Slashdot Mirror
Usability and Open Source Software
Martin Soto writes "This article by two user interaction researchers, discusses many of the usability problems in current open source projects. The nice part is that, unlike many /. readers, it doesn't stop there, but goes into suggesting novel (at least for the OSS community) approaches to cope with those problems in an open source compatible way. Worth a read to those that, like me, still think that OSS should find its way to every desktop computer."
The main problem with open source is and always will be that its created by programmers for programmers. Thats when closed source has the advantage. They are paid to make sure that the user understands the software. Apart from ego what incentive does an open source programmer have?
In the open source world, programming ability is king. If you're a hot coder, you're desired and you can be seen as a philanthropist by developing free software.
However, when it comes to interface design, usability, documentation, and any of the 101 other skills related to developing applications, there just isn't the same level of acceptance.
How many open source apps have good documentation, easy to use interfaces, and professional Web sites? One or two.
There's some darn fine software out there (Apache comes to mind) but where is the demand for good documentation, design, art, QA people in the open source world?
I think that those few writers, artists, and interface people working on open source projects are extremely underrated and aren't getting the credit they deserve.. while someone who comes up with a clever hack in C++ gets their name in lights.
mogorific carpentry experiments
There is no great computer interface, they all pretty much suck somehow. Someday we'll figure it out, but to say it's an open source problem isn't that fair. (Even if open source is sometimes behind some non open source projects)
The combination of OSS underpinnings with the Aqua interface, designed as a commercial project, shows the functional results of one of their solutions. It isn't necessarily the only way, but it gives hope that the other approaches can be successfully navigated.
Closed source, sold software has a strong feedback loop. The developer puts it out, customers complain about it, developer makes it better, customers start buying and give more feedback, developer makes it better yet, more customers start buying, etc, etc.
Money creates a strong feedback loop, which creates a program that fits better with the demands of the customers. (It also lowers support costs. It breaks down in a monopoly situation...)
OSS doesn't have a strong feedback loop. That's why nearly all of the truely successful OSS projects have truely expert programmers, which somewhat make up for not having this feedback loop. Still, there isn't any significant pressure on them to make the product closer to what customers want, instead of what they want, leading to "usability" problems for others.
bullshit. this is *exactly* the kind of elitist attitude that is directly responsible for OSS *not* being more widely accepted.
sure, you say it in a much more palatable way ("... it just has a higher learning curve to the uninitiated user."), but the meaning behind the words is always the same.
people like you fear a "dumbing down" of OSS, and yet you are the same individuals who are first to champion Joe Average abandoning a platform he is comfortable with and "just works" (from *his* perspective, which, as much as you might want to disagree, is all that matters) in favor of one which is completely foreign and threatens the prospect of having to relearn even the simplest of tasks.
wake up. useability is much, much more than "something you add on the backend of a product to market it." it is ultimately what decides, after the marketing hype and initial bandwagon inertia have settled, the success or failure of anything from the simplest script to the most complex architecture.
and, oddly enough, it's actually YOU who really thinks that "useability is a nice GUI where you can get a mouse trail going."
!!!!! useability != gui !!!!!
a gui must take useability concerns into account just as much (in many cases moreso) than a command-line interface.
The paper was meaty, but made its points well. Early on, the authors touched on the difference between the two user communities - the average user and the developer. Sorry for belaboring the point, but that's the problem with the OSS user interfaces - they're not designed to be used by te average user.
My opinion is that the best solution to actually selling OSS software to real users (the 'other' 99%), is to wrap it in a functional GUI that users can use, usefully. If I were Microsoft, that's what I'd be afraid of.
As an aside - there was a good feel of humour to the paper:
Usability is not tacked on at the end. In fact, if you've tried to tack some usability on to the end of your way-cool-code, then your little app is almost certainly not that usable. That's a big problem with OSS, as far as I can tell. You've got to think about *how* people are going to use it and *why* they would use it before you write the code that is the *what* of the equation. I'm an Information Architect, the how and why is what I insist on before my app developer writes the code that does it all. We get much better results than trying to proceed in the other direction.
Somebody in an earlier post said that OSS app coders are just interested in "utilitarian" stuff and that's why they are like they are (the apps, that is). Uh Uh. I don't think so. "Utilitarian" means that someone has to *utilize* the thing. If there's no usability, then utility is a lot harder to acheive.
