Submitted by “The best open source applications and operating systems are more usable now than they were then. But this is largely from slow incremental improvements, and low-level competition between projects and distributors. Major problems with the design process itself remain largely unfixed.” Personal Note: I am not sure how many people feel that Free Software has poor usability. As far as the desktop environment, I find most of linux window managers to be the more user-friendly than Windows and OS X.
http://developers.slashdot.org/article.pl?sid=08/08/03/1847220
Sayng OSS lacks usability is like saing Bob Dylan lack charting hits. Perhaps it doesn’t matter.
Edited 2008-08-04 15:16 UTC
I can think of three reasons why there’s this myth about OSS usability:
1/ Some nerds (like myself) not only don’t mind the occasional drop to command line, but often even prefer it to GUIs. This in itself isn’t a problem as OSS is all about choice – however when some users first exposure to desktop Linux (say for example) would be through their geeky friends, the new users could feel intimidated thinking “I don’t have to edit config files / use the CLI in Windows”. This is the main reason I show off Compiz to my mates rather than show off the fact that I prefer to mount drives to connect to networks via the CLI.
2/ FUD. Horrible term and I apologise for using it, however Linux’s biggest problems seem to lye with public perception – a false understanding which seems to be re-enforced by Windows / OS X fanboys and techies on MS’s (et al) payroll spreading myths and half-truths.
3/ Lack of consistency would be my final point. The interface of some Linux programs can vary massively from other depending on what toolkit a developer has used. Granted Windows suffers from this too (in fact, in some ways even more so as Windows developers often prefer to build their own buttons and skins rather than using conventions – even MS’s own applications break the standards from one suite to the next!), however Windows can get away with it more than Linux can as Windows gets greater exposure thus users are given more time to learn the various application specific GUIs.
There maybe others reasons, but those are the 3 that strike me as the most obvious.
Right but my point stands. Does it matter? Do Bob Dylan need to make more songs that sound like 50cent so he can be more famous and, presumably, make more money? Or maybe there’s something other than being the most popular that matters.
Edited 2008-08-04 15:39 UTC
I guess it depends on whether you think Linux (for example) gaining market share matters.
Personally I do. Personally I think it matters a lot that there is a 3rd option to Windows or OS X.
I’m not sure the analogy works. Dylan’s gone platinum a number of times. See http://en.wikipedia.org/wiki/Bob_Dylan_discography .
Edited 2008-08-04 16:38 UTC
The point is that he does not write what is commonly considered “commercial” songs. He does his thing and sometimes it sells but most of the time it doesn’t. He hasn’t considerably changed his music to cater for the common taste.
If you want to talk about “myths and untruths”, that is one right there.
Here is my install procedure for ubuntu
1) Install it (very simple and easy)
2) go into /etc/fonts/conf.d and muck with the symlinks, since ubuntu defaults to something that looks downright horrible
3) run cabextract on a setup.exe, and then do the ndiswrapper song and dance to get wireless working.
4) completely remove the v4l stuff and compile/modprobe newer modules that will support my capture card
5) install vmware, and half way through the install manually patch vmnet.so so that it will bridge to wlan interfaces on linux 2.6.22+.
After that, I start tweaking settings and building a bigass list of stuff to start pulling down with synaptic and whatnot. But those first five things are for the most part hardware specific, and will be different for every person. I have an hp dv9800, which is one of the most used laptop series out there, so this is not obscure hardware.
Personally, I don’t care about having to do that. Took me about a day and a half to figure out how to get things working on this machine, but I am a developer, an os geek, and have been using linux for almost six years now. By contrast, on windows it is a matter of downloading a few things and going through the next-next-next-finish-reboot song and dance, on osx it is even easier.
Thats not myths, thats not fud, is just reality.
The perception of doing this things is as is the platform market and psychological attachment about something. I am one of the persons that knows right now Linux as high level, means I don’t know CLI at all, excluding some rare case when I really have to read it. For me yet Windows seems confuse and I use it for only two reasons: I want to play a game and I have to work with a proprietary MS stuff (a C# IDE). At that very moment is not a fault of Linux there.
Anyway, when I bought a system that I know from scratch that it will run Linux, with all components selected (I do as Apple do
), I have all that runs at the first key, it boots at the half of time of Windows (SUSE 11.0 or Ubuntu 8.04 against Vista Home Basic), it does not start with: rundll32 application fails, I don’t care about antivirus. Looking on typical person applications, I can say that Windows applications are so different: Office 2007 or 2003 are so different (excluding the gradients) from Vista look and feel, Vista itself is a mix of new dialogs and some old dialogs from Windows 9x era, the Windows Live Messenger is not so Vista-ish and for sure is not consistent with IE7 (and all other apps). So, in large, seems a bad thing, very bad one. Adding that Vista comes by default with no codec, and I use again mplayer for looking to movies, for me seems a waste of that sluggish OS, and the experience is very unpleasant.
The FUD that faces Linux, even in that article, is not worth as much as show the same issues that Windows itself as a platform has.
I see what you’re saying there, but when OS X only runs on specific hardware so you can’t really compare the two and Windows has a stangle-hold on the market so hardware manufacturers are forced to support it.
When you consider that with Windows, users are expected to download the latest drivers manually and reboot after near enough every driver install, Linux does amazingly well. I wont deny that Linux’s support for some hardware is very poor, however the majorety of the stuff works “out of the box” with open source drivers, or a quick and simple package managed download of binary blob where you want propriatry drivers. Plus no waiting 5 minutes while your system shuts down, reboots and configures itself – at worst a restart of X.
Granted when things go wrong in Linux or hardware isn’t supports, it takes real geek to get it working. However, I’m not convinced this isn’t the case in Windows either as I’m sure even you have lost count of the number of times less tech-savvy friends and co-workers have invited you round their house so you can fix a Windows install.
The thing is, to the general public, “its not the fault of linux, its the fault of the hardware vendors” is pretty irrelevant to them.
I’m not saying that windows or apple are perfect, or that linux is horrible or anything, I’m just saying that the kind of stuff I described is very common still, and that is why you get that kind of perception, not because the forces of darkness are out to slander it.
Oh absolutely.
To be fair, it very much depends on what systems you are installing Linux on. I’ve had a range of prebuilt and custom built PCs that have all had Linux on and no hardware issues what-so-ever (though granted they weren’t top end systems either).
