“I’ve been having a discussion with someone from IBM about whether it’s ever possible for Free Software to have a nice human interface. In theory, I think it is possible. But in practice, the vast majority of open-source projects are also volunteer projects; and it seems that the use of volunteers to drive development inevitably leads the interface design to suck. The reasons are many and varied, and maybe one day I’ll turn this into a long and heavily-referenced essay. But in the meantime, here’s a summary.” Matthew Thomas, a Mozilla contributor, explains why not many of the Free applications feature a good UI like most of their commercial counterparts do.
Why Free Software Usability Tends to Suck
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
The real reasonI think that most Free Software(OSS) tends to have less than consistent UI’s is because the members of an open source project tend to come and go and through in their share of the code without any unified leadership. In a commercial project the project tends to have some type of timetable by which certain things are expected to be accomplished by and therefore there is some form of leadership which guides the project through completion. Also another thing that differs between OSS and commercial software is that OSS is designed for a more technical crowd than commercial software. Since poor UI’s are less bothersome for the tech set than they are for the average consumer most OSS projects tend not to be part of the UI hall of fame, but perhaps members of the hall of shame whereas UI’s are concerned.
I haven’t seen any badly designed UI’s in oss, they may not be consistent between applications due to different toolkits being used, but anyone with the slightest bit of common sense can still navigate them. I highly disagree with Mr. Thomas’s opinions, which is all his rant contains. Also, as far as.. “Many hackers assume that whatever Microsoft or Apple do is good design”, is that a joke?
To me, oss (and I’m speaking of the desktop variety) are usually feature-lacking, cheap knockoffs of their commercial counterparts. For example, compare Audacity and Cool Edit Pro, and it’s easy to see what I’m talking about. I personally don’t have a problem with oss and think it’s great that people are willing to give away their source code for free, but I do have a problem with pundets who say that oss software is superior when it’s clearly not. Access to source code does not a superior app make, unless you’re a hacker
I think my biggest gripe about these programs is the lack of a decent way of getting at the help file. In any decently written software, the help file is going to be available via the help menu, with the documentation being hyperlinked. Having to go into the program’s /doc directory (and often times not be sure of what the hell you’re supposed to view the files with) is not my idea of intuitive.
I think the problem is a large part due to the fact that engineers tend to design things for engineers whereas companies employee people to design for consumers.
The general population tend to like simple things which look good. Engineers like power and couldn’t give a stuff about looking good, in fact generally consider it surplus to requirements (i.e. bloat).
Many programmers consider eye candy as bloat but you’ll find users love this sort of stuff. Like ring tones multi coloured backs for mobile phones. Or the looks of OS X.
Other times there is a choice – design something powerful but complex, or simple and easy. I did this once and picked powerful & complex, it was awesome but it didn’t sell.
One is not right and one is not wrong, a bloatless OS would probably be super efficient but a pain in the arse to use – but by the same argument a bloatless building would be really efficient ..but really ugly.
I think that in general OSS software is not so ‘good looking’.
The Main Problem is it don’t work.
Why ? I agree with SSA. It don´t work because hackers come and go (some time just to show off their skills, if you can call it that way) and – speccialy – because there isn’t a leadership and a “conductor” of the specific project.
Meaning slow and chaotic devellopment and no satisfactory final result; unlike commercial software !
What a shame…
trying…
“I think that in general OSS software is not so ‘good looking’.
The Main Problem is it don’t work. ”
how can you say it doesn’t work. just because you don’t know how to make it work, doesn’t mean that it doesn’t work.
On what basis do you say I don’t know how to make it work ?
.. my only concern with the linux ‘Desktops’ is that it doesnt have its own
identity. It tends to copy Windows or Apple. I think inorder for Linux to be
a viable contender in the desktop market, it should have its own unique
GUI. Well… it doesn’t have to be really unique that it will require new users
to buy a book.
from the fact that you seem to think that it doesn’t work
Comparing Audacity to Cool Edit Pro is … audacious. Cool Edit Pro started life out as a dinky little shareware program, Cool Edit, back around 1994 or 1995. It wasn’t much at all. No offense to the author — that’s just how software starts. Well, free software is similar, and Audacity is still in its infancy. Someday it will probably be quite good too. Same story with Paint Shop Pro, WinZip, UltraEdit, etc. In the old days they were pretty limited; now they’re old Windows standbys.
I agree with the article especially when it say OSS didn’t look good and get influenced by Window$ or OSX. However it still usable for me.