I don't think Neilsen is god. I think his usability equations don't give enough credit for software and sites that are compelling, as well as functional but, that said, the usability gurus have a lot to teach OSS creators. IMHO.
dcobbler
REG: Yeah. Well, what Jesus blatantly fails to appreciate is that it's the meek who are the problem.
In many cases users are the problem, not usability...
Due to the Microsoft monopoly, most users are trained to understand how to interface with Microsoft products. Thus, people define usability based on how close it is to Microsoft's interface.
Case in point, I was in a library recently. The library has a bunch of iMacs running OS X. A young woman and her friend approached one of the computers and began fumbling about with it. After failing to find what she wanted after 5 minutes, she told her friend in disgust, "I hate Macs," and left.
To me, it appears that OS X has a fairly straightforward, easy to use interface. To this young woman, however, it is apparent that she finds Microsoft Windows more usable than Mac OS X, because her home PC is likely a Windows machine, and she doesn't want to put any effort into learning anything new.
Sure, OSS usability is a bit rough around the edges in many categories. However, the only way it will be "usable" in the eyes of many users is if it copies the interface of Microsoft's products.
"You spoony bard!" -Tellah
Well, first let's factor in one thing. The majority of people developing open source software are NOT paid for their efforts. They have to hold down jobs like the rest of us.
Now, from that, factor in the actual time to write the base code of the application, the time it takes to patch and fix any bugs reported to the developers, not to mention the time taken to (god forbid) live their daily lives. Writing the actual interface becomes doing the bare necessity to make the bulk of the code work.
Perhaps if they were paid to do nothing but sit at home and code for 8 hours a day on their projects they'd be more useable. In the meantime, since they are (largely) unpaid for their coding efforts don't expect something that's going to be as sleek and sexy looking as something you can buy on a shelf.
If people in the open source community (users or coders) are upset by this fact, then I encourage them for the betterment of the movement to grab their fav. language and their favorite open source program and produce something better for the interface, open source the interface, and make it available as well.
----- I want my LART.
Precisely. Another part of the problem is that OSS developers, typically being geographically-disperse and having little access to funding, have no contact with their end users during the design and development phase and cannot do usability testing.
(There's also a "willingness" aspect -- a developer is often not the right person to be doing usability testing with naive users. It's a touchy-feely kind of task, which most developers, OSS or not, wouldn't enjoy, let alone be able to do it well. In commercial environments, that's what the human-factors folks are for.)
Which is how we got to the present situation on OSS and usability testing:
"When writing Soviet GPL code, user interface tests you!"
They are investigating whether poor usability
in most open source software is connected to
open source licensing and open source development methodology.
That is, does open source = poor usability.
Two alternative explanations for poor usability
should be explored (and need to be disproven
before blaming "open source") IMO:
- by historical accident, most open source
developers are unix programmers who don't
know a UI from a hole in the ground.
i.e. the argument is that more UI-attuned
open source communities can do just fine
with the open source licensing/methodology.
- open source GUI software is simply quite
immature. We're just maturing on the server
side; the UI is still comparable to Linux 2.0 or
earlier, not Linux 2.6.
Lots of our GUI software is very newly-written.
And a lot less people are working on it than
are working on the kernel and Apache and
so on.
This will change as the userbase grows.
I also don't take it as a given that commercial
software is hugely better; some important
commercial packages (such as Quicken) have
pretty awful interfaces. Though some
are very nice, for sure.
I waste my time doing this about once a year. I Before Win95 I used Linux almost as much as I used DOS. But, as my time became more valuable a GUI was in need. Win95 gave it to me, X has never given it to me. I'll keep trying so that I'm "open minded", but when a company can take a couple years and get unix on the deskop right (OS X) practically the first time (admittidly, it was released a few months too early), I become uninterested in the year after year failures of the OSS alternatives.
There is no longer anything that can be done with computers that is nontrivial and clearly legal. -- Paul Phillips
This comment is a classic troll. You've made a blatant overstatement. Let's see some evidence, even anecdotal. Has anybody here really tried to contribute documentation or art to a project, but were dissed or not given credit?
I think that those few writers, artists, and interface people working on open source projects are extremely underrated and aren't getting the credit they deserve..
