Year of the Penguin - el Reg's 2011 Linux-land roundup It's been a tumultuous year for the Linux desktop. Anno domini 2011 saw the release of not one, but two major new desktops, the GNOME project's GNOME 3 shell and Ubuntu's rival Unity desktop. By the time most distros hit their stride in 2011, the GNOME 2.x line had been replaced with GNOME 3. With change comes angst - …
"Activity" and "Activity oriented workspace" are sheets of the _ANDROID_ songbook, not Apple's. Both KDE4 and Gnome 3 are such a trainwreck because of copying _THAT_ obsession without copying any of the state management and workstack management which goes with it. iOS is still application based, not Activity + Intent based.
Just read the frigging Android and iOS developers manuals for crying out loud, look at the changes in KDE 3-4 and iOS and it becomes immediately clear who copied whom.
25 Up votes?
Go on then, I'll bite, name me another distro that is relevant on the desktop?
Isn't this sweet, a distro goes hell for leather for mainstream and the freetards around here just gun it down.
Don't worry fella, I'm sure 2012 is the year of the Linux desktop but obviously we don't want it to be a popular one!
I've been a perfectly satisfied Linux user for a very long time. But the desktop developments are beginning to annoy me.
I'm X fan, and have often used the redirection abilities to remote displays for navigation and science systems. I got on perfectly well with the basic windows managers.
Now we have fantastically elaborate desktops, dbus, half a dozen sound managers, and hardware integration is no better. Perhaps the effort has gone in the wrong place?
Half of me thinks that the changes for Gnome3 and the new KDE are good, because innovation is good. But you tell me now that they are copying apple, not breaking new ground? All I know is that things get more complex, harder to understand, but not /better/. I feel much the same about Unity: change is often good, but it isn't working out.
I like the E17 desktop, and get on with that. I like LKCFE or whatever it is called, I use that on a Suse netbook.
Laptops have been getting some really dismal screens lately (768, 800 vertical pixels in many cases) and if people need to rip up and start again, why not change to a left hand menu to restore much needed vertical screen space (I know about Unity)?
But really I feel that the Desktop problem is largely solved until someone comes up with a dramatically different paradigm (like merging the desktop screen and two or three tablets into a single environment). Can we not get some of this astonishingly skilled development effort into finishing a few applications instead? Fix the bugs in Libreoffice/openoffice. Finish at least one of the CAD projects. Why do we have 6 desktops and only one Gimp? Skype has stolen the public space that GPL started. Voice recognition? Access for the blind?
Can we fix all the scruffiness about power up/hibernate/sleep/restore?
How about fixing the printer problem with something similar to ndis netowrk driver handling? get all those dodgy lexmark printers working just using their windows driver disk?
Stop re-inventing the front offside wheel. Fix the other wheels as well.
but I have 6 machines running OpenSuse 11.4, a netbook, a laptop, 2 workstations, a fileserver and a desktop in our holiday home.
I don't have a problem with KDE although like you I'm quite happy with more 'primitve' managers
I don't have a problem with sleep mode
I have two printers - a Samsung laser and and Epson scanner/copier/inkjet which both work perfectly and network nicely from the fileserver.
I do agree about GIMP but I often process RAW photos with showFoto
I don't use Windows at all and expect to do everything I want in Linux - these days I find this quite easy.
Incidently I'm thinking of building a compute server to offload a lot of intensive scientific calculations & modeling and also for the transcoding/rendering of 1080/50p video
> Stop re-inventing the front offside wheel. Fix the other wheels as well.
Are you willing to provide the funds?
There are basically two kinds of developers in the Linux arena: those who get paid to work on it, and those who are self-funding.
The former generally work on specific projects assigned to them by their employers, who expect to benefit from it within the context of their own business plans--this is mostly not consumer oriented stuff, unless you count Android.
The latter, which apparently is most of us, work on whatever we feel like and are able to tackle within the constraints of our own knowledge, competence, and motivation. Projects with a positive and welcoming "feel" to them also tend to attract more contributions--this is why I for example have contributed to the KDE4 code base and ecosystem, even though I definitely do not consider myself a desktop developer, nor is something that I'm attracted to; it's just an opportunistic reaction.
