Quality rant by Eric Raymond on the topic of bad user interface design in open source software, making some good points. The editorial was inspired by ESR’s experience with CUPS.
Eric Raymond on UI design in OSS
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
Who is Eric Raymond to complain about Free Software? He didn’t pay for it, what does he expect? If he sees a problem, he should submit patches, not whine endlessly! I hope I don’t sound too rude, but nothing is accomplished solely with words.
I really don’t like the whole idea of auto-detecting printers on the network. Perhaps Apple’s implementation works really well, but XP’s doesn’t. Every time I use that damn “My Network Places” program, it freezes for minutes at a time, presumably trying to auto-detect crap on the network. And even though it finds the Canon i850 I’ve got upstairs, it refuses to print to it.
I guess in theory, when it works, its much better, but IMHO, it is a UI disaster in the sense that if it doesn’t work, there is no recourse.
Um. he has written loads of code for the free software movement. He’s got the right to complain.
His software page:
http://www.catb.org/~esr/software.html
Not only Eric Raymond has the right to complain about Free Software. Everyone can, writting or not “lots” of code to “free software movement”. Free software is for everybody, not only for programmers and/or geeks..
Recomending people slit their wrists and such. The negativity of this man is astounding, and not really the kind of guy you want as representitive of your community. The guy rants incoherently at times, and in many places recomends violent and destructive solutions towhat he considders problems.
š – kind of hard to complain about a GUI system when you don’t show it, no?
Haha, kind of funny, cuz one of ESR’s “theories” on good documentation design is: No screenshots!
OK, I sometimes I’m irritated by ESR, but sometimes not. He’s one of the old UNIX heads who makes a genuine effort to learn from his Mac brethren — but still, I’m beginning to think it’s all in vain. OS X has proved something very important (which is largely ignored), but until UNIX and Linux devs at large adopt a style that thinks from top to bottom (fat chance) — end users will always struggle on *NIX.
Many developers only see their application from their own perspective. And that is somewhat understandable if you are a hobby developer who codes for fun. You are missing a tool and you make that tool so that you can use it yourself. Good.
Since you aren’t responsible for making it work for other people, why bother? “Doesn’t work for you? Well here’s the source-code.”
And that’s perfectly fine. It’s their code. But it saddens me that many people are thinking like this. They often have a lot of talent and they could contribute so much to the community, still they don’t. It’s their choice though, but it’s sad.
Rarely have I seen developers asking GUI/usability experts for help with their projects. Why is this? Don’t they know where to look? Or do they have to much pride?
I’m not a good programmer myself. I can code and has worked as a programmer for a few years, but I can’t say that it’s something that I have a lot of talent for. The only reason I started to program again after a long brake was the fact the noone wanted to help me bring my ideas into working applications.
Most of the time when you have ideas post them on the net to find someone to help you code it, the responser are either silence or “sounds cool. learn some programming and do it yourself.”.
So the person tries to learn how to program spending several months training only to discover that he has no talent for it. He leaves the project in frustration.
I’ve seen this case so many times. They community is loosing a lot of innovation and great ideas here. There are so many people that are great at designing applications but lousy at programming. I can’t understand why these two camps can’t learn to work together.
Just look at the way Eugenia gets treated when she points at minor flaws in GUI designs. Those flaws while not so important by themselves are important because they make the complete product so much less polished and makes it look amatuerish. Still people bash her for pointing them out. While I don’t agree with her on all things, she does know her stuff.
There are a handful of people who have both talents. These are often seen working alone on projects and they do a great job. But as I said, those people are rare. So we either have to breed more of them, or learn to work together.
Agreed. User interfaces blocking on connections can be really annoying. I don’t think I’d mind it as much if there was SOME feedback, but half the time I wonder if it’s crashed. Outlook loves to do that, as does/did evolution. And I find Windows even blocks on initializing my CD drive.
But I do think the idea of autodection is great. Maybe not for corporate environments, but certainly for home users. ESR does have a point here. Often, open source configuration is a major pain. Mind you, some areas are improving. GNOME has improved immensly compared to version 1.x, and things like Kadzu, or presumably udev are really good as well.
Ok, I think the guy over exagarates a little. However, he has some valuable points. First of all I have used and played around with linux and most likely all of you have a decent in depth experience on how to configure almost every part of their system. Yes, you get the software for free, however with all the people making all the noise of how the linux is ready for primetime and for the home user this article should be taken into consideration. Even with its polished desktop, after some tweaking by many distros, many individuals are still afraid to use the Linux or Unix. First of all I honestly see linux being a big winner in the long run however not right now. Currently i think Ximian has the right idea in mind as well as some distros but some of the changes that are implemented are not always compatible with what software is released for those systems. And trust me no ordinary Joe will even want to touch a system where just to install a stupid browser he will have to learn how to compile the packages. Even better the printer example learn all the command whick will allow him to mount the mods. and all proper drivers just so he can print out a photo for his girlfriend or himself. I honestly think that opensouce software is great but because of its lack in order such as menus under gnome or kde, –system tools, –system utilities, and –system preferance, most users will have no idea what is the purpose, neither do I, of these preferances. As pad as win**** is it has a very clean and sophisticated GUI and this is what makes windows so likable and OS of choice.
A snippet:
To understand why, you have to stop thinking like a hacker for a few moments. Cram your mind, if you can, back into the mindset of a clueless user.
And here’s my favorite part:
Somebody who not only doesn’t know what a string like “/printers/queue1” might mean,…
Yup…If even a well meaning critique on user friendly UI design in *NIX is just as bad as the case in question, then it’s got a long way to go still.