Then I bought a laptop and had nothing but trouble getting the graphics chip, wifi and even digital sound working on it.
However the chips were rebranded so even windows wouldn’t function without a day of downloading litrally dozens of drivers and installing them. Even to this day (9 months after the windows re-install) I’ve still not got the built in SD card reader working.
I didn’t mean my original post to sound like that. What I meant was articles from Microsoft linked companies are obviously going to release statistics with a bias to MS products. Same goes for Apple and Linux. However Linux articles don’t get the same exposure to the average Joe as Windows / Apple products do.
What I also meant was that hardware you buy from PC World (etc) will have “Certified for Windows and OS X” on it but you rarely (if ever) see the same logos for Linux (despite the fact that said hardware would run on Linux without any messing about)
Like it or not, computing PR for the home user is tailored around “look how great Vista is” because that’s how they make their money.
So perhaps FUD wasn’t the best description, but ultimately it amounts to the same – users are going to be warey of what they don’t understand and you can’t expect them to understand something which they’re not exposed to.
Well, I’m not a ‘real geek’ but don’t have any problems myself. If a peice of hwardware isn’t detected by Windows, I just go to the manufacturer website and download the driver (however, I have recently discovered driveragent.com, so now I get all my drivers from one site). Installing it is as simply as running the .exe file, or going to device manager and updating the driver there.
That’s better than the Linux way, IMHO, where you usually have to resort to a combination of rocket science a voodoo just to get shit working if it doesn’t run out of the box. Personally, I can stand a few reboots, if it doesn’t mean I’m on Google for several days trying to figure out why my f**king printer doesn’t work right.
What do you do when the driver refuses to install? What if there is no driver for your version of Windows or the driver is not available on the manufacturers website? What happens when the manufacturer goes out of business?
You don’t need to worry about these things on Linux but you do on Windows. I have never had to search for drivers on Linux either. The kernel detects everyting automatically, I guess I just don’t see how it is easier on Windows.
This may sound strange but to me that means higher usability than Windows. At least you can get your wireless adapter to work. Try to install Vista on a PC with a Radeon 9200 graphics card. Good luck getting anything better than safe mode graphics…
My PC is in no way high end but it’s still a perfectly fine desktop PC. As I don’t play PC games, I’m in no hurry to upgrade. My PC plays 720p AVC videos fluently — that’s the most resource hungry thing I do with it.
When I try to use Windows on that PC, at best I have to download drivers to get its hardware to work (practically everything when I try XP, incl. LAN) and in the worst case I don’t even get the required drivers.
My PC hardware is fully compatible with any current Linux distro, including fully accelerated 3D graphics with FOSS drivers.
I think there’s a contextual difference between what you’re talking about (compatibility) and the point of the article (usability).
You don’t even need to talk as complicated as your example. Here is a REALLY simple example why Linux just does not get it in somethings.
http://www.howtoforge.com/ubuntu_feisty_fawn_vmware_server_howto
Yes the instructions work, and yes you just need to follow the instructions.
BUT on Windows you download, double click to install, and the maybe reboot with things working right out of the box.
How exactly is that Linux’s fault? Isn’t it more like that VMWare chose not to provide a graphical installer for Linux? It seems more like that the guide was written by a user BECAUSE VMWare didn’t provide a graphical installer, not because they don’t “get it”.
Edited 2008-08-05 07:51 UTC
That seems more like a VMware issue rather than Linux issue to me.
That statement is a bit short-sighted.
Actually, I think a lot of that comes from people who have tried Linux in the past and came away with a bitter taste in their mouths. Many of these people will assume that Linux is now what it was back when they last tried it. Of course, those in the know understand that Linux is progressing at a rapid pace, so it’s not even what it was 2 years ago, or maybe even 6 months ago
Maybe, but i know a lot of people who have never used Linux and wont even try it because it’s percieved as being too complicated for them / too geeky / command line driven / whatever.
If they’ve never used Linux before, their perception of it had to have come from somewhere.
LOL, so where do you think they heard it? Probably from a friend who last tried Linux 5 years ago and told them that it’s like impossible to use. You know, kind of like people who last used Windows on the 9x kernel, and they go around spouting that Windows crashes all the time
Nothing to do with being on anybody’s payroll.
I’m not saying people are paid by Microsoft to slander Linux. I’m saying when techies main income is from Windows and it’s software, they’re not going to start pushing FOSS on to their customers and loose a sale.
Edited 2008-08-04 19:30 UTC
According to many FOSS advocates, making money developing FOSS software is no problem, so then they’d have nothing to lose, right ?
But you might have a point insofar as even though I’ve never had a friend or relative ask me about Linux, I’d probably discourage them from installing it if they did ask. Why? Because they always come to me with their computer questions, and so since I don’t personally run Linux, if they started calling me with Linux questions, I wouldn’t be able to help them.
You might argue that they’d have less issues if they were running Linux, but since I’m the one educating them on the ways of Windows and how to keep it secure and running smoothly, they generally don’t have too many problems.
Myth? I think not…
1—Some nerds (like myself) not only don’t mind the occasional drop to command line, but often even prefer it to GUIs.
Yes, but you are not linux and are not its only users. While I don’t mind if you/we are, I have to admit that we often wonder, quite openly, why others don’t use Linux. The majority of users don’t want to see the CL–or the admin portion of the GUI–they just want it to work. While that may not be reasonable, there it is…
2. FUD?
FUD is relative. Nothing with Linux or FreeBSD may be complex for you–but it might be for the majority of computer users. While some distributions may be easy to install, having all users figure out dependencies, etc. is not something most people want to figure out–again, if we are so “geeky,” why can’t we fix this? That is public perception…
I don’t think Linux has to reach the “mass-market” to be successful. However, if “we” don’t care to get to the mass-market, then we should stop complaining that we are not accessible and/or willing to “fix” those things the majority of computer users want–whether we want it or not…
It isn’t just Linux–it is OSS. Snort and other “free” OSS are not made to be easy to install and configure. Why?