Personally, I feel you have every right to a) use the stuff I create or contribute to (in the FOSS sense, including profiting from it), and b) complain if something I have done doesn't work as it should or is a piece of rubbish in any other way.
On the other hand, unless you're willing to dig into your pockets, I don't see how you can tell me what I or anyone else should be working on. Either pay for it and have it the way you want it (or code it yourself if you are able), or go with one of the various commercial off-the-shelf offerings from Microsoft, Apple, or whoever else if they already provide what you need. They do some good stuff too.
I'm so sick of people saying that! There is no way in hell, I'm ever going to switch completely to a laptop/netbook/tablet and use only web apps, it's just NOT going to happen, so stop trying to tell me I will!
We have a tablet in the house, my partner loves it, but for her graphic design work, she still returns to her desktop. I have both a netbook and a laptop and a Galaxy S, but I do 98% of my daily work on my desktop, that's not likely to change, ever!
I'm with Robert E A Harvey, why does hibernate still not work properly, why do not all printers just plug and play the way they should? Linux has SO much going for it in desktop, which is still just as important as it ever bloody was, and yet it still falls over with simple tasks like being able to plug in a cheap printer or run 2 versions of an app in different desktops, WTF people! Stop whinging about the bloody web and get on with the desktop for crying out loud! Let android deal with the web, it's doing a damn fine job, Ubuntu is a Desktop OS!
"why does hibernate still not work properly"
There's a lot of hardware out there, a lot of design issues.
Hibernate works on my Acer Aspire One, Opensuse 12.1 (it did on the 11s too). An MSI windbook 64bit is a bit more touchy, will be rolling it back to 11.4 but that's because I have the choice.
"it still falls over with simple tasks like being able to plug in a cheap printer"
I use HP, have done for years, on my third having started with 640C (?). Lexmark don't want to support Linux then don't buy their printers. Network based printing will solve all that over time.
"or run 2 versions of an app in different desktops"
bit difficult to argue that generality - does that never happen with Apple? It does with Windows and the only solution is to spend money.
With Linux, possibly we're all in this together and by using feedback to the devs we can reduce the problems.
Hibernate doesn't work on my Aspire One, but that's because I've got no swap space set up.
However, I chose to stick with LXDE for the window manager on it, responsive and perfectly adequate for the small screen. I hope I am never forced to use Unity, and I wasn't too impressed with Gnome 3, but then I always preferred KDE over Gnome anyway.
Developers working on Gnome, KDE and such mostly do not intersect with those that work on the kernel and drivers. Otherwise, Linus would fork whatever he wanted all by himself.
Hibernate work on all of the computers I have seen so far, however with a strange GRUB_CMDLINE_LINUX="resume=/dev/sdax" addition in the grub config. Suspend-to-ram is a little more problematic when kms/dr on i sometimes... I haven't seen a non-working printer for a long time. Like other people said, do not buy from manufacturers that do not support GNU Linux.
My cheapo brother hl-2230 works perfectly and is controlled by CUPS server on Debian. All Ubuntu machines can access it through the route and their CUPS. I heard that this is harder to implement on the Windblows machine. Moreover, almost everyday I find a taken-for-granted feature on Linux to be non-existent or more complex to get on Windows. A fiend asked to salvage data from her burnt laptop said that someone couldn't get her hdd to connect via sata-to-usb to a Windie machine. It got mounted for me right away ...
In general, both MS Windows and even Mac OS X are both so very lame compared to free and open OS's, even with all the "unholy mess" that we have in DE right now.
Since most distros use CUPS (pretty much Apple's only useful contribution to this world), it's usually safe to say that if the printer is advertised to work with OS X, it'll work in Linux. Although this of course is not always the case, it's a pretty safe bet.
Samsung was always good about putting a penguin on their laser printers, and whatever you do DO NOT buy from Lexmark. They have always been a tool of Microsoft, and although a select few models do work in Linux, they involve some sort of badly written Java-based [shudders] driver and installer. Epson is a good one - in fact most of their modern all in ones, like the Artisan 810 that I have actually use embedded Linux behind the scenes. I've yet to find an HP that didn't work (aside from hardware issues).
Printers are practically the easiest Linux hardware to buy as far as compatibility. Just for god's sake don't buy Lexmark, no matter how cheap they are.
