the argument that he should just wait until the distro updates the repository, rather than be able to easily use new software straight off the bat.

When I was trying Linux I had endless issues with the software in the repo’s being massively out of date, and what do you expect?  You have tens of thousands of apps to track and keep up to date and it’s not even like the iPhone or Android app stores in that developers don’t necessarily submit new versions requiring the repo maintainers to find out when there has been an update (good luck with that).

In an ideal world the repo’s would be all you need (and communism would actually work) but in reality the system of repositorys needs to be supplemented by a system of manual installation in the cases where the software is out of date or simply unavailable.  And the brutal truth is Linux falls flat on it’s face here.

Manual Installation

Welcome to the quagmire that is manual software installation in Linux.  As soon as you’re outside the walled garden of vendor approved obsolete versions of software to be found in the repositories then you’re largely out of luck.

How often have you seen Microsoft being berated for making it’s own ‘standards’ for things like word documents and protocols?  How many Linux evangelists have you seem complain that .docx and .doc are not compatible?  Yet these very same evangelists will argue ’till they are blue in the face that somehow multiple incompatible package formats and standard breaking distros is somehow a ‘good thing’ – usually under the guise of ‘choice’.

Hell, I’ve even written a post bashing the Windows software installation method and since this is technically an anti-Linux (and thus visited largely by Windows users) blog I should be getting angry posts saying how I am wrong, and an idiot, yet all I got was agreements and clarifications!  Yet if you say something as BLINDINGLY OBVIOUS as Linux needs a standardized package format you’ll get flamed to a crisp.

The Dead Parrot Horse

I am flogging a dead horse with my point but I think it deserves to be made (and flogged).  Being critical of Linux always reminds me of the Dead Parrot Sketch.  The community is largely unable to accept any** criticism as valid, no matter how obvious or problematic, with the first approach to user complaints to be to deny they exist, then to call the complainer an idiot troll.

This is what is holding Linux back from going mainstream – it’s the fact that the community simply doesn’t care about the needs or issues of the people they are trying to foist Linux upon.

The claims of ‘community development’ are a massive lie in that aside from posting the occasional bug reports (which you can do with closed source) anyone with a problem is abused and faces North Korean levels of censorship***.  You either take what the Linux cult gives you and be quiet, or you simply don’t use it.  Trying to contribute improvements and suggestions just gets you into trouble.

If you’re happy with Linux, sure use it.  If you’re not don’t even bother – it’s simply not worth it.

* Not if I am working there they aren’t

** I think one of the sources of my Ubuntuforums ban was due to an argument where I was trying to propose double clickable .deb files.

*** Preston, try to have the discussion on your blog post on a Linux forum (Ubuntuforums is a good choice) and see how fast the admins delete your post then ban you.

Of course there is absolutely nothing wrong with distributing something under the GPL.  If you are committed to freedom then you’ll accept that people should have the freedom to use whatever license they please.  What is problematic is the popular view that somehow releasing your software under the GPL is somehow ‘ethical’, with some people even suggesting it should be a legal requirement for all software.

If you believe the revisionists they would tell you that ‘free’ is the natural state of software until companies (mainly Microsoft) came in and ruined everything by ‘closing’ the software and charging money, snatching software utopia out of the worlds hands in the process.

But the problem is this: Why is software the only thing covered by these fundamental freedoms?

Books, Software, Movies and Music are all technically the same thing, they are an infinitely reproducible product based upon human endeavour, differing from standard creations (such as making a chair) in that making 1,000,000 is almost as easy as making 10.  Only the first one took the time to make.

The ‘rights’ outlined in the GPL are not just the right to the source code – that is tangential to the issue.  The main ‘rights’ are for free modification and redistribution – If I receive any GPL’d software I am free to edit it and give it away for no cost.  The claims that ‘you can sell GPL’d software’ as a counter to a claim that it creates an unworkable business model are intellectually dishonest since you can only sell it once – after that you’ll be competing with free.

If I were to buy a new book by Iain Banks, should I have the right to edit, make copies and then give those copies away?  Should I demand the original document rather than the printed version as by not being able to do the above easily with a physical book my ‘freedoms’ are being compromised?  If I was to make this argument to most people who support the GPL as a vehicle for software progress I would be laughed at, but it is fundamentally the same thing.