Hopefully in the future there will be good UI designer volunteer themselves to improve OSS UI. Other’s comment might be through that most of OSS contributors are hackers that doesn’t need a good UI to use their computer.
But the end user which is the higher majority of computer user are not expert like hacker.
The hacker and engineer (those who think good UI is bloatware) should accept the computer evolution. That is why most of usr didn’t stick to command line system such as in DOS era anymore. In term of usability, DOS (for example) still can be use fo business puurposes.
The first thing open source projects get into trouble is make stuff look cool first. Let’s spend months and months debating on a cool project name, a cool logo and a cool splash screen. Then let’s make everything theme-able. Meanwhile, no one has actually worked on the main source code.
And good usability isn’t restricted to english-speaking able-bodied americans. People who put flash splash intros on their website should be shot. Accessability for the disabled should be part of the GUI design.
What the F,
not all commercial software out there have good user GUI’s, this rant is just that. Sure MS and Apple have invested quite a lot in supposed GUI research but look at their current efforts. Total crap and resource hogs. Computers for the majority of people are hard to use and many people do not understand the current Desktop paradigm. Try doing a little desktop support and you’ll find out. Anyway onto the topic of this message. What about Sun Micro System’s survey into GUI usability helping the Gnome desktop? This happened half a year ago if my memory served my correctly and the idea was to learn from non experienced computer users what works with them. The knowledge is being utilised by the Gnome team in their development.
Anyway for me I think BeOS had the best GUI (with the inclusion of Open Tracker) of any OS I have used. Simple, clean and efficient. That is what I want although some of the user interfaces on Final Fantasy where pretty cool but I don’t see them happening any time to soon )-:
That is quite possibly the most accurate article I have ever read on the usability problems that are currently facing the open source software development community. Everything that has been bothering me for the last two years, all the stupidities I have seen, have been condensed down into a nice little list. I am glad to see that at least one other person working on open source software feels the same way that I do.
Are you people daft? OSS has crappy UIs but then again most commercial applications have crappy UIs too. 3d Studio being a beautiful example of counter intuitive user interface, a button for rotation… HOW STUPID IS THAT? Windows has a crappy user interface, MacOS X has a crappy user interface. Why? because they only give a fixed set of tools to accomplish a task, they never learn how I as a user work, I have close to zero ability to change the way things work and so I am forced accept their lowest common denominator solutions. In this area I think OSS actually got it right, for instance SCWM. I can change virtually anything in this window manager by using Scheme “plugins”. Granted I have to know Scheme but that is just a matter of hiding complexity by giving you the right tools. Can I do this in Windows? No. Can I do this in KDE? No. Why do I need this? Well since I do some things more frequently than others I like to have those things close at hand. Can I do this today? No I can’t generally since the fascistoid “UI designers” are just interested in one solution that will fit all.
Grrrrrrrr!!!
Were I consider the various proprietary Unices and semi-open projects ( QNX, BeOS, OSX ) against Linux, it does seem that the firmly guided applications lead the way in consistency.
I wont use the word usability.. but I do feel that some firmer interface rules need to be laid out sometimes. I found KDE pretty consistent due to the common toolkit but even so a lot of applications in it have menus all over the place that mean very little to a regular user.
I think the best consistency, though, is where an application seeks to resemble other applications in its class – why have a wordprocessor that behaves exactly like KDE – when you can have one that behaves exactly like any other wordprocessor the first time user may have experience on? etc etc
MS Word might not be ideal but if all wordprocessors emulated its interface the alienation felt by newcomers to the OSS platform might be much reduced. ( Read these comments as general not just specific to wordprocessing.. )
Excellent article. However, I’m afraid that not enough people will understand that issue. A lot of people still don’t realize how much of a science on its own UI design is.
“Comparing Audacity to Cool Edit Pro is … audacious. Cool Edit Pro started life out as a dinky little shareware program, Cool Edit, back around 1994 or 1995. It wasn’t much at all. No offense to the author — that’s just how software starts. Well, free software is similar, and Audacity is still in its infancy. Someday it will probably be quite good too.”
Oh, I have no doubts about what you’re saying and I agree with you completely. Comparing these two programs is insane. However, my point was that what is even more insane are the pundets who stand on the rooftops and proudly proclaim how far superior oss software is, which I say that in many cases (including this one), such a statement is simply BS. Just because I have the source code to said program, does that automagically make it superior? I think not
The problem here is that usability is very difficult to define because it depends entirely upon your target audience. Take the command line tools – for the average desktop jockey, they are nearly useless as these people have neither the time nor the inclination to learn them, but (for example) to an experienced sysadmin, they are difficult to beat. The world doesn’t consist just of people who don’t know what to do with a zip file.