The audience of the article was hackers, not laymen, so using a technical term like “string” makes sense.
Companies like Adobe, Apple and even MS, who “get” good UI design have something going for them that OSS doesn’t; usability testing. It may be time consuming, expensive and you’ll never like the answers to the questions you ask, but it has to be done, and in the long runs pays dividends.
Its not that the folk at these places are smarter or better skilled at UI than the OSS crowd. It’s just that they’re willing at some point to throw out their assumptions as to what “Joe Average” will be able to cope with and actually go out and ask him.
Ask enough people and eventually you build a consensus as to what works and what doesn’t and why, and this is how you can produce software that is “an order of magnitude” better than the well intentioned but ultimately futile stabs at usability that permeates much OSS.
Who is Eric Raymond to complain about Free Software? He didn’t pay for it, what does he expect? If he sees a problem, he should submit patches, not whine endlessly!
The “Who is ESR?” question has already been answered.
I take issue with the idea that individuals do not have the right to complain about free software/OSS because they “didn’t pay for it” and anyone who has an issue should instead “submit patches”
When I started working at my current job, the desktops were about 50% Windows, 50% Linux. Currently, it’s become more along the lines of 95% Windows, 5% Linux. So much of my time is eaten up managing more important systems such as our HPC clusters (running Linux) and Sun workstations that I really don’t have time to deal with the day-to-day administrativa that comes along with Linux systems. Windows simplifies administration to the point that users have no trouble performing minor administration tasks themselves, such as installing software or configuring printers.
So, one clear alternative to “submitting patches”… move to Windows. That’s not to say that Windows doesn’t have problems of its own, but I have a much easier time dealing with problems on Windows (since in many cases, you can simply take preventative measures to ensure problems with viruses/worms don’t arise in the first place) than I do dealing with problems on Linux.
I loathe CUPS. I have about a million nitpicks as my own, such as the web-based administration not providing a way to set the default printer, or the fact that it autodiscovers printers on the network and overwrites its own config per default, which wouldn’t be that much of a problem except for that its autodiscovery of printers simply doesn’t work half the time (ESR ran into this problem as well). Deploying CUPS across a few dozen desktops is a nightmare, especially considering most packages don’t even provide a decent selection of PPDs per default.
The administrative headaches of OSS, *especially* CUPS, become almost intolerable when multiplied by an entire workplace of computers. In my case, for the most part I’m quite happy with Windows.
If all the people who knows UI design would be stuck programming there would be very few of them left to work with UI design. It’s much easier to point out what’s wrong and let the developers who are much more familiar with the code fix the problem.
You just can’t be everywhere.
Sure they aren’t responsible. It’s a free product with no contract. But that doesn’t mean that it’s wrong to point out issues. But it’s up to them to do anything about it.
Companies like Adobe, Apple and even MS, who “get” good UI design have something going for them that OSS doesn’t; usability testing.
Not always. Sun did finance usability testing of Gnome. You can read the port here:
http://developer.gnome.org/projects/gup/ut1_report/report_main.html
Despite this and other contributions Sun has made to the Gnome project, zealots still insist that Sun “stole” Gnome from the OSS community.
Considering I often find everyone here whining about how Linux sux0rs, I figured I’d find more agreement with ESR. He’s right-on about most things. Linux is still climbing the usability ladder, and it’s getting better, but too much stuff is unpolished.
For one thing, I really don’t get CUPS/SWAT’s idea of having administration for a program be web-based. It only makes sense to me to have web-based administration for something it’d be nice to access from my friend’s house or from a WIFI hotspot. I feel like a lot of projects do web-based admining because they hate writing GUI code so much they’d rather just write HTML.
To someone like me, a web-based admintool is no big deal. But most users don’t know what localhost is, nor what a port is, nor do any of them find it intuitive that you would access “a website running on your own computer” to configure a program that is running on your own computer. Configuration tools need to get smarter, they need to get friendlier, and they need to get unwebized
I’m sure a lot of Linux users are gonna scream at this article and say, the day we have Mac-like config tools is the day we dumb down our OS is the day Linux isn’t Linux anymore. I _really_ disagree with this view. The idea is that computer systems tend to be built in layers, and Linux has the most solid underlying layers one could imagine. Above those layers, if you mask the underlying complexity, you don’t dumb down the OS–you just make it easier to use. After all, it’s not difficult to peel away the layers simply by vim’ing it up in /etc.
The important thing is that in the transition to a more usable desktop OS, Linux doesn’t forget its roots and the simplicity and solid underlying design principles hold.
I forgot about the Sun/Gnome connection, but it’s the exception that proves the rule, and in any case, no-one is going to hold Gnome up as the pinnacle of UI design.
It seems like when ever people talk about Linux they say “the end users won’t understand this or that”. That is not the problem, Linux has to be more user friendly for administrators and power users before it is ready for end users. End users are people like my mother with digital camera software, printer software for creating cards, incredimail, bonzi buddy etc. Linux does about 80% of what I need and about 5% of what my mom would need. The truth is that a power users needs are much simpler, why not focus on that instead?
Most distributions have their own non web-based printer configuration tool. (Suse, Red Hat and Mandrake do).
Web-based configuration tools are a blessing because they are OS independent, and use a GUI metaphor that a great deal of users are accustomed to. There is no doubt that there is always room for improvement, but there is absolutely nothing wrong with a web-based configuration tool.
Interestingly enough, I have better results printing with CUPS that any other printing system. I run a network of Mandrake desktops and servers. The desktops just find the printers on the network and people choose the printer that they want to use depending on the task, with the fastest laser printer being the default.