Many nice Free OSS software programs are made (and made well and “easy”) to just work. However, those that want to make money off of OSS have to have a reason why their skills are needed. The easiest way is to have “free” software that requires help to setup, maintain, etc. Don’t think so? See RedHat, Snort, and a plethora of other software–unless, of course, all of their software has been updated since I last looked at it…
The last time I looked, paying for software that was not OSS was cheaper than getting free software+the service required to set it up to work “out of the box.” Of course, that is just the opinion of a small business owner who has looked at many OSS packages and been left wanting…
Examples where OSS was more expensive:
Zimbra, et al.;
Snort, et al.;
Examples that just work
OpenOffice;
Firefox;
Zimbra Desktop;
Look at those packages and how their implementations differ. Then, I think, you can see what I mean…
Lets say for the sake of argument that Bob does lack charting hits (he does not). This does not affect the quality of his music. The lack of usability most certainly does affect the quality of software.
Or perhaps you mean that OSS is perfectly usable to a certain user group. In that case, it does not lack usability, but is simply aimed at different users. No problem.
If, however, OSS is aimed at a group of people who find it hard to use, it has failed, and therefore sucks in a very real way.
I agree that having something like “Best in Show” award each year for a specific catagory would be encouraging for developers.
As for designers and programmers, my suggestion is to move more towards like Glade for GUI development. Then non-programmers can develop the interface with programmers gluing the code. May not be the best practice but still probably beats a good piece of software render unusable by bad interface.
I like that idea too. I am sure it will encourage programmers to at least think about design. But I believe it is easier said than done. First, what will be the factors for usability. How do we differentiate between the heavy-weights(like kde, gnome etc) and light-weights (smaller apps like eog, gqview etc)? If one program wins the award this year, do we let them compete every year?
Interface or GUI designers are available for just about every kind of widget set out there. The problem is there are a few people skilled enough to make a really good GUI with them as hthe article says.
It’s somewhat misleading to talk about “free software” because that covers so many programs. I think that large, well managed projects like KDE and Gnome probably suffer the least from the issues covered precisely because they’re managed more like closed source software. Usability is indeed an issue with many, particularly small projects.
Coding before design. Software tends to be much more usable if it is, at least roughly, designed before the code is written. The desired human interface for a program or feature may affect the data model, the choice of algorithms, the order in which operations are performed, the need for threading, the format for storing data on disk, and even the feature set of the program as a whole. But doing all that wireframing and prototyping seems boring, so a programmer often just starts coding – they’ll worry about the interface later.
I disagree. It’s not that designing an application before coding is necessarily a bad idea but it seems to work better for proprietary apps then open sources apps. A lot of open source apps that spend so much time on design ending up being just a design with little code involved, or sometimes overly complex code. The OSS way is to get something working and go from there. This works in the OSS world because if the idea is good and there is a working implementation there is a much better chance that people will either get involved and eventually fix usability issues or they will create their own, better designed solution. OSS is much more organic than proprietary software and because of this strict design is less necessary from the get go.
Too many cooks. In the absence of dedicated designers, many contributors to a project try to contribute to human interface design, regardless of how much they know about the subject. And multiple designers leads to inconsistency, both in vision and in detail. The quality of an interface design is inversely proportional to the number of designers.
That’s why I use GNOME. I think the HIG is probably the biggest reason why I use GNOME over everything else. Varying interface design elements are not an OSS only problem and my GNOME desktop has a much more sane interface overall than any installation of Windows with a standard set of third party apps.
Chasing tail-lights. In the absence of a definite design of their own, many developers assume that whatever Microsoft or Apple have done is good design. Sometimes it is, but sometimes it isn’t. In imitating their designs, Free Software developers repeat their mistakes, and ensure that they can never have a better design than the proprietary alternatives.
Again, this is not strictly an OSS problem. You can witness this with the majority of software out there.
Scratching their own itch. Volunteer developers work on projects and features they are interested in, which usually means software that they are going to use themselves. Being software developers, they’re also power users. So software that’s supposed to be for general use ends up overly geeky and complicated. And features needed more by new or non-technical users – such as parental controls, a setup assistant, or the ability to import settings from competing software – may be neglected or not implemented at all.
See my comments about GNOME’s HIG.
Leaving little things broken. Many of the small details that improve a program’s interface are not exciting or satisfying to work on. Details like setting a window’s most appropriate size and position the first time it opens, focusing the appropriate control by default when a window opens, fine-tuning error messages and other text to be more helpful, or making a progress bar more accurately reflect overall progress. Because these things aren’t exciting or satisfying, often years go by before they get fixed. This gives users a general impression of poor design, and that may in turn discourage usability specialists from contributing.
Some of the details mentioned are highly subjective and others are again issues with more than just OSS.
The author makes mention of many design flaws, few of which are problems only in OSS software. Others are just the nature of the difference between OSS and proprietary software. The author criticizes some of OSS’s biggest strengths like release early and often, modularity, and scratching your own itch. If you take these things away OSS fails. We already have proprietary software with proprietary methods of development. OSS has been hugely successful so far and it is only getting stronger. I see no need to exchange OSS design shortcomings for proprietary ones.
Edited 2008-08-04 15:36 UTC
while the gnome hig is insteresting, they basically call everyone idiots and meddling children via their removal of choice…
hell, going from firefox on windows to firefox on linux mess up my sense of hotkeys as firefox on linux attempts to match the gnome hig.
and the change is somewhat strange, as most of the stuff is the same (ctrl+t to open a new tab for example) but some common things are moved around (alt+numbers to change tabs rather then ctrl+numbers, preferences in edit rather then tools).
something like that makes one wonder, whats more important? desktop consistency or app consistency?
Desktop consistency
well in this instance i would prefer app consistency…
or at the very least, some built in way to tune it…
Edited 2008-08-04 16:37 UTC
Firefox as of version 3.0 is now a very well designed cross platform application with native widgets, dialogs, and as you mentioned, hotkeys. You can’t have both consistency with the operating system and across operating systems so pick your poision. I prefer OS integration because it makes using the system overall much easier. Interface consistency across operating systems only makes it easier to use one application.
i could get best of both worlds, if they put in some way to tune it, like i have done with konqueror to match firefox on linux (konqueror behaved more like firefox on windows out of the box).
i dont have a problem with defaults, as long as they are changeable defaults…
If OSS has a usability problem, it is because of the freedom for developers not to appreciate something like the HIG, and act according to that.
In free software, you just can’t force anyone to abide these guidelines. So you can only hope someone understands the need for something like the HIG. All the Gnome team can do is not including the software in Gnome, but if it already is, then it will be hard to get the application follow the guidelines.
