Ask Slashdot: How Would You Fix the Linux Desktop? 1154
itwbennett writes "Slashdot readers are familiar with the Torvalds/de Icaza slugfest over 'the lack of development in Linux desktop initiatives.' The problem with the Linux desktop boils down to this: We need more applications, and that means making it easier for developers to build them, says Brian Proffitt. 'It's easy to point at solutions like the Linux Standard Base, but that dog won't hunt, possibly because it's not in the commercial vendors' interests to create true cross-distro compatibility. United Linux or a similar consortium probably won't work, for the same reasons,' says Proffitt. So, we put it to the Slashdot community: How would you fix the Linux desktop?"
Fix the Kernel (Score:5, Interesting)
Android (Score:5, Interesting)
By adopting the Android desktop.
Re:It's not broken. (Score:2, Interesting)
as usual, you can not put yourself as evidence that something works (for others)
unless you agree wearing high heels is great. It works just fine for me.
Better is not good enough. (Score:5, Interesting)
The Linux desktop is far better than Windows used to be.
But we already know ways to make every desktop, including OS X, far better than what we have today.
The Humane Interface by Jef Raskin gives good ways to start:
http://en.wikipedia.org/wiki/The_Humane_Interface
One main unified desktop? (Score:5, Interesting)
I know I'll get flamed for this since it goes against the Linux philosophy, but how about getting rid of competing Gnome and KDE (and now Unity) desktops and agree on one standard desktop with a single API for everyone to write to. And maintain backwards compatibility for the API so an application written for GnoKDE 2.0 still still run unaltered on GnoKDE 3.0.
I know that having multiple desktops gives users choice, but there are many talented developers on the KDE, Gnome and Unity teams, and it seems like they could make a much more polished and usable product if they worked together instead of coming out with separate products. Oh, and stop pushing out alpha releases (I'm talking about you, Ubuntu/Unity) as the default desktop and telling users that it's for their own good.
But hey, don't trust me, I use Xfce since it does everything I need in a desktop.
No one buys a computer to use an OS (Score:4, Interesting)
Normal people don't care about the OS, the "desktop environment," the openness of the kernel or its ABI stability. They don't even know what those things mean. People don't use computers for the sake of computers, only nerds do that. People use computers because they do things like write documents or fix vacation photos. If Facebook only worked with Linux, then everyone would use Linux. Writing some killer app and only ever releasing it on Linux is the only way a programmer can get people to switch. Otherwise your best bet is a businessman like Steve Jobs to come along. Look at all the people using iOS. Do you think people are buying iPhones because OMG iOS!!! No.
Re:It's not broken. (Score:5, Interesting)
You're part of the problem.
If you want to help spread the Linux base, such an attitude doesn't help.
If you don't care, then please continue as you are.
A satisfied user doesn't help "spread the Linux base"? Why not, I ask seriously?
Focus and Polish! (Score:5, Interesting)
I love my linux world. The only part which I would appreciate is more polish in the software. Most software has a great set of features but it seems that all these suites are always missing the last 5% of development (e.g. making the application feel very polished).
To me it seems that the only way we can fix the desktop is to throw money at it. The last 5% of development work is usually boring (finding and fixing all the corner cases, etc...). I think that the only true consumer ready desktop right now is Ubuntu (yes, with the Unity interface). It has become a very polished and stable package with a lot of focus (maybe a bit too much?) on the right things. Don't get me wrong, I am a huge KDE fan (I contributed code), but to me it seems that it is missing the last 5% of development work (e.g. Kwin crashes occasionally, the panel wont stick to the top and will sometimes be in the center of the screen, Kwin seems to be slower than compiz...).
Canonical has the resources to provide a really solid desktop experience (and it already does) for most average users. For the rest of us, there is still Arch, Mint, Fedora, etc which allows for more customization. The problem is, that most people want their machine to just "work" and not tinker with the OS to just get it perfect.
Good job Canonical!
Not asking a very good question (Score:4, Interesting)
It is not "We need more applications" -- that is easy enough.
Getting people to create hundreds of (cr)applications for Linux is trivial and is not a solution and may in part be one aspect of the problem.
A somewhat more accurate strawman would be "We need more *good* or *compelling* applications" -- that's challenging. Still only a part of the answer, but closer. It requires answering "What does 'good' or 'compelling' mean in this context?", etc.