Jimbo
Your “joke” is in incredibly poor, bigoted taste.
First, it implies that autistics aren’t equal and deserving equal treatment.
Second, it implies they would purposely create such a nightmare because they don’t know any better, when that’s not the way they work. If anything, you’d likely get the opposite effect: the perfect interface that’s very clear and very logical.
Third, autism isn’t a disease that can be cured by therapy or drug treatment, it’s a different neurotype with different psychology, and a different way of processing sensory input. It isn’t like it’s a mental disease that drugs can change; it’s an innate part of a person that can’t be changed any more than you can truly do a gender change of a person that didn’t already have fully functional organs for the gender being changed to.
Fourth, you should be careful: many autistics are employed in the computer field, and a large portion of them write the code you depend on! What may be surprising to many is how many have managed to blend into society without anyone realizing they were autistic, and that often includes themselves. Autism and low IQs aren’t linked as commonly as many people believe, and many autistics have very high IQs, though they might lack in the skills expected out of the standard individual for social interaction.
You are surrounded probably by more on the autistic spectrum than you realize, and perhaps than they themselves realize…
The audience of the article was hackers, not laymen, so using a technical term like “string” makes sense.
Heh, I know, believe me…Was just making a funny =D
Its not that the folk at these places are smarter or better skilled at UI than the OSS crowd. It’s just that they’re willing at some point to throw out their assumptions as to what “Joe Average” will be able to cope with and actually go out and ask him.
True, it’s not about skill. Good coding is a virtue shared everywhere. But, it’s not all that simple either. I mentioned earlier: It’s about thinking from top to bottom, instead of bottom to top as in UNIX. That’s all it takes. A momentary paradigm shift. You don’t have to survey “users”, or make assumptions about “Joe Average” — Joe Average and Aunt Tillie is the developer — I don’t think anyone who can write 10,000 lines of C is stupid enough to not be a little “pragmatic” every once in a while. Selfish maybe, but not stupid. Consideration, you know?
Wait, doesn’t OS X use CUPS? Is it hard to use on OS X too?
Yup, it uses CUPS. And no it isn’t a mess. I know I’m coming off as a Mac Head, but it goes back to what I said about OS X — it’s a great, underrated accomplishment, Apple are the first ones to make a truly functional, user oriented UNIX.
The rant starts off from a silly assumption. That aunt Tillie would set up a networked printer. Not in her wildest dreams.
Aunt Tillie would not even know what a damn network is. So ESR has picked a very poor example to drive the point home that better user interface design is necessary.
At best, Aunt Tillie would call nephew joe, who would come and do it for her. You know, the one who went to college to study computers.
I have had perfectly capable neighbors call me to set up a local printer on their box.
What is the saddest thing about ESR’s rant? That it proves he is using a system that he does not understand. He is using an experimental distribution such as Fedora and has not taken the time to figure out that CUPS is the default printing protocol and that lpr has been deprecated, which unless he has had his head stuck in the Linux sand for years, he should know. I don’t know of anyone who has been using Linux for a few years that doesn’t know this.
You may say, but he was not pretending that this was *his* problem, he was just illustrating a point. But he isn’t. He really couldn’t configure a damn printer, and that is what he refers to as his own version of the “aunt Tillie problem”.
Absurd. Actually beyond absurd.
Next to him, Eugenia’s complaint’s about GUI design appear like pearls of ancient wisdom and I usually don’t find myself finding her suggestions all that relevant.
One final word, but Mandrake gets this right every time and it gets right in exactly the way that ESR wants to see. Discoverability of Network printing devices just happens thanks to zeroconf and CUPS. No user intervention is required, so this is strictly a Red Hat problem. I am sure it will be eventually addressed.
I haven’t used fetchmail in a while, but as I remember it seemed like a hackers-only type of tool.
While ESR makes some good points about usability, I don’t know if he’s really one to talk.
The Printer control panel in KDE (invoked through the KDE Control Center or typing “kcmshell printers”) is quite user-friendly.
Fore sure, CUPS is quite complex. I recently had a problem with it, using an unstable package. It had overwritten the “mime.convs” file, which contains associations between MIME types and the corresponding printer filters…it took me two and a half evenings (and two lunch hours…) to find what was going on by scouring the newsgroups and reading the docs. I uninstalled CUPS, removed all remaining files manually and reinstalled an older, stable version, and everything worked.
Was I frustrated? Yes, undeniably. I’m glad that I now know better about CUPS, but I was pretty pissed off. But the truth is that, when it’s properly installed using stable packages, it’s actually quite easy to add and configure a new printer using the aforementioned Kcontrol printer module (or the printer configuration tool that comes with the distro, such as printerdrake on Mandrake).
The CUPS web-based interface is not totally useless, but I still prefer non-web GUI setup tools…
Great article. Thanks for the link!
Eu wrote:
What is the saddest thing about ESR’s rant? That it proves he is using a system that he does not understand.
The point is, there are *many* subsystems on a GNU/Linux system, and the effect of having to hunt around and figure out a poorly-doc’d subsystem is cumulative if they’re all that way.
I was a teacher for some time and am good at explaining things to people — I can tell that ESR is good at it as well. I have a programmer friend who has a lot of trouble explaining things to me. I believe that his problem is a common one: that he can’t think in terms of someone who doesn’t already understand the system (and ESR made this point). I’m not sure my friend knows he has this problem, and I don’t want to bring it up — I’d rather avoid an uncomfortable situation. He’s an awesome programmer, and I know his stuff works.