In a company, people are just as free to disagree, but there are more tools to make sure that the guidelines are being followed. And in a company, there are customers and there are designers, and they have some power.
The focus on the HIG has made Gnome very pleasant to use. And while i have my doubts on the validity of the HIG for the next generation user interfaces, i feel that for this generation, the HIG was absolutely great. And for the open source community it was a groundbreaking effort.
And what do you know, even KDE has a HIG nowadays.
well i have yet to do the move to 4.x so the impact of the kde hig i cant comment on. but it seems that at least the kde people are not set on making the hig a excuse for a “we know better then the user do” kind of attitude that seems to be growing within gnome.
Yeah, well, i think user interface designers should know better than the user. It’s their job. That is why things like the iPhone rock.
And i wouldn’t say it is growing within Gnome: the HIG had far more visibility during the early days of Gnome 2. Posts on the HIG have become rare on Planet Gnome.
Btw, i am not entirely sure that it is a good thing that KDE is following Gnome. I like the idea of having an alternative desktop where there basically are no rules. That makes it easier to experiment on new user interfaces.
Both coding and design are an iterative process, both should go on during course of the cycle of development, and both need to be user driven.
As for proprietary software not being an organic process, you have not ever been exposed to anything but old style development process (we call it “waterfall”, or “big bang” development now.) The place that I work at now is following Scrum religiously, and i’ve gotten to the point where it is hard to imagine any other process being able to deliver the level of quality to your userbase as you can with scrum.
http://en.wikipedia.org/wiki/Scrum_(development)
Edited 2008-08-04 19:13 UTC
I say the people who start developing their app at the design phase would stop developing their app at any other difficulty they can’t find a solution.
Designing is a good thing. It lets you make the better choices for the first time. Those who just start to code without any plans or design will end up at a completely rewriting their software at one point.
I don’t disagree. If you read what I said though I think it is a much more viable solution in OSS software. It’s the organic nature of OSS. Someitmes it’s more about the idea than the implementation. If the idea is important enough someone will fix the design or design a new application that does the same thing in a better way.
Software developement is very different between proprietary software and OSS software. That doesn’t make one better than the other, just different. I think most people are used to the proprietary way of doing things and therefore think it is the better way to go but I believe OSS software has proven that there is another way.
Mostly the high-profile projects get people on them that, if the project is willing to listen, help improve the usability. Sadly though, some projects just are not willing to listen – GIMP has had this issue for years and only recently started to entertain the idea of changing the interface to make it more usable.
Why? Most F/OSS projects are written by developers, typically for developers. The interface is designed with that in mind usually, and will more often than not satisfy a developer’s use, but be very unfriendly to a non-developer – the average computer user.
So the developers think it is “good enough” since it meets their needs; but the average user thinks it is a piece of junk since they can’t use it without going through a huge learning curve. Those that do, like it b/c the realize the “power” that is available. Unfortunately, it just isn’t something that is ever going to go main stream b/c most are not going to be willing (nor should they have to) to go through that kind of learning curve.
There really is magic in the ‘KISS’ principle – something that developer driven interfaces tend not to do.
But then, I guess it comes down to who is your audience. If you think your audience is only you, or people that agree with you on the interface, then okay, guess there’s nothing to change. But if you really want your project to take off, to be accepted by the average user, then you typically will have a long way to go.
Hopefully more F/OSS software will go towards targeting the typical user than only like-minded people. It would really improve the whole “Desktop” viability thing too.
the learning curve will always be there, its just that so many people are now exposed to the windows or mac way, that if one bump into something that quacks like a duck but dont move like a duck, they freak.
i would have tought that we humans had learned that the only constant is change by now…
It’s not simply a “window or mac way” of doing things. There’s a very general way of doing things, of which both Mac, Windows, KDE, and GNOME (GNOME HID) all subscribe.
But then, compare the usability of GIMP to that of Photoshop, or even MS Paint. Or MS Office 6.0 to ’97 to 2000 to XP to 2007 to OO.org 1.0 to 1.1 to 2.0 to the upcoming 3.0. There’s major differences in usability to them. While OO.org is generally close enough to MS Office that most MS Office users can find their way around OO.org; GIMP is no where near the level of usability that Photoshop has; and I’d even say that MS Paint has a higher usability level than GIMP.
It all comes down to how simple the interface is to use. I’ve tried GIMP from time to time, but (even as a developer) I do not find it in any way intuitive about its interface (which has at least historically sucked). Yet I could pick up Photoshop 5.5 (probably even newer versions) and without reading through a manual, find the few things I want to do and do them.
Yes, there is always a learning curve. But you want that learning curve to be as low as possible for your entry-level audience (whoever that may be!). If you’re not targeting anyone but the most advanced of users, then your learning curve is going to be a lot higher – and should be – but even then it has to meet what that group will find acceptable; however, very very few programs will target that level of audience and target the “newbie” or “just past newbie” stage users; thereby the need a very low learning curve for their basic level tasks.
For example, Subversion (SVN) has a pretty good usability. Their basic level tasks are very easy and typically intuitive. However, there are some tasks that are very advanced (e.g. Hook Scripts) which require a more advanced user, and are designed for just that. At the same time, SVN is also a prime example of bad usability for the ‘tag’ task; the ‘tag’ is defined as a “read-only branch” in SVN, but the system does not natively make them “read-only” or present any easy option to do so – rather, you have to go to the advanced task of creating a hook-script that will do so for you. SVN could easily improve its usability by adding a parameter for branching that would enable the new branch to be set by the system as read-only; or even creating a new function called ‘tag’ that did just that.
There’s many ways to skin a cat, but there’s typically only a few ways that can be taught to someone that does not skin cats for a living, nor has the time to get their Phd on cat skinning to skin the few dozen in the back room.
Again, apply the principle – Keep It Simple, Stupid (KISS).
Brilliant! Now, do it with MS Paint! You couldn’t? Well,I guess you’ve got to bite the sour apple and try Gimp, then?
Actually, most of what I would use GIMP for I can easily accomplish in MS Paint. I’m not a graphic artist; I’m just doing some very simple things – like cropping a picture. (Of course, on Windows I typically will use Irfanview to do a lot of things, and MS Paint to do the rest.)