In the terms of “power users”, free tools are often as good as if not better than their commercial equivilents (take Vi and Vim) as many of these tools have replaced their commercial parallels.
Usability itself is not about things looking “pretty” or “kewl”, although for many people the eye-candy is the most important aspect. Usability is about being productive.
However, in terms of desktop software, yes there is a difference, but don’t forget open source developers are more likely to listen to the people who use their software (if you don’t flame them of course) and make improvements.
Back in the “old days” it was generally known that hackers made horrible user interfaces for normal people, we are talking awk or GUIs that looked horrible and were all but intuitive.
The next generation grows up with the Amiga, the Atari ST, and so on, all of a sudden they are more used to GUIs, and start to understand them in a different, non code, way.
Then we got Windows and MacOS, and again the same thing are happening. Being used to consistent interfaces drives people on to knowing about them, and wanting to code them as well. A good example would be KDE and QT (both which are very impressive projects).
But how come OSS applications suck then? Well, it’s in the eye of the beholder if they suck or not, so I won’t go and be the expert on others (very opinionated) opinions. I for example can’t stand CVS but loves perforce. But I just wanted to add a few differences which I thought could be interesting for the few who might read this;)
The quality of the developers for the different projects are roughly the same, in general. You get pockets of much better developers sometimes, but they are not decided by the openess of source. But what you often lack in the OSS community is standard development proceedures. If you work for Microsoft you have to follow their rules. The codestandard and style guide should be followed, and a certain quality (Microsoft has high quality, I’m sorry, that’s just a fact) is expected. Public APIs doesn’t change at a whim, they are documented, they have proper version handling, it’s tested and they know their quality (if you don’t know the quality, you don’t have it, even if the code might have 0 bugs…).
But in (most) open source you are free to do what you want, for good and bad. You have a bunch of things to choose from, which are seldom finished or have any real quality (I for one can’t stand IPC and the lack of fibers in Linux), but you are doing what you love to do, and you hate it if someone bashes your creation or you “team”. But in the end, you decide what it is going to become. That is why it’s inconcistant, and why the really boring parts are not really done (API documentation is boring after all).
Don’t get my wrong though, I think it’s great that people code on their spare time and have fun, more power to them! But please remember that it’s a hobby, and stop claiming superiority like you did with you castles in the sand when you were a kid. Have fun and do what you want instead.
Then we have the few OSS projects that are run more alike to how commercial projects (should) be run. The Linux kernel has a code standard, and KDE is very consistant (someone please donate an extra computer to me, I wanna run KDE. Actually, I am 100% uninterested in Linux/BSD, I just wanna run KDE;)) and only uses one GUI toolkit (it’s the maximum allowed according to me).
Yes, it’s a rant, but hopefully someone had fun reading it, and I rather have it make you think and give you new insights (regardless if you do agree or not) than going into a “oss rulez, M$ suxxor” mode.
[i]from the fact that you seem to think that it doesn’t work[i]
This is not a fact.
This is a mere supposition you made on your own mind.
(Please note that I’m not a native english speaker, ação não é sempre movimento).
from the fact that you seem to think that it doesn’t work
I got fooled by the site’s comment about html.
I’ve seen a couple assertions that open source software projects suffer because they’re constantly changing leaders. If this were true, you should be able to name many major projects where this was the case. Emacs? Vim? The Linux Kernel? FreeBSD? GNOME? KDE? TeX/LaTeX? Tcl/Tk? Perl? Python? If this assertion was correct, one would think at least a third of the ones I’d mentioned would have gone through multiple sets of leaders, if not complete regime changes, but you’ll find that this isn’t the case. There are a lot of valid criticisms of open source software (as there are of closed source software), but if you’re going to make off-the-cuff generalizations you’d better be able to back them up.
Look. Anyone could go over to SourceForge and start a page for an open source competitor to Alias|Wavefront’s Maya, but that doesn’t mean they actually have code or even a clue. “Wouldn’t it be cool if…” pages probably account for three-quarters of SF’s projects; those aren’t failures of open source.
The article author hit the real problem with OSS UIs right on the head, and I’ll rephrase it: good UI design is not cool to most people. If you’ve made one of the “if the UI isn’t skinnable it sucks” comments I frequently see on OSNews in UI topics, you’re part of the problem.
If you don’t like the UI of a OSS program you can recode yourself or pay someonelse to do that.