The argument could be made that books are for entertainment, but education largely comes in book form.  Also there are no clear cut lines between media anymore.  If I draw a sapceship on paper, do you have the ‘freedom’ to take it and give it away?  What about if I make a 3D rendering of it?  What if I make the 3D rendering display as a runtime exe?  What if I make it interactive and flyable?  At what point do your ‘fundamental rights’ kick in and allow you to do whatever you want with it?

The belief is that if the GPL was enforced today then the world would be better off, and while true this fails to consider the future.  If these ‘freedoms’ applied to the world of literature then there would be a wealth of works suddenly made available for low or zero cost.  What would then happen though is the individuals who before were creating a livelihood on the sales of their creation would suddenly have no income and be forced to find other avenues for money and while some may be able to monetize their fame most others would not.  As a result you would get people like Stephen King working as shelf stackers in supermarkets rather than doing what they do well and sure, he can write in the evenings and weekends but he would only write a fraction of what he could under the previous system.

The ‘freedom’ model essentially pulls the rug out from under the system of rewarding artists that has worked for hundreds of years.  People claim that you can ‘sell support’ and that it is just as profitable, but that is blatantly not true in many cases – only certain software can be made profitable.

I want a system where individuals with talent are able to do nothing but exercise that talent, rather than a system where they have to work mundane jobs to simply make a living and are only allowed to create greatness in the evenings and weekends.

I love the self delusion of the Gimp developers.  Look at their site under ‘contribute’:

“GIMP is Free Software and a part of the GNU Project. In the free software world, there is generally no distinction between users and developers. As in a friendly neighbourhood, everybody pitches in to help their neighbors. Please consider the time you give in assistance to others as payment.”

I’ve always thought there was a massive distinction between artists and developers.  I think what they meant to say is ‘unless you are a programmer we won’t listen to you or cater to your needs’.

Here’s a conversation I have over and over again:

Linux Zealot: You should use Linux, Windows sucks and is for morons.
Me: I need stuff which you don’t get on Linux like Photoshop
Linux Zealot: Use Gimp, it’s better than Photoshop plus it’s free!
Me: Gimp sucks, (insert pages of things that it can’t do here)
Linux Zealot: If you don’t like it don’t use it, stop complaining it’s free!

So to the people pushing Linux (and Gimp) onto users, you have two choices:

1. Shut up and stop promoting it (and calling people idiots for not using it) or…
2. Cater for the needs and desires of your userbase and own up to and address the flaws.

You can’t claim it’s better as well as telling people to not complain as it’s free.  As soon as you say it’s better and evangelise on this basis and it’s not actually better it just makes you a liar.  And if you evangelise on the basis that it’s better and people say it isn’t then their opinions are valid and should be listened to.  Also, telling people to ‘fix it yourself’ does not count.

As soon as you tell them they are, in fact, wrong and it is better* – as the Gimp** supporters love to do – then you have crossed the boundary between lies and self delusion.  Enjor your stay, the Ubuntuforums are first on the left.

* The usual excuse is ‘you are thinking in the ‘Microsoft’ way.  If you’d tried Emacs/VI/Latex before Word you’d find it just as easy’.  Or just plain old WorksForMe(tm).

** Oh, and the name is still embarassing, offensive and exclusionary.

So Monty is complaining that Oracle may end up owning MySQL.  Here’s a reality check: If you sell something for ONE BILLION DOLLARS you have no right to dictate the terms of that product anymore.  You sold it, game over.

Sun’s aquisition of MySQL probably helped sink that ship, unlike the flotilla of private yachts that Monty now owns thanks to Sun’s money.  That ONE BILLION DOLLARS he took in exchange for MySQL must be converted into business value somehow to justify it’s price and Oracle are probably going to do just that.

If he really cared about ‘freedom’ he wouldn’t have taken the big businesses money and sold out.  If he really, really cared he’d take that ONE BILLION DOLLARS and simply buy MySQL back as I am pretty sure it’s not worth now what he got for it.

But no, he sets up a site to try to strongarm Oracle into essentially losing all benifits to an investment valued at ONE BILLION DOLLARS because he doesn’t think they can be trusted to keep it ‘free’.

What an idiot.

But first some numbers…

Linux vs Windows (Google Trends)  Observe the steady decline.  Not exactly a pattern that’ll lead to anything significant.

