Novell's Releases Linux Usability Testing Videos 429
sp3298622 writes "Novell is releasing primary desktop research, including over 200 videos and analysis of usability tests, at betterdesktop.openSUSE.org. Vice president of collaboration and desktop engineering for Novell, Nat Friedman: As a programmer, it's sometimes difficult to know how ordinary people with no technical experience are reacting to your software. Linux people tend to know other Linux people. In these usability tests, we selected test subjects who were experienced with Windows, but who had never heard of Linux, and asked them to perform basic tasks using the Linux desktop."
Batten down the hatches (Score:5, Funny)
Re:Batten down the hatches (Score:2)
emerge nvidia
emerge nvidia-glx
emerge quake
exit;
See...easy...
(please note, this is a joke)
Re:Batten down the hatches (Score:3, Funny)
Re:Batten down the hatches (Quake Install Troll) (Score:4, Funny)
Installation was pretty smooth. I had to download nvidia binary drivers to get fully accelerated OpenGL, just like Linux. Windows is a supported platform for the drivers. I had to reboot the whole OS after installing them, because Windows won't let you easily drop back to a command line mode and just restart the GUI. No worries - I didn't have a server running on the machine, and it only takes a bit longer to reboot than to just restart a GUI.
Caution - Windows only comes with a special limited feature browser that doesn't support tabs, or anything. It is apparently only provided so you can download the latest version of a real browser after you install Windows. Windows doesn't come with a lot of useful stuff that you expect from a Linux distro...
So, I start reading docs to find out how you install apps on this new OS. I was having a pretty good time. Then, I learned that there is no equivalent of apt-get. If there is free software you want to download and install, you have to do it manually. So, I used the funny miniature "IE" browser to get the Quake source online.
Ooops, bad idea. Windows doesn't come with a compiler. You can download a free version, but the full featured "Visual Studio," costs a lot of money. I didn't feel like investing the effort to understand the differences. I decided to just get binaries. Again, there is no tool to automatically download and install an app, so I had to manually google for windows binaries. Thankfully, Quake is a very popular game, so it was very fast and easy to find, but still, it is an extra layer of inconvenience.
After a flurry of clicking "next" and eventually "finish," I finally had the game installed. Hooray. I tried to run it and I got a "BSOD." (Crash error screen) Of course, I already pointed out that Windows comes with no development tools, so it wasn't like I could try again with the debugger to see what happened. I had no way to see exactly what the issue was. What's worse, I couldn't get back to the system. This *game* had caused the equivalent of a kernel panic. It wasn't just the app that had crashed, but the whole system! this, from a system that is supposedly really great for games! It lets a game kill it!
Okay, so I rebooted into Linux. I already knew of a website with binaries for Quake, so I went there in Konq (Which came installed by default! I didn't have to go and download it!), downloaded a package...
dpkg -i Q
That was all there was to it. This "Windows is great for games" garbage is just horrible propaganda.
Now, if only I could get sound to work in Linux...
Re:Batten down the hatches (Quake Install Troll) (Score:3, Insightful)
Ok, no offence and no, it was no joke. I suupose you got some of XPSP2 boxes. If you not, well, then you got very well protected with OEM default installed firewall or something else.
In fact, for at least two hours you should spend of downloading security patches alone, if you don't have one of these things by default.
Re:Batten down the hatches (Quake Install Troll) (Score:4, Insightful)
You followed the intallation instructions to a "T" but the card didn't fit any of your expansion slots. You went back to the store to get a different one, but no one was knowledgable enough to help you out. Finally some kid in the aisle overheard you and explained about AGP and PCI-Express. He steered you to the right card.
After following all the instructions you finally get your game set up, but the graphics look crappy. You complain that your $200.00 card isn't even as good as your PS2. You enjoy bad graphics until Xmas a half a year later when your nephew explains the concept of "Native Resolution." You love your gaming PC now and just think, it only took you half a year to get it right!
TW
P.S. I'm a frequent Windows PC gamer, but I don't have any illusions it's as easy as you make it. Newbies have a steep learning curve.
Re:Batten down the hatches (Quake Install Troll) (Score:5, Funny)
-1, Troll.
Users can't understand? (Score:3, Informative)
Cryptic? Maybe. But so would
also check... (Score:3, Interesting)
<br>
"The Tango Desktop Project exists to create a consistent user experience for free and Open Source software with graphical user interfaces."