The problem isn’t that he has this issue — that’s not the point. The problem is that if I ask him to “please explain”, it’s usually fruitless the first, second, and third times through the email exchanges. After that, I decide that it’s not worth making him mad, so I stop asking him to clarify. I suspect that this same dynamic takes place on many dev mailing lists.
The solution is, for devs who have this problem, to realize that they have the problem, and not get frustrated when users don’t understand their intentions. Trouble is, many folks don’t like admitting their faults. Go figure.
So you folks are complaining that CUPS has a web based UI?
Well what about…
Mandrake’s Drake
SUSE’s YAST
RH/Fedora’s Control Center
Come on even KDE has a GUI for CUPS aka Kcontrol.
I understand ESR’s frustration, but I do think the blame is misplaced. CUPS is not the problem. Other distributions work out of the box
Fedora’s implementation is the problem. And this is being generous to ESR, given how unkind he was to the CUPS developers. They are providing infrastructure software. It is the job of the distributor to integrate it well.
Truth is that in Windows, you would have also had to share the printer if you wanted it to be available on the network. Had he done that on the computer the printer was attached to, he would have had the printer show up on his gui dialog without issues.
Hi Eu, just wondering..Are you just “Eu”, or is this “Eugenia”?
He also had screenshots. Like Band in a Box. Ick.
A guy like Eric Raymond can get away with pointing out flaws, although even someone who is a card carrying member of the geek kingdom get snarled at for doing so. However, let a hapless end user who doesn’t program point out the same type of problem in roughly the same terms, and listen to the screeching and cursing.
There is nothing inherently evil about beta testing for usability, but I haven’t seen many OSS projects doing it.
http://darkwyrm.beemulated.net/uidesign.htm
In all honesty, hapless end users almost never specify things in the same terms. They’ll usually say something like “I don’t know, it just doesn’t work!” without specifying what it is, what they are trying to do, or what specifically doesn’t work. Then quite often, they’ll get rude about it in the process, as if the developer is purposefully trying to make their lives harder. Further, they’ll often start making all sorts of crazy generalizations about programmers and how they are unable to communicate with “normal human beings.” If a user acts like that, they deserve all the yelling they get.
No, I am not Eugenia. Eu, means I, in Portuguese, a language whose sound always enchanted me.
Quite aside from CUPS and web admin interfaces et al – I’d like to second Eric in one thing:
Many OSS developers are so wrapped up in matters of code and beta testing and the usual noise generated by people who don’t know how to properly file a bug report that they overlook a very important issue.
Every cry of “I can’t get this to work” is basically a bug report for your interface, and your documentation.
No matter how stupid or arrogant the user might come across: That’s your user, your customer, and your product doesn’t work for him.
Now it’s very easy to sit back and ignore their screaming and complaining because they’re no-wits. And there’s the difference between proprietary software and OSS: In the business world, even no-wits are *paying* no-wits, so you better make sure they’re satisfied.
As long as too many OSS developers rather play with their compiler than listening to their users, the situation described by Eric is unlikely to improve.
“Fit for the desktop” is not measured in how easy it is to breeze through a standard installation. It’s about how easy it is to fix things that don’t work out-of-the-box.
Therefore we should yell at 99% of users because that is how far the knowledge level of 99% of users go. Sure if you want a 1% market share that is a great idea but not if you want more.
There is a dialoge divide happeninng here.
The programmer thinks in his/her world and the user in their world. As a computer technician (the buck stops here type) there is a solution to the problem.
Programmers need to learn computer support skills.
If a user says something to me like “I don’t know, it just doesn’t work!” without any qualifications then I reply by asking very low level qualifying questions until I understand what the is happening to the users machine by just listening to them. They do not need to learn anything about what is going on. The person who is fixing the problem needs to know that.
Examples of low level qualifying questions might be
When did this start happening?
What are you trying to do?
What do you see on the screen?
Explain to me what steps you took?
What exactly does it say?
Which key did you press?
What was the first thing you did?
If they cannot answer these questions then you try simpler ones again. Until you can get the user to start answering questions.
If and only if the user is actually unwilling to go through this process then you can yell at them (but not within their hearing range)
So let me repeat. Programmers need to learn computer support, in order to help users use their program. Becuase only hobbyists do not care about market share and even some of them do as well.
All I have to say is, read it and weep, OSS-ies.
>Who is Eric Raymond to complain about Free Software?
Eric has like everyone else free speech, and can complain about
whatever he wants.
You have to think about what the actual user experiences when he or she sits down to do actual stuff, and you have to think about it from the user’s point of view.
ESR has been reading my thoughts. Thats almost word for word what I’ve been saying for years.
Sadly it looks like a lot of people here have got themselves hung up on CUPS and ignored the larger rant at hand, namely “The point of this essay is not, therefore, just to beat up on the CUPS people ā it’s also to beat up on every other open-source designer who does equally thoughtless things under the fond delusion that a slick-looking UI is a well-designed UI. Watch and learn…” In other words: OSS UI’s almost universally suck, and ESR is going to use CUPS as an example.
One of the worst elements of most user interfaces in OSS is that it almost always breaks the simple idiom: Don’t make the user feel dumb.
“Who is … to complain about …”
The brakes on your car fail in a critical moment, and your car crashes into a tree. Luckily, you survive.
Do you have to be a car engineer to be in a position to complain about the failing brakes?
File a bug report, or look for better documentation. If ESR won’t stop busting on other people’s stuff I’m going to mention fetchmail’s /var/run/fetchmail.pid stupidity (it puts the process number, a space, then 1….so of you do kill `cat /var/run/fetchmail.pid` it kills init.
Every software has things users don’t like. It’s just a fact of life.