The much touted W3 Schools logs.  These are horribly biased as it is only for w3schools.com itself, is cited so often to prove dubious claims of marketshare that itself probably skews the numbers, not to mention that Linux is obviously going to be skewed on technical sites, yet it has only shown Linux use double in the last 6 years from 2.2% to 4.2%.

I was going to show statcounter as well but apparently Linux’ usage has fallen so low it’s now classified as ‘other’.  I could have sworn it had it’s own heading.  Oh well.

And finally hitslink.  Shows Linux usage at 0.98% at December 2008, a peak of 1.17% before dropping to 0.96% in October 2009.  A net loss.

Obviously if you pull up the logs of Slashdot or some site like that you can get some more ‘convenient’ numbers, but I don’t think I have ever seen a positive trend itself.  Usage is not increasing in any significant fashion.  If it is provide me with the stats (seriously, do it) but until this point I think my assertion stands.

As an anecdotal point, across the board on all the websites that I admin, Linux usage sits at about 0.3%, which includes music festivals, dentists, estate agents and the like.  As soon as you remove people who’s interests are not computers themselves (and view them as a means to an end instead) Linux’ numbers plummet even further.

The Myth of Marketing

Of course the usual answer from Linux advocates at this point is that it is because of Microsoft’s advertising money, and that the only reason people don’t use Linux is because they haven’t heard about it.  Which is now provably false.

1: Linux has one of the largest grassroots movements on the Internet.  In fact I am coining “Kerberos’ Rule #8 – If someone mentions OS’s, someone will mention Linux.” You can’t have a technical discussion anywhere without someone trying to ram Linux down your throat.  I am pretty sure the majority of the technical community knows what it is, yet strangely choose not to use it.  Is this marketing?  Are MS suppressing Linux?  What about the Streisand effect?

2: The real killer is this.  Windows 7 managed to, before it’s official release, pass Linux in terms of usage.  Everyone who had it installed at that point had:

• Known what it was and where to get it, despite the fact it wasn’t ‘marketed’.
• Managed to burn it to a DVD.
• Managed to install it on an existing system.
• Managed to get it working with all their hardware and apps.

Yet these are the very people that, according to the Linux community, should be using Linux instead.  Yet overwhelmingly they all went with Windows, despite the fact that they obviously have the technical ability to obtain it and install it which the proponents of Linux often claim is easier that Windows.

Linux Is Faith Based

Obviously pointing at a bunch of facts as I have done doesn’t lead to a civil discussion, despite all I have done is point at some facts – and not even said anything inflammatory.  Even saying something like “Linux uptake is remaining flat despite claims of superiority, which doesn’t make statistical sense.” will get you lynched on most pro-Linux forums.  Even saying “Linux is not suitable for my needs” is often enough to incite a flamewar, and stating why will generally get you banned.  It’s why I started this blog in the first place.

The reason for this is obvious though, once you replace the idea that you are dealing with Linux advocates on a basis of logic and science and instead realise that it is a faith based initiative, and religious rule #1 is “Free software is superior to commercial software” and rule #2 “Microsoft is the antichrist”.  Any challenge which questions these rules will elicit a fear/hate based response as would telling a religious person God doesn’t exist*.  Such a statement is fundamentally incompatible with their worldview.  This is proven in the fact that there is pretty much zero critical discussion of Linux occurring anywhere outside ‘hate’ blogs, despite the claims of a community created OS.  So in a world where ‘everyone is a developer’ there is almost no debate and critical feedback**.  That can’t be either right or healthy – but it’s the truth.

What the ‘community’ really is is an evangelical mission to spread the word.  It’s not about self reflection and self improvement, it’s about embracing the two rules and then trying to bend the world to fit them – and calling anyone who doesn’t agree a ‘noob’ and ‘stupid’ along the way.