I think you overstate this. All documentation writers and artists receive credit on the OSS projects I'm involved with.
Yes, and for its intended audience, that is highly usable: it's concise, it's easy to automate, it can be typed quickly, it works through ssh, and you can talk people through it over the phone.
The "usability" alternative is something like this:
For end-users who don't know what they are doing, I suppose clicking around provides at least some entertainment, even if it's a waste of time. For expert users--people who have to use this stuff every day--however, even a cryptic command line beats the UI any day.
What an odd and disturbing trend. People publicly admitting that they are unable to perform tasks that tens of thousands of other people have performed with little or no trouble.
Okay, I call bullshit. Tens of thousands of other people have performed the task, but they did not do it with "little or no trouble" unless they already had significant knowledge of Debian. Nobody has little or no trouble the first time the install Debian. I've been using Linux since 0.99pl14 and I've written a miniscule portion of the code in the kernel, and I _hate_ installing Debian, but do so because after I install it I feel that I have installed a system which contains only the things I want (i.e. thinnest possible system).
Any corporate entity I talk to I steer way clear of Debian because their tech people would use it to deride linux as completely unusable. RedHat and Mandrake are much easier. In the time it took me to get 1 good Debian install, I could have installed Mandrake 4 times.
Education is a better safeguard of liberty than a standing army.
Edward Everett (1794 - 1865)
Nobody will see this because I posted so late, but I gotta say it anyway.
The reason OSS hasn't taken hold is because of usability, hands down. Mozilla, OpenOffice, CDex, gAIM, all are good examples of OSS that is quality, easy, etc. They install graphically and simply, have intuitive interfaces, and work like professional commercial software.
Most other OSS is designed by one person. That person has an idea for a program and they design it to suit their personal needs. This software often does not suit the needs of 10000000 users the way something like Office does. It usually ends up being CLI or a piss poor GUI. It's difficult to install and only compiles correctly on one specific version of one specific distro of linux with one specific kernel. The rpms don't work. And there is often already a commercial product for windows that does the same thing, better, easier, and is free, can be pirated, downloaded, or otherwise obtained.
OSS doesn't fail because it is open source or because of the free as in speech mentality behind it. It fails because most often, it sucks. Look at Winzip. Nobody pays for winzip. They crack it or deal with the I agree box. But zillions of people use winzip, myself included. They use it because it is a high quality piece of software, that is free as in beer (not in the world of law, but in the real world), is easy to use, easy to install, and it works.
If winzip happened to be open source it would do just as well. When more OSS reaches the quality of professional software in the same way that Mozilla/OpenOffice/gAIM/CDex have then more people will use it.
The GeekNights podcast is going strong. Listen!
Somebody comments: I thought it was just a distribution problem), and since so many people here on Slashdot rave about Debian I thought I'd give it a try. Especially since PGI(progenies graphical installer) is now 1.0.
Big mistake - you should have found out why they rave about Debian first. Hint: it's to do with raw power, not ease of use or nice interfaces.
tshak says: I'll keep trying so that I'm "open minded", but when a company can take a couple years and get unix on the deskop right (OS X) practically the first time (admittidly, it was released a few months too early), I become uninterested in the year after year failures of the OSS alternatives.
First things first, the usability woes of OS X are well documented. The idea that it somehow magically requires no effort to use is a fallacy. I always end up expending more effort when using a Mac than when using Linux or Windows simply because the Mac needlessly breaks habits to which the vast majority of computer users are accustomed to. This isn't me complaining about things being different, I have no problems with things being different, what I have problems with is the Mac doing things differently simply because that's the way they've always been done, not because it's better. Take the non standard keyboard for instance. Why? Apps don't close when the last window closes, meaning I constantly forget to quit them manually. Why? Software only ejects. Why? These are all usability booboos that you have to force yourself to become used to.
Second point, there's nothing hard about making a desktop based on UNIX. Unix, or rather, POSIX is just a set of standard technologies. What's hard is building a truly free (in both senses of the word) collaborative OS that is flexible enough to appeal to everybody, and yet integrates well enough to be very easy to use. It's hard. We're getting there. Comments like that don't make it any easier.