@EU: “I” huh? Eu learn something new every day..I don’t know a word in Portuguese, but the lang does sound cool, so I agree with the “enchanting” part.
Anyways, the UNIX design is such (and has been since it’s inception) an operating system that encourages “open-endedness”. It has an immense variety of functions, but with “no defaults”. And some of those functions are “seemingly” pointless..”but one never knows”. UNIX programs are written for the future and portability in mind, and doesn’t dictate to a user how a program should behave. The only thing it dictates is that the user must “learn”. It’s a good philosophy that’s kept UNIX around for 30+ years.
But there are other good philosophies, even if not as long-lasting.
The problem is that to provide user friendliness, especially in the case of GUI’s, the system must dictate — this requires a totally diffrent type of thinking, and less room for user customization. This type of design may be a little second nature for seasoned *Nix programmers, but it isn’t impossible. Look at OS X, or even bits of Plan 9 — partly created by the Father of UNIX himself.
What I’m also interested in is: With Linux slowly rising to the top of the *Nix family, and being the place where this “usability” thing will eventually hit it’s mark sooner or later —
What’s going to happen to the old way of doing things? Interfaces come and go, but the foundation should never be lost. I hope once Linux and others get better at this, they don’t stop “being UNIX”.
Anyways, I hope that made sense.
@ Rain:
I think it comes down to that most programmers simply don’t care enough about others. They program for their personal amusement, not to help others. It’s a reflection of the world today.
> File a bug report, or look for better documentation.
You missed the point.
1) A GUI that requires the user to read a manual has failed its objective;
2) The best documentation available should be integral part of the product.
Uh, damn, there goes the #1 revenue stream for GPL coders – writing such abysmal interfaces that they can actually sell the book… Sorry, couldn’t resist.
> If ESR won’t stop busting on other people’s stuff I’m
> going to mention…
Again, why do you have to be able to build a car to complain when your current one isn’t working? That’s called “ad hominem”, if you care to look up the term…
> so of you do kill `cat /var/run/fetchmail.pid` it
> kills init.
Well, read the manual… š
> Every software has things users don’t like. It’s just
> a fact of life.
But if it’s the user the software is written *for*, shouldn’t that be reason enough to *do* something about it instead of making smarta** remarks like that?
“File a bug report”… you mean, a customer unable to use your product because your UI and documentation stinks should find out whatever bug tracking you use, possibly having to create an account, file a precise account of what is missing – which he probably doesn’t even know because, after all, it would be working if he found out – and wait for the fix?
After your software annoyed hell out of him?
Beta testers file bug reports. Users just go looking for a different product. The distinction between beta testers and users is another thing too many OSS developers aren’t aware of.
If ESR used KDE, he would see that configurating CUPS is easy. KPrinter is an incredible application.
Just look at the way Eugenia gets treated when she points at minor flaws in GUI designs. Those flaws while not so important by themselves are important because they make the complete product so much less polished and makes it look amatuerish. Still people bash her for pointing them out. While I don’t agree with her on all things, she does know her stuff.
Oh, wah! Usibility designs are like assholes: Everyone’s got one, and they all stink!
That’s a little harsh, but the logic is true. Why is one person so much better a designing usibility systems than another? Hell, we’re all people, and we all have to use the systems. So, I ask again, why is Eugenia better at UI design than me?
I’m sorry but he needs to try out Mandrake. It is as easy as Windows to configure a printer. It is also very easy to configure a network printer.. No command line.. no config files.
What I find annoyiong about him is that he paints himself as some uber-l33t Linux-hacker and a programming god (when he wrote the Comparator his comments were something like “This is an existing algorithm, but I perfected it”), when in fact he’s mostly talk and very little action. Yes, he has written some minor semi-interesting pieces of software, but he’s not the uber-hacker he tries to pass as. When it comes to coding, he’s nowhere near the likes of Alan Cox, RMS, Linus Torvalds or Havoc Pennington.
As to his expertise in designing UI’s… These are UI’s made by him:
http://www.linuxjournal.com/modules/NS-articles/HOWTO/6454f1.png
http://www.linuxjournal.com/modules/NS-articles/HOWTO/6454f2.png
http://www.linuxjournal.com/modules/NS-articles/HOWTO/6454f3.png
I rest my case.
That’s a little harsh, but the logic is true. Why is one person so much better a designing usibility systems than another?
Because some people have more talent for it. Just like some people have more talent than others in arts, music, sports etc.
It takes a trained eye to notice those small things, it takes creativity to come up with solutions to usability and design issues, it takes empathy to get into the minds of the avarage user, and it takes a bit of perfectionism to not settle with the lesser.
Not all people have those qualitys. So why not take advantage of the ones who has them?
I think ESR has a very good point.
While the UI problems are probably more Fedora/RedHat’s than CUPS’ I think he hit it right on the head when he complained about the brain/dead docs.
I’m not the greatest coder on the planet by far, and while I sometimes prefer a good UI that “just works”, I don’t mind firing up a VIM session to config things. My biggest complaint isn’t even at the UI level, it’s at the docs level. Often times the documentation is simply missing. Otherwise it’s (take your pick); hopelessly out of date, disguised, hidden, doesn’t address key or important points, or all of the above.
ESR talked about the fact that the CUPS docs didn’t inform him that he needed to change certain settings to get autoconfig to work, it didn’t hint that he had to set the version of CUPS running on his wife’s machine to listen to a particular host/port combo. Luckily for him, he applied his experience with Unix in general and Sendmail in particular to bear. The fact that the CUPS docs don’t cover important points of their program isn’t Fedora’s fault. If he had used Mandrake, SUSE or any other distros config utility, it wouldn’t have improved the docs the CUPS developers wrote. Though I’ll be the first to admit, if they did a good job, he probably wouldn’t have had to consult them.