I still haven’t been able to figure out how to do such a simple task as crop a picture in GIMP. (And, yes – people have given the instructions. But I never have those around when I actually go to use GIMP b/c I’m not reading the thread they responded to when I do use GIMP.)
Programs should not make people feel stupid when they can’t accomplish even the most basic tasks. GIMP can easily do that.
I don’t want to be rude but you have bigger problems if you cannot figure out how to crop a picture in the GIMP. It isn’t difficult at all. Use the rectangle tool to select a region then just select “Image->Crop to Selection” from the menu. It is similar to how you do it in MS Paint only there are fewer steps.
Edited 2008-08-04 22:49 UTC
I wouldn’t say there are fewer steps. In MS Paint, you select the area, drag it to the edge, and then resize the whole area (again by dragging) to the size you want. Or, you don’t even have to select at all, if you just want to crop the edges – just drag the edges to where you want. And you can select the entire image at once too. So, no – I wouldn’t say there are fewer steps.
And see my other posts as far as how GIMP works and issues with it – it’s not exactly intuitive or straight-forward. And the crop tool you mention does not quite work the same way either.
you do realize that the first method you mentioned requires 2 steps, minimum, which is infact exactly how many steps it takes in gimp
and the second method you mentioned only works if you ONLY want to get rid of the bottom and/or right hand sides?
also
1. gimp lets you select the entire image with ctrl-A same as paint.
2. theres no point in selecting the entire image because the whole point of cropping is to get rid of parts of the image
furthermore, no gimp works EXACTLY like he said it does, your black bars were either a weird glitch or you hit the wrong button, though playing around with it, i can’t even figure out what button you could have pressed.
*blink* are we even in the same world?
anyways, where is the baseline for usability? we seem to continually compare new programs to existing ones. but are the existing ones really the to date best in usability, or is it just that they have been around so long that people are used to their idiosyncrasies?
thats really what im asking. to me, both gimp and photoshop is a mass of buttons and menus that will need trial and error, and some looking up of terms and the maths behind them to fully make sense. before that, im just a monkey hitting buttons and switches at random to see what effects they give, and if its anything close to what i want…
but then im the kind of person that tries to learn the “whys” not just the hows, or the right menus and buttons for something to happen. and ones one have the why down, one can apply that to anything similar to get a similar effect imo.
usability seems to focus to much on the how, its like picking up a course in ms office that will just teach you root memorization about what menu or button does what, and leave it at that…
I typically do play with software to try to figure it out. My problem comes with software like GIMP that makes that really hard to do. I don’t do much in graphics – mostly create some real basic things (e.g. icons) from time to time, and crop photos – nothing complex, and nothing advanced. Nearly everything I do can be accomplished via MS Paint; yet, I can’t figure out how GIMP does it.
For example, I did at one point get GIMP to crop a photo – only to have the portion I was getting rid of just become a black border, not be deleted like I wanted. Try that in MS Paint, or Photoshop and it pretty much behaves as expected.
And per your question about usability – it’s usually defined by being able to the most basic, common tasks of the program with relative ease such that little to no training is required – it’s intuitive by the natural use of the program.
An easy way to test – is to take a group of people that have never used that particular program before and ask them to do some basic tasks. You may have to explain the concept of the task if they have never used that kind of software before (all the better for the testing of the software), but they should be able to figure it out quite quickly and without much hassle. When 8 or 9 out of 10 users can figure it out without any other guidance – and they had no prior exposure that software or any of its competitors – then you’re going great and right on track.
So yes, there is a way to test for usability.
from what i recall, the black border was gimps way of saying that the window was larger then the image in it…
iirc, photoshop does that by showing a grid of sorts…
Either way, the black border was saved when I saved the file. For example, I took a 5 MB file, cropped it, and still had a 5 MB file; just with a lot of black in the non-cropped area. The cropping process should remove the rest of the image entirely. I’m not so opposed if it isn’t saved; but it was. I haven’t seen that in Photoshop or MS Paint.
Good question. The answer is there is no single “baseline usability”. It must be identified anew for each user group, at each new point in time. We cant identify a single objectively ideal UI any more than we could identify an ideal natural language. Languages and UIs evolve in response to the environment of the users.
If you don’t want to throw your design at the mercy of evolution, you have to do some research on the current state of user expectations, if not actual user testing. that’s why existing UIs are important. But ways in which existing UIs fail to meet user expectations are important too.
That said, mutations are sometimes successful in evolution, and innovation is sometimes successful in UI design. Given how rare these events are though, the more ambitious a UI is, the more critical user testing becomes.
so we are back at “you can take the animal out of the ecosystem, but you cant take the ecosystem out of the animal”?
Overall, I do not see much problems, but there is one BIG that is never fulfilled: consistency.
Basically you can have a mix of KDE, Gnome, Gtk, WXWindow and other apps. Now try saving or opening a file in each of those. All will come with a differente file dialog. The file dialog should be a task of the desktop environment, not the application.
If you give openoffice, kate and gedit for a newbie, he will have a lot of problems figuring how each file dialog works and why they are so different. 😛
Windows isn’t any different. MS Office has a custom open/save dialog (albeit not significantly different) to the standard NT5+ dialog. Old 16bit applications will use the old Windows 3.x dialog (remember the font dialog in Vista
) and countless other applications don’t even bother to use a Windows API, instead choosing to code their own open/save dialog.
Yes, windows has differences, but for the sake of everybody sanity – the differences are FAR less than KDE and gnome. Just because windows does it wrong, we all should be happy with it??
As I’ve said before, the solution is actually simple: let the DE deal with it. There is even a working solution for KDE using open dialog in gtk/gnome files, so no need for a dictator.
Edited 2008-08-04 18:25 UTC
Nobody said we should be happy with it – I was just making a point that the same problem exists in Windows and people, for the most part, manage fine.
That’s easier said than done when you have renegade developers on your hands.
To use windows as a base of reference again: if the open/save dialogs differ from Microsoft application’s when they should be setting the bench mark, then what makes you think every other developer of every other OS (both FOSS and propriatory) are going to stick to their DE toolkits?
Please? At least across KDE and GNOME you can be sure that the menu bar is above the tool bar.
There’s a GTK theming engine to make GTK/GNOME apps look like KDE apps and the other way around is possible as well.