Re:also check... (Score:2)
Re:also check... (Score:3, Funny)
Please proofread. (Score:5, Informative)
Comment removed (Score:4, Funny)
Hm. (Score:5, Interesting)
Might this only result in the Linux desktop becoming more like Windows?
Mod parent +25. (Score:5, Insightful)
Then put them in front of a different system (like say a Mac) and see if they have any problems performing that same function.
Of course the "easiest" (and therefore the "best") user interface will be the one that is as close to 100% identical to the only one they've used before.
That's great for Novell because they're trying to get a slice of the Windows market.
But this does not provide ANY information that any person could not get just by spending 10 minutes on a Windows machine and copying down menu locations and order and wording.
Re:Mod parent +25. (Score:5, Insightful)
Of course the "easiest" (and therefore the "best") user interface will be the one that is as close to 100% identical to the only one they've used before.
That's false. The best interface is one that reflects the user expectations. The Windows interface doesn't reflect user expectations in many ways, so it's possible to create a better interface than one which is just identical.
That should be the aim for the Linux Desktop design, not just to attract former Windows users but to best serve previous Linux users as well.
Not with the testing they're doing. (Score:5, Insightful)
You are correct, in theory. You are incorrect in this specific instance because their testing procedure will not yield the information necessary to find a "better" interface.
That is because they are only testing prior Windows users.
Those Windows users have been trained to seek certain items in certain places.
Even if you added a button that said "Complete this test with one click", the users would NOT find it unless they could not FIRST find the Windows button/menu that they were trained to look for of if that button was in that location.Again, I agree with that, but that will not be achievable through these tests.
Microsoft Word used to have an option to use the WordPerfect keystrokes. This was because the people with the most experience found it very difficult to maintain their productivity while learning a new system. Even if that system was "better" for other people. Back then, the most experienced and productive people had spent years learning WordPerfect for DOS.
Novell has learned nothing in these past years. To make it easy to migrate users, you make it an option to have an interface that is 100% identical to what they are familiar with.
Real "usability testing" requires more people with more experience levels on different systems, including people with little or no computer experience at all.
If you REALLY want to make the system easy to use, you have MULTIPLE options:
# 1. Basic level. Almost no menus and lots of "I want to" included in the icon's name ("I want to send an email to someone" or "I want to look at web sites").
# 2. Emulation level. 100% Win2K look-alike.
# 3. Whatever other interface you design.
The key is to build the interface to the user and what the user expects/knows.
Re:Mod parent +25. (Score:4, Interesting)
I would have to disagree... I do phone support for a living and help people that have only used windows computers and are not that technically literate and when I ask them to look for a folder in a list they often will look at the files and expect it the entire list to be alphabetical order like on a Mac. I have to correct them and say to look at the top of the listing for yellow folders and look for the folder name there.
The funny thing is that these people have never used a Mac.
Apparently for those who have never used a computer the most logical expectation is that all the files and folders are alphabetical. When I first started using OS X, I found it quite annoying since windows always had it the other way around but take a person who has used windows but never bothered to actually learn it and you'll find they'll expect something else..
Re:Hm. (Score:5, Insightful)
Sure...
I have seen the page and the different case studies, they seem ok but I think there were 2 or 3 cases that are a lot more common:
1. Scan a picture, create a new document and write something about the picture.
2. Move the pictures of your camera to the place where you save your pictures in the computer.
3. Engage in a multimedia chat with some friend (micrphone+webcam+text)
Of course every linux user knows [although some of they deny it] those are non trivial tasks in a linux distribution
[I can hear the shout of a thousand Linux zealot moderators
Re:Hm. (Score:3, Insightful)
2 - Ok: Step 1 : ask for the name of the software needed. Step 2 : run gtkam, and get the pictures.
For lucky people who have USB-mass-storage cameras, and know how to use mount, it's even easier.
As a matter of fact, I don't know how easy it is on win, because I didn't even try to do it with my own cam, it jus
Re:Hm. (Score:5, Informative)
Putting a CD into a Windows machine pops up a box, one of the options of which "Copy the music to your computer". Your boss must be really dumb.
Claiming it is non-trivial on Windows is fallacious though, it's certainly no more difficult. Windows Media Player is included with Windows (except in some versions in the EU!) and is capable of ripping CDs.
Re:Hm. (Score:3, Insightful)
Parent: No, these are non-trivial tasks in a *ms-windows* distribution
Oh, come off it. These tasks - ripping CDs, installing printers, etc - are things that millions of Windows users, despite the ill-wishes of a lot of Linux folk, manage to do every single fscking day. And yet even some fairly technical users still get confused as to things like which sound daemon to run on which distribu
Re:Hm. (Score:5, Insightful)
Of course the user has to start with a basic amount of computing experience. But you would expect people with windows experience to do well when switching to macs, because the mac interface is well designed, even though it is not the same as windows. So the question of whether the average windows user can figure out the linux interface is a good one.