Just my $0.02 (Canadian, before taxes)
someone247356
If these are the UI’s that ESR put together for fetchmail, he really has no business trying to tell people how UI’s should work. I have a basic understanding what fetchmail does, and I still had a hard time figuring out what some of the options were.
I forgot about the Sun/Gnome connection, but it’s the exception that proves the rule, and in any case, no-one is going to hold Gnome up as the pinnacle of UI design.
True, however, design improves as more testing is done and feedback from lay people are collated, sifted to remove whining, and as a result you end up with a stack of valid complaints.
The problem with OSS software has always been the lack of taking the pragmatic approach. Instead of taking an idea that works well and improving on it, OSS tends to grab and idea, then do the complete opposit in the name of “innovation” and “being different for the sake of being different”, if the majority of people LIKE the layout of Photoshop, why on gods green earth would GIMP look at Photoshop and do their dandiest to ensure that their programme is the complete opposite?
To be successful, it is about taking and EXISTING idea and IMPROVING IT. There are people who LOVE Photoshop, however, their are issues with it, GIMP should make their software look and feel like Photoshop then work to improve this by talking to end users. If an end user goes onto irc.gimp.org and joins #gimp and suggests that a menu entry should be moved somewhere else as it would make alot more sense, that person shouldn’t be kicked then banned from the server/room, they should be questioned further, see how many other people “out there” hold the same view.
All this will fall on deaf ears and the likelihood of the OSS community getting mainstream is pretty slim at best. You can point out rare exceptions but they’re mainly due to corporate sponsorship and not self motivated GUI betterment.
KDE 3.2 solves the CUPS issue Eric describes, IMHO.
re: Rayiner: guide the user.
correct. users need to be asked VERY high level (depends on your perspective) questions. start with “is it plugged to the wall?” if a modem fails to work. if the speakers are silent, ask “are they plugged to the wall? to the computer? are they turned on? see the green light? turned up the volume? see the small yellow speaker drawing at the lower right corner of your screen, just beside the small clock?” etc.
most users have other things to do than hack computers. hacking is YOUR job. your user’s job is operating the local grocery store such that u can buy a coke for those long hacking nights. he don’t have the TIME to read man pages. so NO programmer should expect her users to have the knowledge and/or understanding she has. neither should she expect them to read tons of man pages. man pages ARE for hackers. users use APPLICATIONS. the most they should do is use an intuitive, context sensitive and clear help system.
the terminology is another issue. users speak <put your preffered language here, or at least default to english>. they don’t know what’s lpd, 2.6.3 kernel, compiling, and even as someone mentioned, network. they know “screen”, “that red cable”, “the green button that has ‘OK’ written on it”, etc. applications MUST use user terminology and NOT hackers terminology. because applications are written for users. it’s the developer’s responsibility to use understandable terminology, or else, announce “this application is only intended for people with 20+ years experience in kernel hacking”. but if you intend that your applications is used by regular human beeings, it’s your responsibility to make it usable for THEM.
users should be cooperative with supporters. that’s the minimum requirement for good users. if the user is cooperative, it’s the task of the developers team to understand the issue, decide whether it’s worth modifying the functionality/behaviour of the app (that’s NOT always the case, but for good deal of the requests, it’s a valid), and then act upon. of course, resources, knowledge or interest of the developer might be an issue as well, but that’s another topic.
followup, make sure the issue is not repeated with other users , or that variants or derivatives of the modification don’t harm the usability or other parts of the application.
it’s an ongoing process, and rarely it reaches a state of ‘just perfect’.
re:Eu: unix philosophy.
the modular design of MODULES (read: specific action oriented cli tools) is a very good and proven one indeed imho. BUT in these days of desktop applications, there should be a layer ON-TOP of these ‘perfect’ modules (no mockery intended) that combine these modules into a coherent and usable application. so there’s no contradiction imho between *nix philosophy, and windows/mac (don’t start baching) philosophy. it’s just that they address very different levels or programming/modules. one is hacker oriented, low level, specific action centric, and the other is of an integrated solution, task oriented, USER-FRIENDLY software. and these two kinds of software can be beautifully combined (don’t give bad examples to contradict).
the correct method imho (as beeing used by many modules) is that *nix cli tools provide an independant functionality, with an accompanying lib module that is embeddable in user applications.
then, user apps should only use the correct modules to achieve the desired technological functionality, and then add on top of it all the relevant USER-related functionality (gui, help, wizards, etc).
easier said than done, but programmers should at least be ready to change their perspective, and consider the fact that their users should be able to get away with not willing to invest the time in reading man pages.
cheers
avih
1) A GUI that requires the user to read a manual has failed its objective;
You’ve obviously never used Maya…
I agree with you up to a point. Complex applications will often have complex UIs, and at some point it becomes counter-productive to have “Fischer-Price” UIs. Sure, they may help clueless newbies, but they’ll get in the way of experienced users.
I don’t like to read manuals – but there have been a few apps that, even though they had well-designed UIs, still required me to do so. I think this is a general problem with our society today: no one wants to take time to learn things, we want everything now. It’s as if ADD had become endemic…
regarding my previous post, i may enhance it by stating that there are different levels of users. 95% of all users are the kind i described in my previous post. however, some applications are targeted at experts in <put your target audience>. so the terminology should be modified accordingly.
if a user is not such an expert, he should be given with some starting points to enhance his knowledge.
avih
>Truth is that in Windows, you would have also had to share the printer
This may be true, but at least in Windows, the documentation would politely remind you that the printer needs to be shared by the computer it’s connected to. Poor documentation is just as much a usability issue as a poor interface.