Get some -real- UX experts (Score:4, Interesting)
By which I mean, instead of the sort projects have now, that say "I am a ux expert, and I like [insert totally unintuitive feature in the name of "prettiness", or "looking like [Apple|chrome|a phone|whatever]", so that is what it has to look like"... instead the real kind, that goes and does useability tests with a wide range of its potential userbase, and then designs based on that.
Once you have a great product that people actually want to use (and yes, I know Linux is technically the kernel, not its window/file managers/etc., but the UI is what people actually -see-), more people might actually want to use it (I am aware that this is a tautological statement, but shut up.) More people using it = more desire for programs = more better. At least assuming some of those application developers also go the route of doing proper useability testing.
Powershell (Score:2, Interesting)
Not strictly (well, at all) GUI related, but I'd love to see something more like Windows Powershell [wikipedia.org] in Linux.
I love the Linux pipeline and being able to pipe text streams between tools is very powerful, but the more I get into Windows Powershell, the more I like it. The ability to pass objects through the pipeline and operate on those objects can be much more powerful than processing text streams.
Re:Minor suggestions (Score:2, Interesting)
Hmmm. So reading between the lines here (and I'm just wildly guessing) I think you are saying that the Linux desktop should be OS X.
I use OS X every single working day and although it is pretty good and pretty much gets out of my way, it is not ideal in many respects. Finder, for example, is a pile of steaming junk, but I live with it, rather than installing some 3rd-party solution - because if I did, I'd soon become dependent on it and would be lost if I had to use a Mac that wasn't mine (which I do often).
The OS X desktop also has many usability and consitency issue. The fact you can only resize a window by dragging the bottom-righthand corner is just one example. It's just lame. When I use Windows for a while (and I have to) then it just annoys the fuck out of me when I get back to OS X that I have to locate and reach the bottom-right corner to resize the window - this should have been fixed YEARS ago.
As far as Linux desktops go, KDE and Gnome are not brilliant, and not as good as the Windows/OS X experience, but they are certainly more than usable.
Blaming the desktop software for Linux not being able to run MS Office or Photoshop is just plain silly.
Re:It's not broken. (Score:5, Interesting)
Nonsense.
The main roadblock is that the market has been dominated by a single vendor since long before a single line of the Linux kernel was written. This dominant vendor was nearly able to kill off Apple with an OS that has no GUI and required MANUAL MEMORY MANAGEMENT.
It seems like some people have not been computing long enough to realize just how BAD Microsoft products have been while being an overwhelming force in the industry.
People put up with Microsoft because of it's perceived monopoly and just deal with problems as if they were unavoidable and inevitable. The same goes for companies and 3rd parties.
Some people are under the delusion that magically turning Linux into a Windows clone or a MacOS clone would help anything.
Even real Macs still have trouble getting traction.
Re:Simple (Score:5, Interesting)
I don't think Linus is interested in owning the desktop. Its pretty clear he wants his desktop to be geared towards his workflow as a hard core coder/developer , and that is not the same desktop you would want for ordinary people. That is a key problem with the Linux desktop, the only people that care about it and develop it are hardcore geeks/programmers and the stuff they want is diametrically opposed to what ordinary people want in a desktop. Its seriously got old 10 years ago listening to Linux heads demand the Linux desktop be a few windows with shells in them, or listening to them as they forked and developed 100 different window managers almost none of which gain critical mass and none of which will ordinary people use.
If you want Linux to succeed with the general public my suggestions would be to:
A. Get rid of some of the fragmentation, relgious wars, and wasted time caused by the GNOME vs KDE conflict in particular. I understand why the split happened but its done nothing but damage over the years and its time to stop it or Linux will never succeed on the desktop.
B. You need very well written core API's because everything else flows from those. A good IDE helps too, Eclipse is OK but its not great, Xcode is awesome, DevStudio is pretty good.
C. All apps need to use the same API's so they interoperate and look and feel the same. Constantly writing variations of existing API's. and fragmenting them, is not a wise thing to do on the desktop. A lot of Apple's success can be tied to the fact that Cocoa and Objective C are very well done in a lot of areas, and they make it easy and a joy to develop applications. If its a joy to write apps, more developers will do it and the a quality of the apps for the time spent is consistently higher. Writing apps on Linux by comparison is a poke in the eye with a sharp stick, its painful, inconsistent, there is constant wheel reinventing, everyone does their own thing and it shows in the inconsistent apps that don't interoperate.