Pragmatism

I like to think of myself as a software pragmatist – I don’t care which faceless corporation that pretends to care about me produces it – I’ll choose whatever is best for the task at hand.  At the moment I have…

• A Windows 7 laptop,
• An iPod,
• An Android based phone,
• A Linux (Smoothwall) based firewall,
• I also develop websites for LAMP on Windows,
• Plus numerous managed Linux, Windows and BSD servers

The point being I’ll get whatever fits my needs the best at the time.  I’ll get all the available information about all the options, compare them to what I need and base my decision upon this information.  If someone asks me for help I use the following process:

1. Ask them what they will be needing it for
2. Find out the skill and experience of the user
3. Make my recommendation based upon the above

Here’s how your typical Linux advocate makes their decision:

1. Use Linux, it’s the best!

You can see the problem here.  It’s faith based, they have ‘belief’ that Linux is good (and correspondingly that Microsoft is evil), thus they have their answer.  Actual user requirements are irrelevant as they know that Linux is superior, and that Windows sucks.  From this point on it’s simply a case of trying to convert the person they foisted Linux onto to their world view – which generally involves telling them they are wrong and bashing MS whenever a problem arises.  The question ‘would this person actually be better off with Linux’ is rarely, if ever, asked.  Someone being happy with what they have is never even considered.

You see it all the time.  ‘Why aren’t you using Linux’, they ask – the assumption of superiority is in the question.  ‘It doesn’t run Photoshop’, I reply.  Then we go through the whole ‘try Gimp’ rigmarole, then the ‘try Gimpshop’, which is the technical equivalent of putting a Humvee bodykit on a Ford Fiesta and sending it to Iraq. Eventually Wine is mentioned – which is fine if you are 3 versions behind and don’t expect everything to work or be stable.  The fact that I’d have to invest large amounts of time and effort to simply be at, best case, a par with what I had before I ‘switched’ means nothing to these people.

Which is the crux of the matter – you cannot expect a Linux advocate to give you honest advice.  Pretty much regardless of requirements they will suggest their favorite distro.  I’ve seen people recommend Gentoo as an easy beginners system. ‘Gaming is fine with Wine’ and other absurd claims.  The importance seems to be getting people to use it, not getting people the best system.  It would be like asking a fundamentalist christian in a bookshop what book you should read, or asking a Vegetarian*** what restaurant you should go to – the answer almost certainly would not be based on your requirements.

Lies, Damn Lies and Linux Advocates

Another side effect of the blind belief effect is the ability to play fast and loose with the truth.  If it makes Linux look good, it’s true.  If it makes Microsoft look bad, it’s true.  Anything else is FUD and lies.  A once run benchmark from some random idiots blog somewhere will be taken as proof of Linux’ success if it is favorable.  A large companies benchmark will be called lies if it is unfavourable.  Facts are not rated based upon their veracity nor citations nor testing methodology.  They are rated based upon how well they fit the agenda.  The words ‘scientific method’ are seen cowering in a corner somewhere.

Take for example this review of Windows 7 – I think someone linked to it on LHB originally.  It’s one of the most unbalanced hatchet jobs I’ve ever seen, yet is put forwards as being ‘fair and balanced’.  Gems include:

In the RAM usage front, Windows 7 when idle takes a good 1-2GB of my 4GB of RAM, while Kubuntu takes between 200MB-500MB

It’s pretty well known that Windows Vista and now Windows 7 use the RAM you have.  Any half decent programmer knows you can sacrifice RAM for performance and vice versa (largely through caching).  If you put Windows 7 on a box with 512mb it’ll idle using about ~250mb.  It’ll run fine too.  It’ll generally take half your RAM for itself to speed things up but that’s fine – it’ll relinquish it if programs need it rather than using as little as possible and thrashing your disks constantly.

Yet this bit of information, despite being infinitely sensible, and which has been common knowledge for years, is somehow not known to an alleged IT pro.  If Linux treated RAM in this way you can be guaranteed we’d see a spiel about Windows inability to take advantage of your system.  Also, nobody runs an idle system.  Even if you did use a RAM-light version of Linux (which would not be the fully-featured Ubuntu this guy is chatting about) you’d be stuffed as soon you open Firefox or try to do anything as that’s where the real RAM usage is.

But everyone always trots out the RAM argument as it’s on the pro Linux talking points list, despite RAM being dirt cheap and plentiful it’s still viewed as somehow better to spend weeks tweaking a low use system than just buying 6gb or something ridiculous.