For once I am in complete agreement with ESR. Reading the article brought back painful memories of an afternoon spent doing what he rightly says should be a 30 second task. His experiences absolutely typify the sort of thing that acts as a serious brake on the acceptance of OSS into general user land. Confusing and irrelevant configuration options, undocumeted key administration features and unhelpful Help files compound the problem further. It is absolutely the wrong response to say ‘hey, cups is great on mandrake, or in kde or whatever, because that is not the issue, because the distributors have in effect rejected the standard cupsd GUI and made their own. The issue is that for every X app that has a good GUI, there are dozens others that are frankly abysmal.
I also fail to see what relevance his own past designs have. They may indeed be less than exquisite to look at (some might say plug-ugly) but actually they are very functional, and there’s a great big help button prominently displayed. Firstly, a man is allowed to learn from his past, and secondly some programmes are indeed very complex because they have to be, and no matter how good the GUI there are some things, like email server functionality, that will always require at least some advanced user input. Also, this article is about programmers and GUI designers putting themselves in the place of Joe User, which too often in the OSS world they don’t. Ad hominem attacks on the bearer of the bad news does nothing towards diminishing the problem.
I don’t like to read manuals – but there have been a few apps that, even though they had well-designed UIs, still required me to do so. I think this is a general problem with our society today: no one wants to take time to learn things, we want everything now. It’s as if ADD had become endemic…
I’ve used quite a few apps in my time that have required thick books to understand and use well. Most of those were justified, because the aim of the application was to allow the user to perform complex tasks with a lot of options. AutoCAD, 3DSMax, Photoshop, even some IDEs (perhaps most, but once you’ve learned a couple they get easier to use intuitively). The point, though, is that the tasks themselves are complicated, and the manuals (the exception being the IDEs, though 3DSMax may be another) are written knowing that the applications are complicated as well. Most good books for beginning VB users will spend far more time telling the user how to use VB’s interface than telling them how to use the language, because the language is more transparent than the interface and only requires a small amount of teaching. I used to have a set of AutoCAD manuals that would take up nearly as much space on my bookshelf as books for the .Net framework do currently (which is to say about a foot wide, and that’s considering that there’s a book specifically on dealing with handling XML in .Net that isn’t here and a book on managed C++ that I haven’t picked up yet).
The point, though, is that these are complex applications that allow the user to do complex things. Configuring and using a printer is not (and should not be) a complex thing that requires a lot of user interaction. If I want to hook up a printer, I should be able to tell the program to try to detect the printer, select whether it’s directly connected or on the network, and be done. If it can’t find a network printer, it should be able to explain to me, in a couple of steps, the issues that are most likely to cause a printer to be undiscoverable. If I’m using a GUI to discover the printer and both systems are using the same GUI to handle printer management, I shouldn’t have to drop into a command line or text editor to make the printer discoverable, it should be a simple set of options available on the computer to which the printer is connected.
For one thing, OS X uses CUPS and nobody is complaining about how unfriendly OS X is in configuring printers. Obviously this problem isn’t inherent in CUPS, but in whatever goofy utility comes with Fedora. Second, I set up CUPS to print on another computer running an lpr daemon using CUPS’ web interface. It was completely brainless; I just entered the server type (LPR), print queue name, and host name and was done.
“The issue is that for every X app that has a good GUI, there are dozens others that are frankly abysmal.”
Yup — that happens because Open Source is a meritocracy. Unfortunately the contribution of the docs people, bug testers and reporters are, alas, without merit. That means most suggestions to simplify things and get the bugs out are dismissed by the developers. Only developers want more features; end users just want stable software that works in a consistent manner.
“The rant starts off from a silly assumption. That aunt Tillie would set up a networked printer. Not in her wildest dreams.”
(1)
I don’t know about Aunt Tillie but I *do* know that my Uncle Mike does this (He has two Windows boxes and a HP printer).
I also know that he does this with ease. Connect the printer to the port, open a dialog, click some preferences and it’s ok.
And he’s a regular user. He would not have found out how to do *anything* in CUPS, in the scenario Eric describes. So, Eric is right.
(It’s not the CUPS system that is at fault, though, since it much friendlier Mac OS X, as I hear. It is the design (GUI/configuration) that is to be shown to the end user that sucks.
(2) Forget Aunt Tillie. Even Raymond got baffled with this system. A well know hacker and tinkerer for Christ’s sake!
So, Eric is right once again.
(3) The assertion that Aunt Tillie “never in her wildest dreams would set up a networking printer”, is based on the assertion she wouldn’t WANT/NEED TO, or in that it IS TOO DIFFICULT FOR HER? If it is the second then Eric is right again. If it is the first, then why do you think that?
“Aunt Tillie would not even know what a damn network is.”
Yeah? Then how she ended up with two connected computers? It takes minimal tinkering to do it in Windows, and in Mac OS it is very very easy too. In the “not having to know about IP, eth interfaces and stuff” sense.
A graphic designer the other day connected with Mac OS X to the company’s printer (where I work). DEAD EASY (although he did had to choose “Windows printing” as the protocol, which he didn’t knew what was -he figured it makes sense, since the printer is connected to a W2K box).
So Eric is right again.
“So ESR has picked a very poor example to drive the point home that better user interface design is necessary.”
No, you are wrong. See above.
Therefore we should yell at 99% of users because that is how far the knowledge level of 99% of users go. Sure if you want a 1% market share that is a great idea but not if you want more.