On Windows I just have to look at IE7, Office 2007, Windows Media Player 11, and to see bigger differences across applications from one developer on a single platform than the differences between the KDE and GNOME HIGs.
consistency requires a dictator…
Overall, I do not see much problems, but there is one BIG that is never fulfilled: consistency.
Basically you can have a mix of KDE, Gnome, Gtk, WXWindow and other apps. Now try saving or opening a file in each of those. All will come with a differente file dialog. The file dialog should be a task of the desktop environment, not the application.
If you give openoffice, kate and gedit for a newbie, he will have a lot of problems figuring how each file dialog works and why they are so different. 😛
I have to argue this everytime I hear it because it is crap. Windows has the most inconsistent interface among desktop operating systems. First of all Windows provides very few built in application compared to either OSX or any standard Linux deskotp OS. Third party apps use their own toolkit, widgets, icons, interface, etc. Even Microsoft’s own applicaions differ greatly from each other. Compare that to a GNOME distro with strict Human Interface Guidelines, standard icons that are often used even outside GNOME specific apps, and only two commonly used toolkits. There is definitly less interface variance on a standard GNOME desktop then there is with a standard Windows one.
//Even Microsoft’s own applicaions differ greatly from each other. //
Especially in, say, Office 2007.
/sarcasm.
Please provide proof of your idiotic tripe.
I’ll reply despite your rude behaviour. Just take a look at the command prompt, WMP, Office (as you so conveniently mentioned), the Vista control panel, MS antispyware, etc. The list can go on and on. To add to this 3rd party apps rarely adhere to any kind of toolbar scheme consistent with Windows conventions. What proof do you have to the contrary? A condescending attitude will not suffice.
I have to argue this everytime I hear it because it is crap. Windows has the most inconsistent interface among desktop operating systems. First of all Windows provides very few built in application compared to either OSX or any standard Linux deskotp OS. Third party apps use their own toolkit, widgets, icons, interface, etc. Even Microsoft’s own applicaions differ greatly from each other. Compare that to a GNOME distro with strict Human Interface Guidelines, standard icons that are often used even outside GNOME specific apps, and only two commonly used toolkits. There is definitly less interface variance on a standard GNOME desktop then there is with a standard Windows one.
first of, im glad osnews has some new blod, as had the editor been one of the old hats im sure the list would have been praised like a gift from god…
the second thing is that a lot, if not all, of these projects start with one or two programmers sitting down to scratch their own itch, and after putting the code online have it snowball on them at a impressive rate.
most commercial software on the other hand sets out to create a appealing product.
the idea of designing first, coding later, is counter to scratching the personal itch.
kde came about because someone wanted a more modern desktop for their *nix install, and none of the available toolkits where up to scratch at the moment, enter qt. gnome then started up because some worried that kde had a leach from its use of a non-free (gnu sense) toolkit.
neither really set out to create what we see to day.
to potentially go a bit over the top, this is like the *nix version of evolution vs intelligent design.
hell, the linux kernel started out as a one man project aimed at creating a working unix for x86 quickly. and it took of like the designed from the top hurd cant seem to match at all. hell, how many times have hurd been restarted because of some new way to design a kernel?
what we really need is for more people to seperate ui from code, so that anyone can create a new ui for the existing code without having to do much coding at all. in that sense we need something like the visual basic interface builder. drag, drop, shape, then hook it into the programmers creation and be happy.
That article goes on an on about why open source software is not usable, and even explains how he thinks that the issue can be resolved, but it begs the question:
What is usable software?
Here’s my answer: there is no such thing as usable software.
There is a number of reasons for this, but the big one is that every user brings a different conceptual framework to the table. The most glaring example of this is the difference between people who love WIMP based interfaces, and those who are infactuated by CLIs.
But even in those worlds, people will differ. Some people are more spatial (i.e. where you find stuff is important), while others are more sequential. Think of the old debates over the Nautilus file manager. In the CLI world, contrast something like Squeak (where commands can be recalled from a spatial text editor buffer) and bash (where you cursor through the command history sequentially, or use bang and some sequential number).
Of course, those aren’t the only conceptual frameworks, but I think that it provides enough variation to make any UI designer’s life a nightmare.
So if we cannot answer the question “what is usable software”, how are we ever going to be able to fix the “problem.”
I agree with your comment, and I would point out that Bash has one more method of retrieving a command. You can type CTRL + R to search through the Bash history to retrieve a command.
In my opinion, Linux, and it particular Ubuntu, is one of the easiest OSs to use.
OS X is easy if you just want to run GUI programs, but if you want to install the latest version of PostgreSQL, it is a sharp poke to the eye. On Ubuntu, it is a quick trip to Synaptic Package Manager. Windows is great for games, but if you want to write out a quick Bash script to automate something, it is a surprise knee to the stomach. At least automation scripts on Linux will work on OS X.
I guess my main point is that usability is highly subjective and depends heavily on who you are and what you want to do with your computer.
Thing us though, how the hell is anybody supposed to know that? There’s a HUGE difference between being easy to use and being intuitive (where usability comes in).
And what if the app you’re looking for isn’t listed in the package manager?
There are tons of ways to automate stuff in Windows, depending on what you want to do. Thing is, Windows is a lot like Firefox: by itself, it isn’t much, but when you start stacking on 3rd party apps (like extensions in Firefox), it gets much more interesting
For automating tasks in Windows, try bash via Cygwin, or Powershell. Or even Perl for Win32. For GUI automation, there’s AutoIt and it’s distant cousin, AutoHotKey.
man bash?
(Duplicate post)
Edited 2008-08-04 21:02 UTC
What the hell is ‘man bash’ … is that connected to the flux capacitor?
Note: That question is rhetorical.. no need to answer it. I remember the first time I sat in front of a Linux CLI, I had no idea how to pull up help. You would think the system would give you some kind hint if you typed ‘help’ and pressed Enter. But there were apparently no aliases set up for that
last time i typed help into bash, it gave me the basic help for bash. yes, not the most useful of help…
i dont know, would it be any easier to find the help file in windows? ah yes, i recall, it gives you the option of a guided tour on the first boot.
i dont know, maybe a welcome text that said something like:
“help is found by typing man bash and hitting return/enter”
Install MacPorts, open the terminal and type
sudo port install postgresql182
or something similar depending on the version desired.