Then this classic:

This is a common element of debate, and one I happen to agree with. Windows 7 seems to have borrowed a bit from the KDE4 series. Now days, KDE has changed it’s look to the “Air” theme, so it no longer looks as close. However, the Oxygen theme that was used in KDE around the same time as when Windows 7 was in development is mysteriously similar. Even the selection of widgets is similar, with both shipping with an analog clock, news reader, sliding block puzzle, and hardware monitor by default. Take a look at this quickly put together collage, with widgets from both Windows 7 and KDE versions previous to 4.3

The dude doing the review has apparently never so much as seen a screenshot of Vista, yet alone used it. The clock, sliding puzzle, black gloss, RSS reader, and every other thing he mentioned are stolen from Vista.  Unless Microsoft owns a time machine it’s pretty clear that KDE ripped off Windows.  Hell the guy proved it himself – but it makes Linux look good so fact checking isn’t important.  I even linked to a video ages ago showing that Compiz is an early Vista ripoff – wobbly windows and all.

As stated earlier though this reviewer had obviously already got his conclusions before Windows was even installed, it was simply a case of hunting for the facts to fit.  If the aim was impartiality or fairness he would have attempted to find out why Windows uses the RAM it does, and would have tried it on a 512mb machine, and would have at least used Vista to get a feel for the improvements in 7.  I’d also bet he is a ‘Vista Sucks’ cheerleader – despite never even having seen it before.

The rest of the his points are the airy “I prefer Linux” type of conclusions where he just looks for a bunch of things he can criticise and write up.  I mean a ‘fair’ test would be to use it for a week or two for actual work (although a month or more would be preferred) and then report back on the experience.  This review looks like he installed it, played for an hour or so then ‘reviewed’ it.

Yet crap like this never gets called out.  It’s standard practice.

I have a point? Wow!

The whole point of this post is not to stoke some imaginary ‘Linux vs Windows’ war, but instead to try to get the supporters of Linux and FOSS in general to take a look at and to hold themselves to the same standards they accuse Microsoft of breaking.  You can’t claim MS are bad for spreading FUD while making BSOD jokes.

Also instead of getting defensive when someone says something bad about Linux consider that although you think it’s not a problem they certainly do and thus it is important.  There is a reason nobody is using it – find out why and it will rule the world.

P.S. I don’t imagine for a second anyone will actually listen though, so carry on writing that angry comment/email ‘proving’ that I am wrong so we can get back to flaming each other. 

* It’s an example, let’s not go into it.  I also don’t have a problem with the religious, be it FOSS or actual religion, as long as they keep it to themselves.

** If you think I am wrong provide three examples of healthy, critical, debate that isn’t heavily moderated or a flamewar.  I dare you.

*** I am a vegetarian, but I did make a sausage-crust pizza last night, so I am probably not the best person to ask about fancy restaurants anyway. And I am biased.

Anyway that’s not really the point.  I used to be a nearly full-blown freetard a long time ago.  I had the whole ‘hate the man’ thing down (Microsoft) and bought into the whole ‘by developers, for developers, lets cut out the corporate middle man’ movement.  This was before I had even really used Linux – but the concept seemed sound.  And how can a developer not fall for that idea – software utopia, plus never having to deal with anti-user crap like artificial limits and activation again!  But then I actually used it for an extended duration and moved into my ‘the idea is sound, but it just needs more work’ phase.  I believed with the concerted effort of like-minded people and by including designers, UI experts and artists it was only a matter of time.  The third phase was that it was a shame that a great idea was being ruined by a defensive community of non-developer, non-contributing idiots intent on scaring away anyone who didn’t believe it was already perfect.  At least I was partially right that time.

And finally, I am where I am now: It’s a bad idea to the point of being dangerous and it just won’t work anyhow.

Here’s the thing.  Nobody owns GPL’d code.  The code is free.  You are not.  For example you, as a user, have more freedoms when it comes to BSD licensed code – effectively you can do what you want with it, provided you provide credit.  GPL’d code on the other hand has a slew of limitations on what you can and cannot do and attaches a larger burden on you in terms of distribution of changes and source.  The code is more ‘free’ under the GPL.  By GPL’ing code you effectively say ‘nobody owns this, it belongs to itself’.  I can sell BSD code and deny you your ‘fundamental right’ to freely distribute it further.  I can also grant you the same right as the GPL.  The GPL simply ensures you cannot stop anyone from modifying and distributing.

The problem with the above is that it entirely negates ownership.  You cannot ‘own’ free software.  It doesn’t matter if you wrote it – it’s not yours.  As soon as you release software under the GPL you have no more right to it than any of the other 6 billion people on the planet.  It’s why FOSS advocates call closed source (distastefully) ‘slavery’, as closed source software is owned and controlled by someone.