Re:Hm. (Score:2)
Re:Hm. (Score:5, Insightful)
Might this only result in the Linux desktop becoming more like Windows?
Maybe... but...
There was a brief comment in an article in, I think, last month's Linux Format (UK magazine) (I'm at work, so can't get at the article, sorry). Usability testing had been done on Evolution, and it was observed that one volunteer repeatedly used the "send/receive email" when they wanted to create a new email. The testers realised that the traditional "send/receive" button was not particularly intuitive. To my mind, that's the kind of useful information we might well get from this kind of testing - not assistance in turning Linux into Windows 2.
I mention this only because I believe there's still hope ;-)
Re:Hm. (Score:2)
-matthew
Re:Hm. (Score:3, Insightful)
Re:Hm. (Score:3, Informative)
# Issues encountered:
1. The date and time configuration tool is not easily discoverable from the menus, and is not listed in Personal Settings.
2. Users assumed the root password request meant they had to log in as root.
3. Users wanted the click behavior of the clock applet to be similar to Windows.
# Recommendations:
1. Fix time and date settings to not require root access.
2. Add tim
Re:Hm. (Score:2)
So much for this (Score:5, Insightful)
If you don't have a diverse testing population, you aren't going to produce meaningful results. The idea is fine and all, but the results are mostly useless.
Re:So much for this (Score:4, Insightful)
It's how they solved it, you are trying to find problems not produce useless statistics.
Re:So much for this (Score:2)
Re:So much for this (Score:5, Informative)
The required population size depends on what you're trying to test and how carefully you select your population. If you're trying to test the failure rate of moderately experienced Windows users performing tasks on a Linux system, and you can accurately identify and select moderately experienced Windows users with no prior Linux experience, then you only need a tiny population.
Testing the failure rate is important: 100% of 11 users succeeding at a task can give you at most ~ 90% confidence that all similar users will succeed. 1 of 11 users failing is a far stronger result, telling you that you can expect at least 9% of all users to fail.
Various usability experts suggest that as few as 5 or 10 individuals are required for usability testing, and the remaining usability issues are discovered and resolved via the bug reporting and maintenance processes.
Re:So much for this (Score:5, Informative)
89% sounds like a very good success ratio for the date and time test. However, RTFA and you'll see that only eleven people participated, most of them female.
Eleven people is a pretty good sized group for a usability test. This sort of testing is pretty expensive and time consuming, it's not like a survey or something. From a group that size, you can get a pretty good idea of how the average person will try to accomplish a task and some problems they may encounter. I've worked on projects where usability tests included only three people to test the interface to a product costing tens or hundreds of thousands of dollars. I'm going to have to disagree that these results are meaningless. The important thing is how did the user who failed try to do the task. What stopped them? What problems did other users have?
Like fish in a barrel.... (Score:5, Funny)
And we all know that programmers have no frickin idea how to satisfy a woman.
like shooting karma-fish in a slash-barrel. :)
Re:So much for this (Score:4, Informative)
So only one of them had problems? Sounds good.
You shouldn't let the small numbers put you off. Respected usability professionals say you only need five people for meaningful results [useit.com].
Re:So much for this (Score:4, Interesting)
Re:So much for this (Score:4, Funny)
Parse error: Expected verb, got '?'.
Re:So much for this (Score:5, Funny)
Woot! And we got video of them, too!
Finally! (Score:5, Insightful)
Developers, you don't get to check in code until you've watched the video of users struggling with your program. OK?
Re:Finally! (Score:4, Insightful)
Re:Finally! (Score:3, Funny)
this is needed (Score:4, Interesting)
opensuse.org? (Score:3, Interesting)
That's funny, that url points to betterdesktop.org.. Is this subliminal advertising now? o_O
The day the internet choked (Score:5, Funny)
Fortitude (Score:5, Insightful)
It is difficult, but it's vitally important. These people aren't stupid losers- they are fluent in another operating system, where they can achieve whatever it is they want.
The problems on show here are ours, not theirs.