Tackhead writes: Precisely. Another part of the problem is that OSS developers, typically being geographically-disperse and having little access to funding, have no contact with their end users during the design and development phase and cannot do usability testing.
This applies to most software: any software in fact that isn't produced by a large group usually will not have dedicated usability experts on the team. I don't see people flaming the Windows shareware scene, despite it being home to some of the worst UI atrocities in history. And what do you know, the largest open source projects (gnome, mozilla, kde) have usability teams. It mirrors real life. The idea that all commercial software is more "usable" than open source software is imho a stereotype that's only loosely grounded in reality.
ChileVerde: "It raises the question though, how will the need for usability specialists fit in the current model for developing OSS? AFAIK, most of the usability/interface work on the projects are handled by programmers, who doesn't necessarily have the background on this topic."
Havoc is a great example of a programmer who "gets" usability (though perhaps a bit over the top). I always think of usability when designing my interfaces. Programmer != GUI monster. Often though they're not experts, but that's why we have experts such as the guys from Sun working on GNOME. They already are fitting into the open source model.
There may be a connection. A closed project allows one person to impose their will religiously throughout an interface. Open source ultimately is about concessions and cooperation, which may negate this type of centrist control.
No, it's about cooperation. That doesn't necessarily involve concessions. An open source project is like any other project - the leaders can impose their will with an iron first, or they can be weak and agree with everything. This happens in the commercial world as well.
ACK!! says "The other side of the coin that these folks do not take into account is the fact that OSS application developers for all the desktop adoption talk are not coding for the masses. They might think they are but they are not."
Important insight here - the GNOME flamewars demonmstrate this very well. Some people felt GNOME2 was being taken away from them and retargeted at the corporate desktop user. It had a lot of "crack" features stripped out. It took balls to do this. The flamewars on the lists weren't pretty, and still the trolls keep trolling on forums like slashdot and FootNotes. This is a good example of a large open source project (that doesn't even have one leader) taking the initiative with usability. GNOME proves that a lot of the FUD in this thread is simply wrong: open source can be very usable, and it can be written for non-developers.
I have seen open source overcome every problem it has encountered so far, back when I was excited about this new new thing called Windows 95. I have seen it go through "toy OS", "can run web servers but will never get enterprise acceptance", "good at servers but will never get enough apps for the desktop", "too hard to install" and now "software isn't usable enough".
Every single one of those problems has been solved. This one is being solved too. Tomorrow I release autopackage 0.2 - it's CLI interface was designed with usability in mind. It uses colour to make the text easier for the eye to process, it uses simple, obvious command names (with aliases to facilitate guessing) and it comes with documentation. Open source is dead. Long live open source.
All I can say to this is that I've had a VERY different experience with Linux installers.
My very first Linux install was with a purchased copy of RedHat 6.0. Even then I was impressed at how well it worked. Literally 30 minutes from CD in the drive to a working desktop getting on the net.
Later, I ran a Suse 8.0 install. This one had some problems with the drive which required a low level format from an OEM utility. After that, I was again extremely impressed with both the presentation and functionality of the installer.
I can honestly say the same for Mandrake as well.
I have other issues with all of these that keep me using FreeBSD, which doesn't have the same super-slick installer, but provides for many other benefits. Even still, I managed to get it installed and working properly on the first try without anywhere near the kinds of problems you had.
I suppose the appropriate response here would be to illustrate the many frustrating hours fighting various Windows installs that didn't play nice due to a variety of reasons. How many folks here intuitively knew about the F6 trick to get SCSI loaded properly for NT? How about changing out a motherboard from underneath an already installed system. Oh yeah, Windows just loves that!
Why just pick on Windows though? I've run into all kinds of interesting glitchies with Mac OS 9 and X in the past. Various formating gotchas, or extension conflict finding sucking away the hours.
Go have yourself a visit on any newsgroup or mailing list for OS tech support. All of them have horror stories or odd gotchas that impact every darn thing out there. Coming up with one for Linux is hardly that noteworthy, escpecially when the vast majority of folks are able to get their installs to work properly.
The line must be drawn here. This far. No further.
What an odd and disturbing trend. People publicly admitting that they are unable to perform tasks that tens of thousands of other people have performed with little or no trouble. [...] I used to get pissed off at people who would glibly and gleefully admit that "they did not understand math" or "computers confused them" as if math and computer skills were somehow optional in this world.