The ramifications of this are obvious – you can’t make money selling free software.  You can sell it, but so can everyone.  If you’ll need to sell 10,000 copies at £100 each to reclaim your investment and the kid in the local computer store will sell them for £5 to anyone that wants one then you’re simply never going to break even.  The whole ‘you can sell free software’ excuse is intellectually dishonest as you have to compete with people with no sunk or running costs.  The best you can do is put up an online tip-jar and rely on effectively begging – and I’ve yet to hear of that working well.

Effectively, what the GPL and the Free Software movement says, is that developers do not deserve to be compensated for their effort.  Unless they can manage to sell the first copy for £1,000,000 then there is no way to ever get paid for the time spent.  It simply can’t be done.  If you want to man a phone line, do email support, work as a call out technician, then you can (according to ‘software freedom’) demand a fair hourly wage, but if you are a developer you can’t.  It’s ‘immoral’.  Simply on the basis that you cannot own code, thus can’t charge money for it.

The argument often made about the above is that businesses still need programmers.  Which is true.  But this is where the whole thing unravels fully.  I am a programmer – say I have an idea for a great new CMS tool*.  I now have two options:

1. Quit my job, rely on my family to put up with me for a few months while I spend 12hrs+ a day working on it.
2. Take my idea to a large company such as IBM, Oracle, Microsoft or someone else like that.

As above, according to the ‘software freedom’ camp if I picked option 1. as soon as I released it it would be mine no more.  The second I tried to charge money for it it would just be forked and given away.  Instead of relying on the future income I could have gained from selling it at launch to pay back the debts I would have no doubt incurred, and to fund new development, I would be forced to ‘get a job’.  Any development work would have to be done in evenings and weekends – and I would effectively be forced to decide between programming what I want, or my marriage.  Plus the software would progress more slowly as I would not have much time to devote to it.

Now with option 2. there is pretty much zero chance of getting taken up on my idea.  You don’t tell a business that you want to hire you what you are going to work on.  Unless you are really, really famous.  Chances are unless your surname is something like ‘Carmack’ you’re going to be writing the mundane stuff that they want you to do – not your own exciting ideas.

There are companies that are based around and heavily involved in free software.  Names include companies such as Red Hat, Mozilla, IBM, Canonical, Google, Novell etc and the one thing about these companies is that they are not in the software sales business.  Red Hat is in providing support and SLA’s for businesses and servers. IBM is similar to Red Hat – they sell ‘solutions’.  Mozilla makes money from advertising for Google, Canonical is a billionaire’s plaything, Google sells advertising and Novell is just a slightly more pragmatic Red Hat.  You’ll never see a company such as Adobe adopt a FOSS business model as they are in the business of selling software.  I don’t imagine anyone can make the argument that Adobe can go FOSS and remain profitable with a straight face.

It’s the following realisation that made me realise what a disaster FOSS really is.  Free Software only benefits large companies and the rich.  It is almost impossible to be a developer and not work for ‘the man’ under the GPL.  Sure there are exceptions to every rule but the simple fact is you can’t be a developer unless you can get an alternate revenue stream.  Support is good, but not a lot of apps will require support and you don’t spend months programming to be forced to make money manning a phone line.  There are also the dual license options, but this is effectively shareware, and you are still making money selling closed source software.

If adhering to the GPL was a legal requirement then it’s not like software would all suddenly be free and open.  What would actually happen is that people would simply stop making software.  All the games available on Steam would not suddenly be free – they simply wouldn’t have been made in the first place.  I pay money for SmartFTP because it is the best FTP program I have ever used.  By paying money I help fund further development.  I am happy with this.

It’s not even like the mass piracy and commoditization of  music, as it’s not like developers can make money selling tickets at £50 a go to live shows.  The software itself is all you’ve got.

Now I simply don’t see how anyone who thinks pragmatically about a future in which the GPL is accepted as the way to distribute software can possibly support it as in reality the people it hurts the most are the very people that support it – individual developers.

* True story, I do.  I am working on it and plan to launch it in a few months.