Martin
Re:Fortitude (Score:3, Insightful)
Re:Fortitude (Score:3, Insightful)
Less technical people don't install nvidia cards. Less technical people use whatever came with their computer. And anyway, nvidia have done a quite stellar job with their Linux drivers - the only objections I have are ideological. The procedu
So ... everything should run like DOS? (Score:5, Insightful)
I have users who were quite skilled with Win2K who are lost with WinXP (until I show them how to make it look like Win2K).
So, which interface should Linux emulate then? Win2K or WinXP? Or Mac? Or something else?It is difficult and it is important
This approach will give you completely different answers depending upon whether the group you select is familiar with:
a. Win2K
or
b. WinXP
or
c. MacsYep. And so the "best" interface for Linux would be
Novell could have saved all that time and money and just spent 10 minutes with a Windows machine, copying down menu locations, order and wording.
There is NO "usability testing" being performed here. No one will learn whether a specific Windows implementation of a menu is less optimal than a different one.
All that will be "learned" is whether those users can find the Linux equivalent and that will always be easiest for them when the Linux menues are 100% identical to the Windows menues that those users are familiar with.
Yes but ... (Score:3, Interesting)
First, I don't want users to mess with system settings unless they are allowed to (e.g. unless they are admins in 'wheel'). I'm happy to support regular users, but not regular users that think they should be adminitering a system they don't understand. I'm not try
Re:Yes but ... (Score:3, Interesting)
Tell the person to choose a time zone and NTP does the rest. You could have this for every user on the system if you want.
Either way, a DESKTOP/WORKSTATION situation has issues such as this which normally you can change (see: all other operating systems)
Re:Yes but ... (Score:3, Interesting)
Re:So ... everything should run like DOS? (Score:3, Interesting)
Not sure I follow you. Are you saying flat out "Windows has the best user interface"?
Because I call bull.
The best user interface is the one that your user can use easily and be most productive in. Sure, for some that will be Windows. For some, Macs. For some, Linux. Heck, I was rather fond of RISC OS.
IME, the great majority of Windows users a
Re:Fortitude (Score:2)
While I do agree a bit with this, we do need to keep in mind that it is a completely different Operating System which is not the one they are use to using. I think you'd see similar results if you take a person that has only used Windows and put them on a Mac, or vise versa.
What we need is people that are well trained on all Operating Systems and be able to compare them without zealotry or bias. We know that MS or Apple won't change their OS layout unless 3 mil
Re:Fortitude (Score:3, Insightful)
I have always had the greatest patience for someone starting off learning something, but I'd suggest that the users you're referring too are indeed stupid, and that your use of the term "fluent" is confusing the issue. The average Windows user is fluent only to the degree they have learned to recognise certain icons (on their desk
Re:Fortitude (Score:3, Insightful)
So because they are fluent in english, we should do away with french even though it creates the better poems?
Sorry, blindly copying is not improvement. Intelligent copying is not doing everything the same way. Take what works, leave the rest. And the definition of "works" is not "what people are used to". A lot, a huge amount of the things that "work" on windos are actually cludges
Desktop tasks? or config? (Score:2)
For the former, both Windows and Linux are equally simple, because it's a simple task. For the latter type of task, Linux is substantially more complicated than Windows, but for Joe Bloggs it doesn't really matter much because they hav
Re:Desktop tasks? or config? (Score:2)
Eh... to open, and write into, a word processor: yep, both just as eas
Re:Desktop tasks? or config? (Score:4, Interesting)
Re:Desktop tasks? or config? (Score:3, Funny)
Those are very interesting questions. If only there were an article somewhere that answered them.
Of course, what would be really great would be if some people would not just assume certain tasks were "simple" and move on, but actually watch some real live users try to acomplish them. They could even videotape it to see exactly what the stumbling blocks were. Then someone could write an article about it so others would understand what they assumed was simple actually causes problems for people. Of course
It's one thing to do an analysis... (Score:4, Insightful)
Go to the web on a Linux PC (provided you've got a browser pre-installed), and download a tarball of say, Firefox. You are a Windows user but you're 'elite', so you use Firefox, and since it's available for Linux, you want to have the same browser.
You have downloaded the tarball, presumably to your desktop. You double click on the file, and it gets opened by Archive Manager. And from here, you can bet that 99% of the Windows folks that would like an alternative to their PCs will not make the adaptation to Linux.
It has to be EASY. Apple set the benchmark for this -- and if imitation is the greatest form of flattery, then do it. Who cares about inflating Apple's ego? If Linux makes a breakthru on the desktop because it's as easy to use as an Apple, or even as easy as Windows, how does that hurt anybody? The true geek can rely on the the commandline only distros, or drop to terminal to get their tasks done using regular expressions and grep or whatever they want, while the 'idiots' (and I would venture to say, that I'm one of them) can use the nice GUI that's simple to follow and easy to use.