You're right. Math and computer skills are absolutely necessary for everyday life just like reading and writing, although perhaps not to the exact same degree. Similarly, and increasingly as time passes, they are essential skills to being an informed citizen (who votes with his or her rights) and consumer (who votes with his or her wallet). Anyone without a solid understanding of computers and digital information is set up to lose out big time.
Now I am going to have to make room in my big ass loser bag for people who are unable to install linux.
BZZT! Wrong. You need to make room in your big ass loser bag for yourself and anyone else who understands technology. When there is a widespread ignorance of and resistance to using Free software, everyone loses. The existance Palladium itself is not a threat to the survival of OSS and the freedom of information; however, the result of combining a cleverly devised public key crypography scheme, terrible legislation like the DMCA, and a bunch of ignorant, sheep-like users is that the universal perception of information is perverted into something false and harmful. You may see that, "If you can see it, you can copy it," is a basic principle of the physics of information, but with enough experience with commercial software and DRM "solutions", I can guarantee you the average consumer can be convinced otherwise.
There are two ways you can approach this problem:
1. People are too stupid and lazy to migrate en masse to free software. Since people won't switch, everyone's screwed.
2. Free software, on the whole, is not usable enough to tempt the average user or create large numbers of converts. Unless software gets easier to use in the near future, everyone's screwed.
If you go with #1, you give up all your power to correct our disasterous course towards unbeatable proprietary domination of the software market. You lose. Please step into the bag. If you go with #2, the opposite happens. You may not be able to change the way everyone thinks, but you sure can develop usable software. When you bitch and complain about how incredibly stupid people are, you pull yourself and everyone else away from a constructive solution to the very problem that irritates you. When you realize what you DO have the ability to accomplish, you are empowered.
Help make newbies comfortable. Help save your future.
Compare KDE/Qt apps to their GTK/toolkit neutral competitors and often there's no competition: Mozilla is a better browser than Konqueror (I've often wondered why the Konq team is still reinventing the wheel; GNOME noticed there was a beautiful, easily-embedded rendering engine available and we got Galeon. Konq could embed Gecko and advance by huge leaps and bounds). Evolution is a better PIM/e-mail program than anything in KDE. OpenOffice beats KOffice on so many levels it's not even funny (how about "actually works with MS file formats" for starters?). There may be some flashy, shiny, GUI IDEs available, but that doesn't make good apps by itself.
Sure, KDE is pretty. Sure, Qt is nice to work with. Sure, the development tools are great. But the KDE team isn't accomplishing anything with them. That's why big companies that use *nix desktops go with GNOME. That's why Red Hat set Mozilla, Evolution, and OpenOffice as defaults in Psyche. And that's what KDE's developers need to realize and deal with if they want to compete seriously for desktop market share in the future.
Like them or not, Microsoft dumps billions of dollars into research. Some of this money is spent on seemingly 'stupid' things like, what shade of blue they should use for their kernel lockup screen. But some money is spent on really useful additions.
Take the mouse scroll wheel. When I first saw this thing I sniffed at it and decided it was hokey. Now, years later, it's like I can't live without one.
I think the OSS community should at least pay periphery attention to developments on other platforms. See what works, and what doesn't and 'borrow' it where applicable. This is certainly what both Microsoft and Apple have done in the past.
The Amish think that technology creates idle hands, however this doesn't seem to stop them from using 'outhouse' phones and generators for their milking machines.
"...Well, there's egg and bacon; egg sausage and bacon; egg and spam; egg bacon and spam; egg bacon sausage and spam..."
Your analysis implies a lack of understanding of what usability is. You seem to say that "graphical" somehow equates with usability, but that is not the case.
A truly usable solution would anticipate that users would rather have one or a handful of useful options rather than a large variety of configurable options that will go mostly unused. It is challenging to figure out the optimal subset of configuration options, but once done it becomes much easier to provide easy access to them.
You also described fixing a problem, whereas the original example was about changing a configuration option. Truly usable software would simply not offer options that create problems.
This is not as trite as it sounds. Many times developers do what Alan Cooper calles "putting might on want." Which means offering an option that really isn't very useful or even actively harmful on the 0.01% chance that someone might have a special case where they need it.