In general, I find the mac combines the ease of use of Linux’s command line with a very well designed UI system. Not perfect, but less room for improvement than the alternatives.
I define usable software as being software I can figure out how to use just by looking at/playing with it. If you stick somebody in front of the CLI, unless they know the name of the command they’re supposed to type in, that ain’t happening. There’s really no reason why you can’t have a little of both though. For example, WinRAR has a nice GUI, but it also has enough parameters to it so that you can run it from the command-line if you want. I actually do a little of both, depending what I’m doing.
“Personal Note: I am not sure how many people feel that Free Software has poor usability. As far as the desktop environment, I find most of linux window managers to be the more user-friendly than Windows and OS X.”
I thought something much the same when I first read the article, and then I thought – hold on. Don’t think of it as a comparison. After all, we’re not trying to be just 1% better than Windows or OS X: we’re trying to make the best software possible.
Forget the willy-waving “but we’re microscopically better than Windows!” and view it as a simple factual statement: does most F/OSS software have poor usability? Yes, it does. It doesn’t matter that most commercial software *also* has poor usability – that doesn’t make the original statement false.
You could say, then, that the question of whether the software is F/OSS or not is irrelevant, and we should be asking why software as a whole has poor usability, but that’s also not the best way to look at it. Fact is, F/OSS actually potentially could have a much *easier* time of improving usability, thanks to the inherently collaborative (or at least potentially collaborative) nature of most F/OSS projects. Most F/OSS projects are already set up in a way that accounts for the possibility of outside contributions.
If you look at the original article, this is not what the author argues, and I’d say the original author is wrong on this in most ways when he draws comparisons between the F/OSS and proprietary worlds. But his takeaway point is actually a valuable one that points to a potentially useful avenue of development for F/OSS projects – especially small ones.
The thing with apple is that user experience isn’t important to them. its an obsession.
Whilst clearly many people that frequent this site have ideological issues with them its not relevant to most people. They appreciate the attention to detail and the user experience.
I was wondering why firefox is so successful and Linux (relatively speaking in the home desktop market) isn’t.
I guess if there was a linux distro that you could install without rebooting from windoze and had a migration manager AND automatically configured dual boot it might help. It would certainly reduce the technical barriers to adoption. Live CD’s are a step in that direction but again how many people would use a live CD for home use?
A sure sign when something is purely subjective is when there is zero consensus on what the term even means.
It seems to me that what designers think is usable is what works for the lowest common denominator. Frankly, I’m not willing to be limited to that. I don’t want to use the same interface that my 78 year-old grandmother thinks is “intuitive”.
No, usability is not subjective. For the individual it may be subjective but usability studies are not done just for fun. When you conduct a professional survey with enough people then usability can be measured.
You can measure things all you want, but it doesn’t mean what you measured means anything. Usability studies are about as scientific as psychology –which is to say not very.
You talk a bit like those Intelligent Design guys. You discard objective measurements from professional surveys just as those ID people discard the findings of palaeontology and genetics.
Of course the measurements from usability surveys mean a lot. Here’s a simple example: Show 50 people a a red sign with an exclamation mark and and another sign that’s just a green circle. Ask them with of these two mean “pay attention”. I guaranty you that most of them will point to the red sign. Yeah, you can throw those results away and design an application that displays a green circle on every important question. Your user base will most likely be confused.
Nice ad hominem. You’re honestly trying to tell me that the “field of usability” is on the same playing field as genetics?
A hallmark of science is being able to get reproducible results. With any given usability study –in the rare event that a study is revisited– they inevitably get different results from each trial. And then there’s the ridiculously small sample sizes of ten or less. I mean, if you’re going to base your observations entirely on statistics, at least have an adequate sample to go on.
Human perception is not objective. All you can do is talk about a population’s average which has zero predictive power when applied to individuals. Like I said before, I have no interest in being limited by the lowest common denominator of whatever population you managed to slap together. A UI is an inherently personal thing, and all I care about is _my_ user experience.
To really solve this problem, the UI should be completely decoupled from the application so that I can build whatever UI suits me. I shouldn’t be forced to use whatever the usability nazis deemed the most usable by they mythical average Joe. That is decidedly the wrong solution.
No, they don’t. If a survey is conducted professionally (enough representative subjects etc.), the result is firm and can be confirmed in another study.
In experimental physics you also don’t get exact the same results every time. There are environmental factors that can affect the results marginally but the overall scheme doesn’t change. Different conditions of the air may result in slightly different falling speed of an apple but an apple doesn’t fall up and people don’t think that a green circle is a warning sign.
I have no idea where you get your knowledge about social sciences but it has nothing to do with reality.
If a study concludes that the GUI of a given program only appeals to 20% of the subjects and 80% like an alternative, that app’s user base can’t grow beyond 20% if you don’t have means to force more users into using that program. 80% will dislike that program.
God, I wish I could mod this comment one point up but I can’t for some reason today. Every time that this usability thing comes up, there are lots of people claiming that Photoshop (or GIMP), Maya (or Blender) or *insert your specialized application (FOSS or non-FOSS) here* is too hard and that the entry barrier should be lower for the layman but what these people fail to ask themselves is: if the person in question does not understand the basics about a given subject (say, image editing or 3D modelling), how could they ever understand how to use the application at all?
To stay on a subject that I understand a little better and have some anedocte evidence to support my statements, if all that you wanna do is to let your users to easily draw some primitives and a floating logo on a 3D window, then fine. I guess that anyone can come up with a UI really easy that anyone can understand, Bryce-or-SketchUp-esque. But if your application is intended to be used as a general modelling tool and some working knowledge is assumed from the userbase (such as what are vertices, faces, normals, meshes, extruding, subsurfaces smooth rendering, splines, etc.) then how the hell are you going to create an app that the average grandmother can pick up and use?!?!? The answer is, you can’t.
I’ve heard several times on the web – including on the comments section of OSNews – that Blender is too hard to use. Fine, I guess that it cannot be compared to the average GNOME app these days. However, pretty much everybody complaining has close to none experience with 3D software at all whereas the Blender boards are filled with people coming from other apps such as Lightwave and 3DSMax that pick up the basics in no time and praising Blender’s highly efficient workflow and are impressed with how much power was packed in such small package. The latest open movie, Big Buck Bunny, entirely made on Blender proves that if nothing else.