Then folks, when developers see that they can cross develop applications that work in Linux (with little overhead), and that people will be able to easily use and access them -- THEY WILL. The open source community just needs to see that fact and start making solutions happen. With the extremely fast and accurate nature of Open Source, the feats that have come from it are amazing. It's more amazing, that the basis of Open Source -- Linux -- remains fundamentally unchanged to accomodate the eager Windows users (read: ME) to switch fully to Linux. Until the snobbery stops and changes start, Linux on the desktop is going nowhere fast. And that's upsetting for a Windows user tired of his OS, and not wanting to get tied into another corporate entity (Apple).
Re:It's one thing to do an analysis... (Score:3, Informative)
I mean really, pullleeeze. Installing winzip or somesuch is something that I nearly always have to do when getting my hands onto ANY WinDOS machine for the f
Well I hope you're right! (Score:3, Insightful)
The GUI form however, is still difficult to use for the average Windows user (me). Fix that, and you have another convert to preach the word
Re:It's one thing to do an analysis... (Score:3, Insightful)
RPMs are a start, but there are so many problems I've had in the past with them, they are relatively useless. Install packages need to be like OS X. Drag the app in, and it just works. I don't know how much easier it can get.
Re:Windows apps are NOT distributed as source code (Score:3, Insightful)
Why do you assume that Windows developers will package their apps correctly but that Linux developers will not?
Well, possibly because 99% (conservatively) of all Windows programs are packaged correctly - and this holds true on everything from, say, MyMinesweeper to DB2. At best I'd say that 90% of the software I install from source works first time (sure, all the rea
Missing... (Score:5, Funny)
1) Ooops! Find your kernels source, kill X, and install the drivers for your video card. Oh, and updated XF86Config. Or Xorg.conf. Whichever one you happen to have.
2) Damnit, another kernel panic. Find what obscure change caused it to happen this time!
3) Ah, so now you have a wireless card? Try to get it working! You might need to use ndiswrapper. If you get another kernel panic, go back to #2.
4) Ah, can't get above 800x600 resolution, eh? Yeah... find your monitors horizontal and verticle refresh rates. Google it, and you might get lucky.
5) Figure out how to resolve RPM dependancies. Shit, that package needs Python 2.4.2, huh? Ah well, 2.3.9 is installed. Guess you're out of luck.
All joking aside, this was a pretty intresting study.
Re:Missing... (Score:2)
My brain couldn't help but spit out additional steps to the ones you listed just like a stock ticker during the crash of 1929:
Re:Missing... (Score:4, Insightful)
Linux needs to work on getting software and hardware to work together reliably. That means without having to edit configuration files and without going to a command prompt. Simple basic things are missing. We need to work on drivers, resolving dependencies properly, and making packages that just work (including installing icons and adding documentation).
After we get that stuff resolved, then tweaking the UI will become more relevant.
Re:Missing... (Score:3, Insightful)
You are either a troll or arrogant.
S'top the press'es! (Score:2, Insightful)
Re:S'top the press'es! (Score:4, Funny)
http://angryflower.com/aposter3.jpg [angryflower.com]
Re:S'top the press'es! (Score:2)
Obviously!!
Re:S'top the press'es! (Score:5, Funny)
I just can't parse it.
"Novell's Releases Linux Usability Testing Videos"
Let's take it a word at a time.
Novell's Ah, it's got an apostrophe s on the end, so it's either possesive or contractive. As Novell is an entity I'm assuming that we're talking possessive here. Something belonging to Novell. Good start...
Releases. Well, this can't be a verb as we're expecting the noun that is possessed by Novell, so while it might be nice to think that "Novell Releases" is the start of the sentence, instead we're looking at somethings (it's plural) that Novell owns. So Novell's Releases. Some items owned by Novell that have been released. Excellent, now what about these mythic Releases...
Linux... This isn't so good. Linux is a noun, and not a verb... Three nouns in a row? It's probably not unheard of, but in this case I'm expecting a verb. I want to know what Novell's Releases do... Well, let's soldier on and see if the verb appears later... Perhaps Yoda wrote this.
Usability... Nope...
Testing... Hmmm, perhaps test is being used as a verb and the entire portion in front is being used as a compound noun as favoured by Germans...
Videos. Yes, that's it....