I almost completely agree with you, except that we should not be thankful to Apple for CUPS. It was only 2007 when ESP was bought by Apple. They've been using CUPS since 2002. We should be thankful to Michael Sweet though and all the developers that contribute to the project (including Apple, HP and Brother are even more) .
I agree with you , Robert. It is amazing how many devices out there are not advertised to work flawlessly under GNU/Linux when they do, including the mentioned printer.The installation through cups (+some front-end) is pretty straightforward, except for the suggested driver might not be the best, so a little tweaking is needed. Like in the case with hl-2230, older version of CUPS did not know which driver was best for it. The newer found it right away. The manufacturer's poor job...
I find it just as irritating to be told by a manufacturer that some device 'only' works with Windows 'X' or maybe Apple and to find that actually it works perfectly well or better with Linux.
Recent examples include a USB 3G dongle, a firewire video camera and a USB/serial converter
>Developers working on Gnome, KDE and such mostly do not intersect with
>those that work on the kernel and drivers.
Except, perhaps, in organizations where they are paid to make linux work.
Yes, Mr Cannonical. I'd have far rather you came up with a Linux version of TWAIN that will work inside all graphical and photo apps, than bugger about with the desktop that isn't broken.
Or stop pitching and rolling from one audio driver, patch panel, and music app to another like a drunken sailor who has the price of one bottle of rum and is equidistant between three of them.
A Linux version of TWAIN????
Are you in any way familiar with TWAIN's architecture? I thought even Windows had got rid of it years ago (but I wouldn't know about that).
I presume you know about SANE. What's wrong with it? And more importantly, how are you planning to fix it, or you just feel like having a whinge about a product that is made available to you in an incredibly permissive way and at no charge, should you wish to use it.
Or scanners.
My wife bought me a nice cannon scanner that only works in Windoze, because the one in my HP all-in-one is so bleeding slow.
Or stupid android phones that don't use USB bulk storage, but expect MCP instead
Webcams.
TV tuners.
Sat navs.
> Or scanners.
It's a long time since I've had a scanner that didn't work under Linux.
My old Mustek stopped working under Windows (apparently the cable is broken), but works just fine in Linux.
> stupid android phones that don't use USB bulk storage
Haven't seen one of those - which ones are you talking about?
> Webcams.
I've yet to acquire one that doesn't work.
> TV tuners
I had a problem with a Realtek chipset in a tuner. I asked them for a datasheet so I could write a driver - they sent me a driver instead. Then they said I could release it under GPL.
> Sat navs
The only sat navs I've got access to just give me a filesystem. Which ones are you trying to talk to?
Vic.
If you felt a bit brave and had some spare time, ask 100 strangers two questions:
1. Do you know what Linux is?
2. Do you know what Ubuntu is?
Perhaps shift the demographic in favour of the computer literate and ask the the same question to strangers in say, PC World or an Apple retail outlet.
I feel certain that right now, you'd be lucky to find a single person who knew what Linux or Ubuntu was.
However, go back a few years and ask "Do you know what Android is?"
You would probably get a reply along the lines of "It's a movie about robots?"
These days, most people would know it has something to do with mobile phones - you'd probably get a few people saying "Oh yeah, that's HTC?" and more savvy people "That's google, right?"
Where I'm going with this is that it's taken Google with all their billions and eyes on computer screens to realise Android and get it to market.
Ask a further question now:
1. What is Google's mobile operating system based on?
How many people would say 'Linux' ?
Very few.
For Ubuntu to have any chance in the Tablet market, they will need to increase market awareness massively.
They've made a tiny dent in the Desktop market amongst the geek set, which was no mean feat, but it's taken them 8 years to get to this point and still the total Linux desktop market share is tiny - figures vary, but at most it's 3%
Canonical are right to pursue the mobile market, but to do this effectively, they need brand awareness and that's going to cost a heck of a lot of cash.
In the interim, Apple, Google & Microsoft aren't exactly going to be standing still and with massive war chests they'll drown out any noise that Canonical may make.
If I were a betting man, I wouldn't put any money on Ubuntu getting any significant share of the mobile market, unless they sold out to Google.