In a nutshell, what I’m trying to say is that for a veteran 3D artist, Blender UI is fine although it may appear daunting for the people just getting started with it. The same can be said for other apps, too. While the GIMP UI really could use some love, as someone coming from a Photoshop/PhotoPaint/Painter background, it wasn’t that hard for me to learn my way around it but I would never expect my wife to learn how to change levels on a picture!
Yes, it took me a little while to get proficient with these tools but the point is that it is usable for me on their present state, with minor inconsistencies here and there taken into account. If GIMP is too hard for you, why don’t you give KolourPaint a shot?
Anything is subjective until an objective, empirical system measures it.
“8/12 users took three minutes to complete this task, the other 4 users gave up.” is an objective measurement which can be compared to the same measurement of other products.
It takes special expertise to produce these numbers, which you may not have, but that doesn’t make it subjective. It does make it expensive.
“As far as the desktop environment, I find most of linux window managers to be the more user-friendly than Windows and OS X.”
I have to agree! I can’t live without kwin anymore. Multiple desktops, always on top/on all desktops/… enforcements, per window/per application enforcements of settings etc. Move/resize windows with alt+left/right mouse button (no need to have the window decoration inside the desktop). etc. etc. I love it.
I just want to throw this out there:
I think GIMP gets a bad rap for a couple reasons:
1. it’s not photoshop. that’s been argued about plenty.
2. It’s All There Is:
Over here in windows land there’s hundred of image editors designed for every task imaginable, sprite editors, vector editors, raster editors simple things like paint, more complex ones like Paint.net , enormous things like photoshop.
Over in linux there’s what mtpaint and The Gimp (yes there’s krita, but it’s pretty new so it hasn’t sunk in yet) and nobody uses mtpaint cause it looks like nothing else.
Almost every one on Linux uses the GIMP for almost everything, even when all they want is to make some stick figures really quick to visualize a point.
People use the GIMP for stuff it wasn’t made for and so isn’t convenient at all for.
the GIMP is complicated, with a lot of features, you can’t make a simple interface for, it will never actually be usable for the majority of Linux uses, Photoshop isn’t really much better.
if Linux had a bunch of image editors of various levels of complexity and focus. GIMP would be left to people that actually need it, the sort of people that understand the need for a slew of drop down menus and can actually devote time to learning an interface.
It reminds me of AutoCAD, ludicrously complicated interface, but that’s irrelevant because it’s not meant to be used by anybody that wasn’t trained for it.
I leave you with a link to a post/discussion from OK/Cancel (possibly the only usability themed webcomic)
http://www.ok-cancel.com/archives/article/2004/07/on-unix-command-l…
I wouldn’t say that GIMP is all there is. You mentioned Krita and mtpaint but what about fotox. Then there are vector editors like Inkscape, Xara, and Skencil? There are also RAW photo editors like Lightzone, Ufraw, Rawstudio, and Rawtherapee. Then there are specialized photo applications like Hugin. GIMP is not even close to the only image editing software available for Linux.
well lightzone, raw therapee, bibble, etc are digital dark rooms they generally don’t give you even basic drawing functions so they don’t exist in the same headspace as the GIMP paint etc.
i admit inkscape, xara kind of poke a hole in my pithy summation or a broad problem BUT I feel my general point stands and is restated by Deadfishman here: http://www.osnews.com/thread?325659
To restate myself in a way i’m willing to defend:
“Much of the criticism directed towards The GIMP come from users who do not or should not need to use it.”
that paragraph of my original post was just trying to point out that on Linux there is a “should not need to use it” problem.
Your post just made it seem like image editing in general on Linux was limited to the GIMP and it is not. It is probably the best raster editor for Linux but there are others and when it comes to vector graphics Linux actually has more than one very good editor.
I think the first thing that needs close scrutiny is this blog:
http://linuxhaters.blogspot.com
Amongst the flames, the author of this website actually has a point to his rants. It is too bad that Linux advocates who reply are more concerned about appearances and perception that addressing the short comings of their software.
With that being said, I find it funny when people complain that their *FREE* software, and I do emphasise *FREE* software. These users are getting, after all, a free lunch. Yes, in the last post I made, I moaned about Linux – but there is one thing I do accept – if I want *MY* requirements address, I either chip in and do it myself OR I go out and pay for a turn key solution.
This expectation by end users to have ‘good as commercial software’ for no price tag is just ridiculous (we aren’t talking about server software or non-desktop related stuff, so don’t bring up Postgresql or SQLite – which, btw, are bloody awesome pieces of software – but that is an entirely new topic altogether).
As for the usability, same situation. These are free application, generally maintained by volunteers to address an issue that affects them. What does disappointment, however, is little development is being carried out by the distributions in lieu of expecting ‘volutneers’ to do the work – one only needs to look at the Xorg debarckle not too long ago at the mountain of show stoppers which were ignored. Its all very nice for writers to blame the programmers, but at the end of the day, distributions bundle them with their distribution – it is up to the distributions to do their part in maintaining and fixing bugs in the application rather than just simply pushing the responsibility off onto the original programmer.
Isn’t the reason simple? Linux is not being developed by other users then mostly Linux users. No flame intended.
Of course, Linux does have it’s semi-pro interface designers but you really have to get quite high-level for such a complete experience. And that doesn’t come for free. Look at Apple or what Amiga have (done) for user experience?
Like the article says, Linux has unaddressed problems with this and I really like to see these handled for Linux.
Usually these articles will amount to whining about how FOSS doesn’t have a comparable design or culture to Windows, but be skirting that point.
The points are pretty much dead on. Lack of incentive hits the nail pretty well. KOffice, FI, rocks. All that work, all those mockups, screenshots, competition, hype, hype, hype…and it delivered. With MS making a new and better IE, no doubt we’ll see a better Firefox in the future. Cursing Nero and Roxio for their UIs most certainly helped give rise to K3b. And so on.
My one suggestion, for general application UI: give a scriptable layer to the toolkits, focusing on layout, not what gets executed. That would then be compared against a known checksum, and if different, compiled for a small-but-speedy interpreter. Then, people could muck around with much of the UI of apps, without having to look at C, C++, or Python for that matter.
20 min edit, wee. Obvious changes of tense to the Koffice bit needed. I guess I need to rush and proofread again in the future, ’cause it makes that paragraph look pretty idiotic. *sigh*