The "Releases-Linux-Usability" (whatever that is) owned by Novell is testing Videos!!! Are they testing VCRs? Video Codecs? Movies? Perhaps if I read the article it would tell me.
Or perhaps they REALLY meant "Novell Releases Linux Usability Testing Videos" NAAAAH!
Z.
Re:S'top the press'es! (Score:4, Funny)
If you apply that rule to Linux, then it's clear:
Ok, now we just have to find out what it means to linux a video
Re:S'top the press'es! (Score:3, Funny)
The problem of "it's good enough" (Score:4, Insightful)
Desktop Linux needs to grow up in a hurry. That means it needs to be as easy for the average user to use as Windows XP is by the time Vista comes out. I've used a beta of Vista and was incredibly impressed... and I'm a Mac fan first and foremost. Vista is a major threat to Linux and will solidify Microsoft's control, not end it, if things don't change.
Re:The problem of "it's good enough" (Score:2)
I know I'll hang on to my XP Pro box for a while before I rush off and go to a Vista based system, even with all the eyecandy and new features.
Re:The problem of "it's good enough" (Score:3, Insightful)
Cheap commodity PCs that need a GUI to run them. It's the only reason Windows got big in the first place.
The results are in (Score:2, Interesting)
Not ready for desktop until usablilty addressed (Score:3, Funny)
What's worse, some frequently used apps don't conform to any options standards at all. 'ps' takes a confusing mixture of options, some with dashes and some without, which are mutually incompatible. 'tar' needs some options without dashes, and some with. 'dd' uses a totally different keyword-based scheme like 'foo=bar'. And 'find' has its own little expression language on the command line.
Clearly, grandma isn't going to be able to use Linux until all of these confusing option schemes are made more consistent.
Folgers? (Score:2)
Sounds like an elaborate advertisement, eh?
-matthew
I'd go a bit further than that... (Score:5, Interesting)
How about "always completly fscking impossible" for more precision.
Not that there's nothing to gain from training and experience in usability design. Far from it: it will let you skip many obvious problems, and help you resolve others that users find for you in better and more efficient ways. But until your interface is tested on "real people" in at least a couple of iterations, there is no way in hell you can call it "good", "finished", or anything of the sort. If you don't agree, you've probably never done any real usability trials. There are always surprises. Often really big ones.
Your fine tuned detail somewhere may work just as plannned, but it will easily be swamped by problems stemming from inadvertent signals the interface is sending which never occurred to you, or from assumptions you never questioned or even spotted, which utlim ately make people (rightfully!) misunderstand the whole metaphor and do the wrong thing.
There are good news though: If you are willing to really really accept that the user is right (the way people percieve your product is in fact the way they perceive it, and you won't be around to explain to them that their thinking is wrong), and have set aside reasonable time to correct the problems you will find, - usability trials are fun!
Seriously. Fun, enlightening, and humbling (but in feelgood way), and they will broaden your horizons by illustrating just how differently from your daily assumptions it is not just possible but common to think. Do them. You'll like it.
Just resist the urge to explain the problem to the subject (except to be able to move on to test other things). Write down the problem in stead. The trial is for your instruction, not theirs.
Flash abuse (Score:4, Interesting)
GNOME is horrible broken! (Score:3, Insightful)
Over the years I realized that the request of contributions is just a poor excuse to avoid conversations with the developers or users who want something to get changed.
Some stuff in gnome-vfs for example was so utterly broken that it wasn't touched for a really long time. There wasn't even a maintainer for it (only a guy who kept putting some stuff in there whenever it was needed). Now some other people seem to have taken over the maintainance of it and the process continues.
But within the GNOME development team I found out (due to own experience) that it's quite difficult if not highly impossible to get some ideas through or to convince a developer that a different approach would have been wiser or better. Not to say save a lot of time. But people kept using the broken components for years.
Even now not everything inside GNOME is sane or reliable and a lot of stuff seem to be reinvented over and over again. See DBUS for example or basic things like "specifications" as found on freedesktop.org. GNOME makes freedesktop.org sound like it's a place for developers from GNOME and KDE to met and declare specifications but this is not always true since KDE had solved most of the necessary things that GNOME still urgently needs years before and their specifications and solutions are often by far better thought through and much more mature - and over the years proven that it also works practically and not just as concept.
For example you can compile KDE with a static prefix in say
Another bad thing about GNOME is that the developers do have nice ideas at time but they lack the power or durability to make the changes or visions they have complete. GStreamer for example is indeed a nice technology and it somehow made it's path inside GNOME but still it doesn't feel like it's truly part of GNOME since some apps use it, others avoid using it and stick to xine. Now if these apps stick to xine then chances that GStreamer gets fixed and a whole part of GNOME is low.