More accurately some one asked me. Working in library services she asked me if a new e-book service she was testing worked on Linux (not Android), as she was concerned about universal access. (No, it didn't).
She didn't know that Android was Linux based. I used the growth of content consumption devices based on Android as a reason to pressurise the supplier - who replied with the gem "he didn't think that Linux for Android had been released yet"
I don't know the outcome of that one, yet, however this isn't a race, it's erosion. Ballmer was right.
Once market share gets to a level much higher than it should need to be (15%?) universal access to online public services will create political pressure to level the playing field a bit more, thus removing one more barrier. Possibly Android for tablets will enable LOTD as it becomes more possible to refurbish old hardware for general use.
Well, when someone really thinks that a big non-touch monitor needs to run the same interface as a small touch one, I think it's time to say "sure, go on with this madness" and promptly choose another distro or another desktop manager.
I like the Android touch-friendly interface on my phone and on my tablet, but I DON'T WANT the same interface on my 28 inch non-touch monitor.
Interface designers in commercial products think that users are stupid, suffer from attention disorder, and cannot focus on more that one simple task on one big window that covers all of the screen, with no more that two big buttons at a time. Everything more complex is absolutely too hard to use. And while there are smarter users in the commercial software world, there are also a lot of brain damaged users.
But, if we keep helping the brain damaged users, sooner or later the smart users will die of boredom.
Have you seen the movie "Idiocracy"? It is a perfect example of where we are headed.
Windows is at least going in the right direction here with the "clip to the edge of the screen" thing.
Ok, I hate the implementation and want to find how to turn it off on my laptop, but at least they are thinking about things like this.
Likewise, full screen apps on Lion - probably good for 10" and 13" laptops... not so great for 27" displays.
Size does change behaviour and one OS behaviour does not meet the needs of a 19" and a 27" screen user. Come on, Linux desktops are renowned for giving options. All we need is a bit of logic to test screen size and a manual override button.
respective quotes
"I like the Android touch-friendly interface on my phone and on my tablet, but I DON'T WANT the same interface on my 28 inch non-touch monitor."
"Likewise, full screen apps on Lion - probably good for 10" and 13" laptops... not so great for 27" displays."
Yeah. On my 28" (actually dual 28 inchers which makes the interface issues stand out even further) Unity totally falls apart. Windows 8 doesn't do very well either, and separately I've been subjected to what the modified "Lisa" interface does with larger screens with multiple windows. But for the dual 28's, the modern/stable UI's of KDE 4.7 (now 4.8beta for fun) and Win7 work fine.
For a 15.6" laptop monitor or smaller screens on netbooks and such with 1366*768 max resolution, then Unity and Lisa (Mac) UI's work. But get them on large screens, and they're unfriendly. The mouse gets tired running ALL the way across the screen for a menu option that isn't attached to the window in the lower right. :) Then Win8 isn't friendly on anything without a touchscreen...
What I've liked on KDE is it does detect when it's on a 10" netbook, and adjusts accordingly to the "netbook" interface which saves real estate, and also switches to the "globalmenu". On a normal desktop, it is a fully equipped desktop. I don't see why that strategy isn't adopted by MS, Canonical, Apple, etc etc. Why have exactly the same UI across all the netbook/laptop/desktop computers?
"Interface designers in commercial products think that users are stupid, suffer from attention disorder, and cannot focus on more that one simple task on one big window that covers all of the screen, with no more that two big buttons at a time."
Only ONE big button, thanks. OFF!!
"Not everyone wants to relearn how to use their computer just so Canonical's designers can show off how they think the desktop ought to work."
Summed it up nicely, IMO.
Very glad there's a fork of Windowmaker (windowmaker-crm) now in active development again so there's no reason to switch to Gnome or KDE in the foreseeable future.
It seems to me the Linux crowd need to decide if they want a desktop THEY can use, or one which will make Linux popular to regular PC users.
Linux users are often quite strong evangelists of Linux but they seem to fail to realise regular folk NEED an easy interface, along the lines of Win/Mac. However if they make an 'open source Windows/Mac' shell that _would_ actually attract those they evengelise to, they could end up with something they don't themselves like.