D. As much as I hate to say Qt is probably the best API you have but you need to wrest control of it from the people who've been developing it, and stop the major code breaking changes between revisions. The core API's need to develop like Apple develops them, add new things carefully, deprecate old things gradually, and STOP breaking code doing huge somewhat, gratuitous changes. GTK is just not a good API to base a desktop on.
E. Miguel De Icaza needs to be cut out of his position of authority. His track record in recent years, his Microsoft affiliation, his blaming the desktop on Linus recently, has shredded any credibility he had to lead Linux desktop development
F. You have to fix audio and video so they just work like OSX and Windows. This is a steep challenge. The ALSA audio API was a total mistake. An API that contorted, hard to use or write drivers for never should have happened. Linus is partly to blame for that. Getting good audio drivers is a hard problem, everytime a new audio chip comes out you have to start over making drivers for it. Making video work tends to end in a lot issues with patents, proprietary codecs, etc, which isn't easy to solve in open source.
In summary, the chances of Linux happening on the PC desktop are slim. None of these inherent structural flaws are likely to be remedied. Besides which the PC is rapidly starting to fade except for content developers and coders. Everyone else is switching to phones and tablets. Linux is already winning with Android on thosse, and IOS is Unix underneath. Rather than fight a losing battle for Linux on the PC just switch to Android.
SUPPORT is what is missing (Score:3, Interesting)
A successful computer needs SUPPORT from the manufacturer.
The users NEED a place where they can get their problems fixed.
This is why Apple succeeds, despite their prices. They provide clear avenues for help and assistance, both hardware and software.
Linux has no such support from manufacturers. If you put Linux on your computer, they will void your warranty and/or find reasons to avoid dealing with you, if you've installed Linux on their hardware. Their tech support people are not trained to deal with you. You are a total money sink as far as they are concerned, because every support call must be escalated.
OSX is stealing away the desktop by nothing more than basic competence.
When I fire up my Mac and run software updates, I am confident that my system will keep running. When I have Linux on my desktop, software updates are always frightening. Will my wi-fi adapter still work afterward? Will the VMware drivers compile? I've lost many hours of work, backing out linux software updates that trashed my ability to get work done.
For another, every single terminal program on Linux is just crapola compared to OSX terminal. Really even the old-fashioned shell users are much happier on OSX. Try developing on OSX for a few weeks, gnome-terminal seems little better than xterm.
Until these problems are fixed, linux on the desktop is doomed.
Re:What, exactly, is broken? (Score:4, Interesting)
I'm still on Ubuntu 10.04 as my main OS, but have VMs that are 12.04 for some projects - Unity seems fine to me.
Bugs are inevitable, and without specifics I'm not sure what I can possibly say to change your mind.
My point was that I think that if you find a Linux desktop environment you're happy with, then the rough spots are much more likely to be far outside the desktop owner's realm - hardware support and 3rd party software support.
Desktops are much of a muchness these days. The dark days of CDE/Windows 3.1's Program Manager are behind us, and everything's reached a certain base level of usability that meets 99% of people's needs.
People may quibble over details, but generally we're living in a veritable age of plenty.
My mother can use Linux, and does on her netbook. It's far better than Windows XP on that hardware.
All I had to do to get the netbook to a usable state for her was get the wireless network drivers working. And if you'd given her the netbook and a base Windows XP install, she'd have had the same hardware support issue - and more. And with no CD drive, it probably would have been befuddlingly impossible for her.
But it's running, and meets her needs just fine. She's even connected it to other wifi networks on her own when travelling, and rarely asks me for any help - everything on it Just Works(tm) for her needs.
By such a measure, Linux is a success, and getting more successful each year.
Re:It's not broken. (Score:3, Interesting)
I started on 0.95. Came on 13 floppies.
I've also lost my will to upgrade constantly. Look how many people still use Windows XP; Its UI hasn't changed in over 10 years. Why should the Linux desktop have to change every other year? I don't care about social desktop experiences and all that nonsense. I just need my DE to run a few apps and not actively try to annoy me. I'm not going to run it on a mobile device, I'm stretching it across two 24" monitors. I don't need a database running in the background. I want to be able to start apps intuitively, and run them separately. I want to be able to configure it easily.