Another thing is that plenty of the developers seem to have rotating focus on stuff. Today they work on this one, then tomorrow they focus on hacking on Mozilla or hack on 'dead ideas' they have that no one really takes serious so all the resources of working and fixing GNOME get's lost with playground stuff.
We all know that GNOME was meant to be a corporate desktop. But then a corporate desktop needs different resources and a different approach. Serious project leading is required, strict guidelines are required, and people with brains to enable them.
It can not be (now that the HIG as guideline exists for some years) that applications developer still ignore it. I don't care for third party stuff. But I do care for the important and key elements of GNOME software that should be a good example and follow these guidelines.
GIMP, DIA, Evolution, Abiword, Gnumeric only to name a few are in no way HIG conform. Some are, but others not. I filled in a bug for Gnumeric not long ago pointing the developer to the HIG v2.0 where it says that the Toolbar should obey the rules of Toolbar & Menus capplet (which is a core part of GNOME) unfortunately the bug was closed as not a bug and no further comments have been given to it.
Also printing is a necessary importand thing in GNOME imo and it can't be that I load up GThumb to print a *.gif file and it ends up in printing a totally black picture on a white sheet of paper, wasting nearly 1/3 of my black ink cartridge.
It's also inacceptable for a corporate desktop to have a document reader and viewer like Evince that prints a whol
Re:GNOME is horrible broken! (Score:3, Funny)
Just say it, no need to soften the blow for 4 pages...
the driver hurdle (Score:5, Insightful)
Linux is certainly making progress synaptic does a great job of alleviating dependency hell and almost entirely masking it from the end user. I'd like to see the linux community not necessarily looking to emulate the functionality in Windows or Mac OS X but instead looking for what would be the most elegant solution. Perhaps something like an online database of drivers that manufacturers could update, which could be automatically 'pushed' onto your computer overnight and silently rebooted (with your permission in a preferences box) so that you don't even have to worry about having the latest drivers it all becomes automatic would be neat, in the event it failed to reboot it could roll back to the previous driver and notify you in the morning of its attempt.
You could allow users to rate drivers and add the ratings to the database, this way you could specify you only want to automatically update to new drivers that are rated 3/5 or higher for example. This could be like linux's answer to Windows update only better.
Tainted Sample (Score:3, Insightful)
Putting people with no computer experience there would be much more enlightening, especially when it comes to finding what things are intuitive and which aren't.
Where's the data? What about KDE? (Score:5, Interesting)
Sounds interesting! But I can't find any data regarding that comparison. Sure, there are tests about logging in, but no data about comparing KDM and GDM.
3 Gnome-apps, 2 neutral apps. Where's KDE-apps? Looking at the data-section, I see this:
Again: Where's KDE? Going thropugh the test data I see that every single test was with Gnome. Where's KDE? So instead of being called "Better Desktop", maybe this should be called "Better Gnome" instead? then again, what can we expect from having a Gnome-guy running the show? So much for equal handling of the desktops....
Re:Where's the data? What about KDE? (Score:3, Insightful)
Both desktops suffer from similar problems anyway. Test one and it very likely applies to the other as well.
Also, more importantly, whatever was tested was a mix. It was a distro with both "KDE" and "Gnome" a
It's No Less True (Score:5, Insightful)
Getting the desktop to look like anything except blurry ass requires an hour of reading about how to install your video drivers. Why? Because after installing your package using the really nice script, it still doesn't work. So you google again and figure out you need to edit that ghastly xorg.conf file. And then Google to figure out why the resolution is stuck. And then Googling again to figure out why the refresh is stuck at 54 Hz and giving you a massive headache. Dual monitors? TV out? You may as well just go cry yourself to sleep unless you're an uber-leet nerd, because that stuff takes hours to set up. That shit is a matter of one click in Windows; my mother can do it.
Then there's networking. Support for your wireless adapter may or may not even exist. If it does, it's probably in one of the generic Prism2 drivers or something like that. Great, but it doesn't help me a whole damn lot - mine says Netgear on the front. Back to Google again. It's also intresting to note that Linux's DHCP client and the server in my Linksys didn't get along real well, even on a wired connection. There's no way someone who doesn't know how that crap works would be able to troubleshoot that.