So do they want Linux to go mass-market, or do they want to keep it for those who understand computers? The former seems an obvious answer, but if they're not in it for the money then perhaps the latter is better.
Why should we choose between a mass market interface *or* a power interface? Do hot rod enthusiasts decide between a tweaked out vehicle *or* a vehicle their mom can drive, or do they enjoy a world where both are well-supported?
The point of FOSS and Linux to me is *choice*! I want total power over my computing experience, and an app-compatible mass market platform for the non-geeks I love (and support).
I borrowed a Mac Mini a few weeks ago to see what the fuss was about.
I was expecting a life changing experience after all the hype.
What a disappointment. It just felt clunky and looked clunky. I'm a long term Windows user and after using Win7, OSX just felt awful.
I even prefer using the Ubuntu 10.4 Netbook edition on my Acer Aspire One netbook for getting stuff done.
> I borrowed a Mac Mini a few weeks ago to see what the fuss was about.
Think yourself lucky. I bought one for a job.
> I was expecting a life changing experience after all the hype.
As was I.
> What a disappointment. It just felt clunky and looked clunky.
I don't know about "clunky", but there were definitely a number of things that didn't "just work". And finding the fix was less than easy :-(
> I'm a long term Windows user and after using Win7, OSX just felt awful.
I'm a long-term Fedora user. I now have three Macs (for various work projects), but I would not consider changing my personal machines for Apples.
Vic.
... as long as one has a proper shell and perhaps a few GUI applications like Browsers.
What would be interesting, instead of re-inventing the same wheel, to find a way to actually have _functional_ graphical user interfaces. Interfaces which are more than just forms and switchboards. Where I can draw a command, and the computer executes it and gives me the result... just like I type a command, and the computer will execute it.
> Where I can draw a command, and the computer executes it and gives me the result
Visual programming paradigms have been around for donkeys' years. I last used one in anger in the mid-90s.
Sadly, they don't really work for long. If you have the richness of interface that a typical command achieves, you have a very complex interface to try to support that richness. It very quickly becomes much easier to learn the "traditional" text-based interface than to try to work through the mappings from that interface to a visual paradigm.
So the visual thing gets you through the first few weeks of development, as you start up the learning curve, but soon becomes an impediment to progress, not an aid.
Vic.
... so nobody ever will.
Which makes absolutely no sense—especially in the IT industry, where Apple has proved *all* the mobile phone industry's veterans flat-out wrong about how touch-screen interfaces *should* be designed.
The reason most visual programming tools suck is because they've been designed by programmers. Most programmers have all the graphic and interface design skills—and you really do need such skills for a _visual_ tool—of a used sheet of toilet tissue, so expecting them to get something like this right, is like expecting a congenitally blind child to paint the Mona Lisa.
A programmer will only ever create a visual interface to a traditional programming paradigm. The result is just a bunch of visual stand-ins for APIs that you wire up like a circuit, or slot together like a bizarre version of Tetris. Neither is a particularly good representation of visual programming because both rely on a fundamentally flawed assumption: that a visual tool must simply provide a like-for-like representation of the underlying, traditional, written-code-centric API.
Before you can create a visual programming tool, you have to first redefine "programming" to suit the new medium. That means hiring *non-programmers*, and having them lead the team. Programmers already *know* how to code; they cannot possibly be expected to approach the problem with fresh eyes.
> ... so nobody ever will.
No.
My opinion is that interfaces are inherently wide - either a small number of classes[1] with a lot of options, or a large number of classes with fewer options.
To create a visual paradigm for this, you either have a vast amount of options on each item, or you have a very large pallette of items from which to select. Either of these situations ends up with the user needing to know so much about the API that he might as well just be coding to it.
It's the same old GUI/CLI argument: the reason we developed speech, rather than just pointing and gesturing at images, is that speech is very much more expressive. You can convey an *accurate* message with far less effort.
And that is the end-point of programming; the language doesn't matter, it's all about describing the solution to the problem space. The more expressive a coder can be, the more effectively he can fulfil that task. And so a GUI is excellent when getting to know a system, but the CLI becomes the tool of choice when the user is more experienced with its capabilities.
Vic.
[1] Choose your term here; "class" is appropriate, but I'm not going to get into a semantic argument about how to term a collection of programming elements.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