And then you declare the basic desktop DONE for 3 years or so, and work on apps. Maintain the desktop in terms of bug fixes, and internal reworks and anything else you need to do, but religiously keep interfaces static for 3-10 years. And instead of going all 2nd system on the interface, work on other things.
I think the main problem is that people (especially "volunteers") want to work on the new and shiny stuff. They want to put in new features, because they can, not because they should. There's an urge to make a new something instead of make an old something improve. That's why the newest versions of both KDE and GNOME are terrible. They went mucking around because they could, and never asked themselves if they should. Maybe they are afraid people will stop using your somewhat older-looking but very stable and well-liked DE if you don't constantly add crap to it?
-B
Opposite of my experiences (Score:4, Interesting)
Plug-in a new mouse while you're in a game on Windows 7? It won't work. Works fine in Linux (for any game ran in any way).
WinTV card? I have my choice of apps to use it with in Linux that cut commercials, reencode, etc. In Windows it's WinTV.exe (worthless) or nothing. (and most of those cards work despite being obscure hardware).
You basically just more economically said (Score:5, Interesting)
the same set of things I suggested above. Kudos to you.
I started using Linux in '93 but stopped in 2009 because, frankly, I was exhausted. I had forgotten that in 1993 I started using Linux because it let me do the things that I wanted to do at a cost (free) that significantly beat ($thousands) what was on offer in the Unix world at the time.
In 2009 when KDE took a shit on everyone and news that GNOME was about to do it, too, hit the netwaves, I suddenly realized that the situation had become inverted. Now being a Linux user kept me from doing the things that I wanted to do—not in theory (in theory, everything is possible—hell, you can design and fab out your own damned CPU and architecture and create a platform port for it if you want), but in practice. I was spending 10 percent of my time re-learning every major subsystem in Linux that changed every 6 months to 1 year, and another 20 percent of my time constantly fighting to get apps installed, keep them installed across distro releases, support my slowly evolving hardware (which required upgrading to new distro releases or doing backports by hand), and getting those apps to do the things that commercial apps could do easily.
Linux was no longer saving me many $thousands, since consumer-level OSes were now adequate to my needs and the applications I needed to use were only in the $hundreds camp. The capabilities that I wanted—working multimedia, powerful apps that shared file formats with the rest of the world, set it and forget it tools that I didn't need to build myself and that could manage my data—were right there, on the shelf at affordable prices, in every way that they weren't in 1993.
It was like a light bulb went on over my head—and I suddenly realized that Linux was holding my real career back, rather than enabling it as it had done in the early '90s. Bye-bye, Linux.
The culture of Linux remains the culture of 1993 mid-range computing—but we no longer live in a world in which CS students can't afford the hardware/software they use at school and mainstream OSes can't do the fun stuff. Quite the opposite. It's funny to think back at how thrilled I was to have X11 on the desktop (compared to Windows 3.1) versus how I feel now, twenty years on, comparing KDE or GNOME on Fedora or Ubuntu to OS X 10.8. The tables have been exactly turned. Linux is still essentially the same in architecture and philosophy, while the rest of the world has moved to a completely different paradigm in which computing is essentially appliance-driven. In 1993 Linux was ahead of its time. In 2013 Linux is a decade behind.
These days, I want an complete, polished, turnkey appliance at low cost and with no labor time investment, not a set of building block. Today's appliances are fast, intuitive, stable, durable, powerful, and integrated like the iPad (which I do, yes, use for serious work about 5-6 hours a day). For most users (which is where I have always ultimately fallen), Linux is solution in search of a problem that no longer exists.
Re:Powershell (Score:2, Interesting)
PowerShell is interesting, but seems like a classic Microsoft solution.
By which I mean that Microsoft has some fantastic developers, some great minds - but they work in a very monolithic way. They set out their paradigm (ugh, I feel dirty for using that word even when it's appropriate!) and they stick to it.
So when I'm using PowerShell to administer a Windows environment, it's great. It works really well. For example, the other day I wanted to get statistics for some Exchange mailboxes. With a few pipes of mailbox objects into Where-Object for filtering, and a bit of sorting, and finally outputting to ft, I had a nice little report.