Of course, there's always multimedia playback, right? The install I liked best so far, Unbuntu, couldn't play anything out of the box. I know it should have been able to, but for whatever reason my install was futzed no matter how many times I reinstalled it. I never could figure out how to make it play videos. There were several settings for decoding and such (as well as about 10 different players to choose from), but nothing seemed to change no matter how I tinkered with those settings. Oh, and Unbuntu comes with several options for audio input and output including ALSA and ESD. WTF is the difference? I've heard of ALSA before so I'll use that one. Oh wait, that one doesn't work, but the ESD one does. Well, as long as I hear sound I don't really care. At this point, I don't even want to Google it.
This is why there aren't more Linux desktops: there are severe usability issues. I find it easier to get a webserver complete with PHP and MySQL up and running on Linux than a desktop. Why? Because I don't need video drivers, audio, or wireless networking. I also don't change my server hardware every month or two. Linux makes a great server, for sure. But as great a server as it is, it's a shitty desktop. And you'll please excuse my anger, I just got finished configuring my Linux install and promptly broke it...again.
Here's what desktop distros should be working on:
Re:It's No Less True (Score:3, Insightful)
The problem is I want my father to be able to configure and install things. For him to do that he needs a system that works. When he buys a new camera or scanner or webcam he wants a single app on a CD or website that he can download and clik and automagic it work
Windows Clone or something better? (Score:3, Insightful)
The solution would be to think long and hard about whats the best way to do things and then stick to it since change seems to be the biggest problem. Just dont change to much and try to KISS.
There arent that many parts i feel must be changed in Linux. For mass adoption a common third party package format for Linux applications would probably do the trick. Make it easy to install applications and drivers that arent managed by the dists repos. Other than that i really cant think of something thats hard to do on Linux.
I think I understand what I thought you said (Score:3, Insightful)
Where I use to work the software development people sometimes were not engaged in what was happening on a manufacturing side. They "developers" thought they new how to do the manufacturing technician job but it was of those, "I think I understand what I thought you said".
They would start a job trying to get a specification together and so the people they would talk to were the managers of the manufacturing technician. Well guess what - they did not really know the job ether and what was ended up being developed would drive the technician up the wall with how things in there words was "screwed up".
What happened on latest projects was before getting to far into the project spec, they also included the technician in the interviews. Then once a somewhat rough spec was put together and some idea of the direction it was going. The next step was to videotape the technician doing the job as it was currently being down. One month was spent on just taping various people doing various aspects of the job. Each taping session went through a post-mortem review with all parties involved, the spec writer, the software developers, the managers, the technician, and anyone else they could drag into the meeting. The tape would be gone through and question like "Why did you do that? That's not written down anywhere" would be said every five minutes. Even the managers were asking what was going on.
What was brought out in all of this is that unless you are actually doing the day-to-day job in manufacturing, you do not understand the process no matter how many design meetings you have with them. This became the standard method on following projects.
Finally (Score:3, Informative)
This isn't hard. Usability labs like this aren't necessary. I only have to sit any member of my immediate family down, my parents friends, co-workers, etc. to get an idea. Maybe more serious testing needs specialised workers but by no means do we need these specialised facilities.
If you want a cool way to benchmark in terms of speed, acuracy, and rate of habituation try GOMS [wikipedia.org]. No testers needed. (For those who know about GOMS, please clean that article up. I haven't had time.)
Better evaluation needed (Score:3, Insightful)
One example: In one of the tests the users have problem setting the time. The recommendation is that this should not require root login. And sure that would make the task of setting the time much easier, but it would also possibly break things like kerberos or NFS file sharing. There is also other users to take into account. Letting ordinary users change the time also have security implications as it makes the track record of various loggs useless.
The proper question to ask, would be why should an ordinary user need to change the time in the first place? Why not make it simpler to hook up to a time server. That way the user wouldn't need to worry.
What the ordinary user should be allowed to change would be what timezone used in his clock.
Re:Define basic tasks (Score:5, Informative)
OK Here is (I think) the complete list from the article:
I found it interesting that eight out of twelve succesfully completed the "Find out if the computer is online" task. I also wonder if these users could complete all these tasks in Widows.
Re:Define basic tasks (Score:3, Interesting)
First Subtask: Understand what this task actually means. Edit the EXIF data of a JPEG file? Add a symlink to the file? Or what?
Re:Excellent Idea! (Score:4, Funny)
=Smidge=
Re:Excellent Idea! (Score:4, Funny)
Re:Right idea - wrong execution. (Score:3, Insightful)
Since 90+% of computer-users use Windows, I think it's only natural to look at the problems from their point of view.