Plain and simple.
you can recode yourself or pay someonelse to do that.
You wont get desktop users by repeating the mantra of being able to do it yourself or shut up
Even when the product is free, people have expectations of quality, which if not met will cause them to migrate to alternative solutions
People wont use Staroffice or KOffice when its so easy to download Word 2000 off kazaa filesharing. Linux now has to compete against pirated Windoze software as well.
“If you don’t like the UI of a OSS program you can recode yourself or pay someonelse to do that. Plain and simple.”
That was one of the author’s points. People who know how to design a good interface, i.e. dedicated interface designers, are often not programmers because they’re *dedicated* interface designers. So they can’t change the interface themselves, and when they make suggestions, they get answers like these from the OSS programmers.
Hmm –
there are a lot of things to think about in Matt’s weblog. Point #4 is an interesting one, as it focuses on the negative half of an issue which is a two sided coin. Copying MS or Apple makes sure you won’t be any better than them – but it can also make sure you won’t be any worse. Say what you will about the stability, bloatedness or whatever of their products – but do realize that both MS and Apple are spending more on usability testing than most any other corporation…
And – it works. Sure, it may not work for you – but for 95% of the population – it meets their needs. People don’t want to have to _think_ about how to use the computer to solve the problem at hand. They want to use the computer to do a task w/o thinking about anything but the task. We’re not there yet.
A lot of usability experts prattle on about these issues, and how they want a “Transparent User Interface” (i.e. so good you don’t have to make special consideration for the fact that you’re working on a computer). OK – lets think about that for a second… Most interfaces which are virtually transparent to the user are for single purpose, relatively simple tasks. Now while a great many things can be broken down into a series of simple tasks, the fact is that a lot of problems people want to solve with computer assistance are difficult and complex. If they weren’t, they wouldn’t need a computer to do it.
A modern example of “usability” which most UI designers won’t appreciate would be Java. From a programmers standpoint, Java is a fairly impressive improvement over C++ (not trying to start a language war). Not having to think about things like memory management while solving a task allows you to concentrate on the problem at hand, instead of wrestling with the computer. Determining how to do this for the usage of end-user interfaces is a difficult, time consuming task.
Sure, test groups of 5 to 10 people can be used to help eliminate 95% of usability problems, but you’re going to need to go through a few iterations. And, this is before any real coding starts – with mock-ups of the UI. This research is then used to produce a “standards” document for the UI which people then need to adhere to. Hmm – how many OSS projects are going to keep people on with that sort of direction?
So – if OSS keeps successfully copying MS and Apple – and their UI is no worse (even if it’s not better), then they win. OSS is free, why spend money if an alternative which is just as good is available?
“I’ve seen a couple assertions that open source software projects suffer because they’re constantly changing leaders. If this were true, you should be able to name many major projects where this was the case. Emacs? Vim? The Linux Kernel? FreeBSD? GNOME? KDE? TeX/LaTeX? Tcl/Tk? Perl? Python?”
Didn’t the kernel just announce a new leader (or whatever term is “correct”) not too long ago? And aren’t those programs the exceptions instead of the rule? How many open source projects are there, that are actual working projects? Lets say 1000 (I’m sure its more, but thats a start), so lets double the ten you listed to 20, that means that about 2% have steady leaders and a direction (I know both numbers should be much higher, but I think the relation would be approx. the same).
The assumption here seems to be that the commercial software world suffers less from bad UI design. In my experience there as just as many examples of bad UI design in non-free software as in free software, and I must say I’ve found IBM to be one of the worst offenders. To me the problem is simply a lack of skill in this area. There seem to be very few people around who know how to do it well, though there seem to be a lot of people around who think they can do it well.
They’re not exceptions because of precisely what I said in the second half of my post: an <em>announcement</em> of an open source project is not an open source project. If a “project” on SourceForge has 0% activity and no released files, it’s not a project.
And, let’s be frank: when people talk about commercial software’s advantages over open source, they are talking about commercial software <em>successes</em>. If you use commercial programs which have been around through several successive stable versions and have built up a measurable userbase as a point of comparison, the representatives you pick for open source must also be programs which have been around through several successive stable versions and have built up a measurable userbase. To do otherwise makes the comparison invalid. (After all, if you let me selectively pick from the commercial market, I can make open source projects legendary for their user hostility look like UI works of art.)
“Everything that has been bothering me for the last two years, all the stupidities I have seen, have been condensed down into a nice little list.”
I am glad someone fells the way I do, also. I tought I was alone on this. And it bothers me so much I cessed to use it.