|Home » Industry Watch » The Technological
Some people never learn; some don't want to learn; some want to learn the wrong things.
Mark Shuttleworth is at it again. Speaking at the O'Reilly do in scenic Portland he called on open sauce developers to create a Linux GUI that surpasses Apple's.
Of course the fanboys will gladly rally around anyone who defends their precious computers as the best in the world but this isn't that type of article. Heaven forbid.
'The great task in front of us over the next two years is to lift the experience of the Linux desktop from something that is stable and robust and not so pretty into something that is art', said Shuttleworth on 22 July.
'Can we not only emulate but can we blow right past Apple?'
You go Mark. Sorry but you're being really stupid now. Darryl Taft of eWeek comments.
'He made no mention of whether Apple intends to simply sit idly by while desktop Linux catches up to and surpasses the user experience that Apple has become so well known for.'
Some comments from the two Digg threads.
Dude, I love Linux and open source. The Kernel, Apache, various scripting languages (PERL, Python, PHP, Ruby, etc..), Mozilla, are all great open source tools to work with. But seriously, there is only so much work you can rely on people to do for free. If you want the eye candy it's time to open up the wallet.
Its not a pretty theme that attracts me, its what I can and cannot do with the OS.
One of the biggest problems when Linux UI developers try to compete with Apple is that they somehow think that the battle is about what looks better. It's about usability for normal users. Looking good is something that comes along for the ride when you do good design. As long as UI designers for Linux (and Microsoft) think that they can just glue a pretty layer on top of the way things are already done, Mac OS X is going to continue to be seen as 'prettier.' It's the difference between doing a unified project with artistic integrity (a term that most gearheads would laugh at, since they don't understand it) and doing something to just put on a layer of paint to cover up problems. It's VERY possible to compete with Apple on UI, but the problem is that the people in the Linux and Microsoft camps don't seem to understand (deep down) that the process starts at the core, not just with what the user sees.
I'm not a total n00b and I had trouble figuring out at first how to use the package browser in Ubuntu. Nevermind if I didn't understand the concept of packages... I would *never* look at something called 'Synaptic' to install other applications.
How difficult would it be for a company to make an OS similar to Leopard yet be able to run on the variety of machine configurations that linux and windows can? Also why has it not been done? I'd like the stability of unix/leopard but still being able to run windows and windows apps like you can on a mac with ease compared to using wine.
Ubuntu hippies somehow think they can change the world, right? Wrong.
Summer of 2006
Two years ago a number of geeks and programmers at the CLIX Exchange decided it would be a good idea to take up the OpenStep challenge: produce an open source GUI based on NeXTSTEP that would run on Linux. The project took the name 'NeXTbuntu'.
Thirty or forty people were involved: three programmers, a few admins, a project manager, and over thirty others who said they'd be willing to 'test'.
The object wasn't to write the whole thing from scratch: the idea was to help disparate forces come to agreement and produce something substantial.
GNUstep is the oldest GUI for Linux but it's also the one farthest behind. Whilst first KDE and then GNOME made it out the door the 'developers' at GNUstep - who basically inherited their code from Paul Kunz before he went over to Windows - are still moaning that it's so much work to make a finished system and that they are working very hard at it.
A recent effort called Etoile attempted to take the GNUstep code and make something substantial; however the prescribed programming standards are so ridiculous not many want to contribute.
The CLIX Exchange crew attempted to contact Mark Shuttleworth. 'Give us two years and $200 K', they wrote, 'and we'll put something together for you. For your current GUI bites aardvark testicles.' No answer - but a few months later Shuttleworth was in the headlines again, this time singing the praises of RMS.
Not a Clue
Shuttleworth might be stinking rich and he might have been a bit of a programmer and he might be a great humanitarian too and he's certainly to be thanked for making his Ubuntu distro free to your mailbox but when it comes to human computer interaction he simply doesn't get it. In fact few of those in open sauce have a clue.
They're still trying to shamelessly copy Windows. Right down to copying the same blundering mistakes Microsoft made when they tried to rip off the Mac. Right down to the obnoxious menu on the window mistake. Right down to the bloody INI files.
It's not about how it looks - it's about what it does. NeXTSTEP uses a floating point screen coordinate system; vector graphics; a fully - a brilliant - object oriented API. People might gawk at how fantastic the iPhone is but what none of those punters realise is that this is possible solely due to the fantastic 'GUI' NeXT and later Apple have had all along.
To discuss how the NS classes work in an article like this is way beyond scope but suffice it to say there are literally innumerable examples of how and where the NeXT system leaves both Microsoft and the people in open sauce in the dust.
And that was over twenty years ago. Skip to 2008 and see how the distance to the also-rans increases.
Not a Chance
Shuttleworth and his minions don't have a chance unless they make major changes. Here are some of the changes they'd have to make.
- Don't ever forget: you will never beat Apple.
- If you have to copy then copy from Apple. Shamelessly.
- Get good graphics artists. Programmers can't draw for shit.
- Scrap GNOME and KDE both. They truly bite aardvark testicles.
- Stop worshipping Richard M Stallman. The waste of space's a total loser.
- Stop copying Microsoft. Microsoft's way of doing things sucks. It always has.
- Get some good OO code with a good language like Objective-C. Borrow from GNUstep.
- Make sure your developers are experts at OO both from the developer's and the user's perspective. And C++ and C# are not OO. Make sure they know Objective-C in their sleep.
- Stop thinking it's about how 'good' it looks. If you build it right you can make it look like anything you want. Start hitting yourself with the clue hammer: it's about how 'good' it WORKS.