Some Further Thoughts and Ugly Truths:

The most overlooked point with regards to software development is that it is generally the result of a few people sinking a large amount of time into it, not as a result of a lot of people doing a small bit.  You simply can’t throw developers at a project and expect it to flourish – it just doesn’t work.  If you want something well written and cohesive it’ll take a small, dedicated, team – not a large bunch of semi-skilled volunteers – not something that really happens in FOSS unless you are being funded by a rich 3rd party.

Capitalism, in a nutshell, is providing people what they want.  If you don’t provide it you don’t succeed.  If enough people don’t want it the provider can either improve and adapt, or die.  In the FOSS world you take what you’re given and have no right to complain.  Gimp is near useless and doesn’t even match the decade old Photoshop 5.0 yet people not using it will not spur on development.  There is simply no motivation to cater for a wide audience – if Gimp was a commercial product the company would have been bankrupt years ago but catering for your users needs is simply not important in the world of free software.  Free software is about developers scratching their own itch, not finding out what itch other people want scratched.

The bulk of FOSS development work is done by one of these people:

• Students, getting paid to learn.
• Unemployed, getting paid by the state.
• The Rich, resting on their laurels.
• Large companies, adding value to their other services.

Free software is categorically not made by these people:

• Software companies.

If it is name one.

Software development is hard.  It’s incredibly time consuming.  If most people truly appreciated the difficulty and time required to create a truly great piece of software I doubt they would have bothered.  It requires people with a true passion for development to make truly great software, and I personally find the fact that such people would be relegated to only doing what they love in their spare time, rather than as an actual job, disgusting.

And finally it is the software that is free, not you.  All this talk of ‘freedoms’ and ‘rights’ is utter crap – it’s not your freedom or your rights so the whole idea that commercial software is taking them is simply untrue.  If a developer wants to release something under a Libre license that is fine – but it should never be considered ‘ethical’ or expected for this to be the case and if you really want to make a case for your own ‘freedoms’ then the BSD license is far more appropriate candidate.

Open Source is simply providing software with the source code so the receiving user can modify it to suit their needs.

Free Software is having the ability to freely modify and (this is the key) distribute software they have received from a 3rd party.

Free Software is being *unable* to restrict what a 3rd party does with your software.

You can very easily have non-free Open Source.  In fact a large amount of software made, mainly custom stuff and most of what I do, is technically ‘Open Source’ – that is I am paid to to do it, and I give the client the source along with any binaries.  It is *not* Free Software though, as the client is unlikely to give it away.

For example a large amount of commercial web apps are Open Source in that you have access to the source code, but are not Free Software as you do not have the right to redistribute the source freely.

This is the main bone of contention between the Commercial and Free Software camps – it has nothing to do with source – which is a red herring – it has to do with control.  The main point of free software is to remove any and all control from the original developers and give it to the users – the software is owned by the community.  Supporters of the commercial model (obviously) have a problem with this as it makes making a business out of developing software impossible – If you sell support you are in the business of support, not software.  Free Software is fundamentally anti developer.

Personally I support Open Source, and plan on releasing my next project as a dual licensed Free Non-Commercial and pay for Commercial offerings.  You can use it for free and give it away if you do not use it commercially, and if you want to use it commercially you will be required to pay a license fee.  Which is perfectly reasonable but it is not ‘Free Software’ and would not be supported by the FSF.

Free Software is an ethical line for it’s supporters.  Non-free (as the example above) is considered immoral by supporters and according to the FSF I am unethical in my request for compensation for commercial usage.  According to the FSF doctrine I must relinquish all ownership over the code and give any user full source and distribution rights.  If I want to make money I should sell ‘support’.

The thing is if (and many people suggest this) I was legally forced to make it ‘Free’ I would not even bother to make it in the first place.  In that world the job of ‘developer’ either would not exist, or would be relegated to a role in a huge corporation where you are told what to work on.  The small, independent software companies would be dead – and the software that they would have produced would not now be free, it just wouldn’t exist in the first place.  So instead of having the choice to pay money for something or not have it, you’d simply not be able to have it at all.  Not a situation I can support.

I think if people stopped thinking Free Software and Open Source are interchangeable terms, and that if more people found out what Free Software really is, and the implications, significantly less people would support it.

Update: I do not think that Free Software is bad, should be banned, or anything like that.  I believe it is up to the developer to choose how to distribute their software and that there is nothing wrong with closed source, commercial software.  It is up to the developer to choose the license, and the user to choose the software.  It’s all about choice.