But the reason that the word paradigm is validly used is that the moment you leave their use cases, you're screwed. PowerShell is a very nice walled garden for administering Microsoft products, but for doing anything outside of that it's pretty poor. (So for example if the statistics I wanted aren't served by their cmdlets/objects, I'm straight back to writing a program rather than a script for it.)
This is why I keep UnxUtils - a Win32 port of essential GNU utilities - on my machine at work. Whilst I administer no *NIX boxes at all in my job, and am technically not a *NIX administrator or even close, having grep/sed/awk saves the day a lot more often than PowerShell ever does.
PowerShell is great for day to day admin of the Windows-based products, but the moment I find myself with a non-Microsoft logfile or similar, PowerShell is no longer the solution.
By comparison, the other day I solved an issue with reporting via a (slightly convoluted) chain of grep->awk->sed->tr. The issue was getting statistics from "XML" report files that a process dumps when it's completed. My other options were to write an actual program using an XML library and VBscript/C#/$whatever.
Or to chain together 30 year old UNIX tools to create a CSV file from (slightly dodgy) XML.
I tell no lie when I say that it took me longer to convince Excel 2010 to produce a usable chart from the data after importing than it did to extract the raw data. It was under two hours to get the data out, and a lot longer struggling with Excel's poor interface for multi-axis charting....
And I'm not a UNIX expert - I had to refer to the O'Reilly sed/awk manual several times.
So in a nutshell, PowerShell is a great environment - but it's designed by programmers for one specific environment, and doesn't work at all once you leave their specifications. By contrast, the UNIX utilities are so broad and useful that people are still applying them to uses that the authors probably never even imagined would be possible, let alone probable.
I'd like a PowerShell style UNIX environment, but I suspect I'd still end up calling grep/sed/awk many times within it precisely because an object isn't going to be as useful as I'd like at that point in time.
This isn't to say that PowerShell isn't very powerful and nice - just to say that until it gets a decent set of tools that allow it to step outside its own paradigm and just work with anything it's given, it will always be a walled garden.
(Albeit a garden of national park proportions.)
Re:Add Support for Visual Studio (Score:5, Interesting)
One thing I will give credit to Microsoft about is their software development tools. It is also useful to remember that Microsoft started out as a compiler developer that happened to end up in the operating system game due to (for them) a fortunate series of events.
I've been using Microsoft products since 1979 in one degree or another, even though I do think they are an evil company sucking the life out of the American computer industry. Still, your comments about the quality of their development tools seems to be pretty spot on with my own experience as well. They use these same tools (Visual Studio) to write MS-Windows, so they get a whole lot of internal attention within the company where it is co-workers complaining about nasty bugs and not just outside customers.
I fell in love with C# because the design team for C# is a bunch of guys that I like and are some of the best compiler/language developers in the world. They were the original developers for Borland Delphi, and if you are familiar with both Delphi and C#, you can find a whole lot of similarities in the language design including underlying philosophies for how they work with data structures. The chief architect of both languages was Anders Hejlsberg [wikipedia.org], somebody who I have come to admire. What I also like is that Microsoft pretty much let him do what he wanted, and C# has become a pretty successful language on its own merits.
a few things that would help (Score:4, Interesting)
1 every X releases have some sort of LTS release and then fix everything DON'T EXPERIMENT WITH NEW BETA STUFF
and have it work as much as possible (so no switching to some new version unless EVERYTHING is working in that version)
then use the run up to the next LTS to get everything ready
2 if you are in the Thou Shalt Not Use the Root Login EVER camp then eliminate every time something needs ROOT you can
(set automount to mount portable drives READ WRITE WORLD not Read Only Root Owned (unless that is set in the drives meta data))
3 even if its a Python clicky shell eliminate the need to go to the Command Line Shell as much as possible
4 have the desktop "control panel" slurp any generic control panels it can
5 God help me for this one: Have in the Help system some sort of Clippy type Agent that you can use to invoke "wizard" type things for the times you have to do stuff across multiple control panel items
6 any person that uses JFGI in your distro forum WITHOUT GIVING A CORRECT SEARCH PHRASE should be banned and if an employee does so that person should be fired.
7 and finally INSTALL A LOCAL COPY OF THE HELP FILE WITH ALL APPLICATIONS never assume that your user currently has a network connection