Its not a question of knowledge level. Its a matter of the fact that our culture encourages people to be rude and ignorant. Even if you understand nothing about software, it should be self-evident that if you want help from someone, you should give them a detailed description of what you are trying to do. Even if it is stated in the users terms, it is very helpful to the programmer. Further, it should be self-evident that if you want help from someone else, you should be nice about it, rather than rude.
Consider this analogy: most people know nothing about their cars. Yet, when they go to a mechanic for help, they have the sense to give a detailed description of what’s happening: “its making this kind of sound” or “this light on the dash is blinking.”
The fact that people defend such behavior is particularly amusing. Being a “good user” doesn’t require any knowledge, or skills. It requires plain-old common sense — something that our culture seems to go out of its way to discourage. Of course, that makes sense — people without common sense make much more obedient consumers!
“We are now deep in the trackless swamps created by thoughtless, feckless UI design ā full of glitz and GUI, signifying nothing. This is the precise point at which I decided I was going to write a rant and started taking notes.”
When we, geeks, stop looking at the same direction all the time, or stop rationalizing, we can grow up just like him, that is, we begin to understand that the we are all humans after all.
We can be hard core hackers, but we are users as well. And we attract other users, by the means of our programs, support, etc.
So let’s not pretend that Linux is the perfect solution to anyone anywhere. Let’s not deride others that are having difficulties with their systems.
Just for the record, I lost an NTFS partition in one of my 3 harddisks, and I don’t really know what went wrong. It could be a hardware problem, but I was messing with ntfs and captive ntfs to try to get some stuff from one NTFS partition on one HD to another NTFS on another HD. And the partition that I lost I was to READ from it. I don’t remember trying to write to it. “Se la vie”.
I tend to go straight to the config files rather than bothering with the UI frontends. More often than not they’re more trouble than they’re worth. Even in distributions like Mandrake most seem incomplete, buggy and user hostile compared with similar utilities in Windows.
When I added a PC running Mandrake 9.2 to my Windows network I had no luck with the GUI tools and it was impossible to work out what was going wrong. But after spending a few days reading HowTos and guides I could get it working from the command line. I had a similar experience with configuring my graphics cards for a dual headed display and TV-out, the XF86config frontends I tried did nothing but screw up my settings.
It’s something that really needs to improve if Linux is going to become a popular OS for normal home users. IME most people just give up if the GUI tools don’t work.
It is difficult to emphasize too much the importance of GUI & usability design in the world of open source. The truth is that many open source software GUIs are still quite poor usability-wise.
In fact, I think that the not-so-good state of GUI usability design may be the biggest obstacle preventing wider Linux desktop adoption too, at lest after certain proprietary software monopolies and hardware support problems.
The sad thing is that those who should read stuff like this, don’t, or they refuse to learn.
Examples? Screenshots? Just run your Linux like KDE and you have plenty of examples right in front of your eyes. Though, of course there has been much good development recently too, KDE is getting better gradually, Gnome has long had its HIG etc. But there sure are still plenty of examples of badly designed GUIs in the open source/free software world.
As to his expertise in designing UI’s… These are UI’s made by him
Oh the pain…
Although, in defense of ESR, one need not be a good GUI designer to know a bad GUI design when you see it…
Yes, Rayiner, it’s quite true that, particularly as people use more impersonal forms of communication such as e-mail, it seems to be easier for them to be rude and demanding.
But first, think of things from the user’s point of view. If you are the programmer, though you don’t know the user you may have just caused him hours of frustration. So for him, the interaction between the two of you has already started out negative. Of course you don’t even know there is any interaction, and when you receive an angry message from a stranger it’s quite natural to feel angry in return. You are free to let things happen this way, but it doesn’t turn out profiting anyone. Sincerely asking for details of what is wrong in language that users can understand may, jujitsu-like, turn the user’s energy to your advantage. He is, after all, keenly motivated at that point to help you make your program better (at least in the sense of easier-for-him-to-use).
“We talk about world domination, but we’ll neither have it nor deserve it until we learn to do better than this.”
I think Eric has a point here. If Linux would replace all Windows boxes tomorrow, many users wouldn’t be able to perforom simple tasks, that they usually could (like installing printers, burning CDs etc). And that would be bad.
The problem with Linux is that you need a lot of knowledge in many different areas to be able to perform different tasks. Many users don’t have this knowledge and never will, this is why software has to be ‘discoverable’, to make it possible for less skilled people to use it.
“Apple are the first ones to make a truly functional, user oriented UNIX.”
NextStep.
[Richard james]
“Therefore we should yell at 99% of users because that is how far the knowledge level of 99% of users go. Sure if you want a 1% market share that is a great idea but not if you want more. ”
And who are these people wanting “more” and why aren’t they scratching their particular itch? As someone pointed out over on slashdot, the coders code, and the distro’s polish.
Maybe the “users” are yelling at the wrong thing?
[Anonymous (IP: —.ihug.co.nz) ]
It’s called “scratching an itch” OSS is based around it.
[Solar]
“Beta testers file bug reports. Users just go looking for a different product. The distinction between beta testers and users is another thing too many OSS developers aren’t aware of.”
That’s why distro’s exist.
[rain]
“If all the people who knows UI design would be stuck programming there would be very few of them left to work with UI design. It’s much easier to point out what’s wrong and let the developers who are much more familiar with the code fix the problem. ”
That’s why I think the GUI and the “code” needs to be decoupled as much as possible. Glade is a nice start, but this can be carried much farther. The designers “design”, and the coders “code”. Each accorded their strengths without stepping on each others toes (much). You already see this in website development.
“Aunt Tillie would not even know what a damn network is.”
Yeah? Then how she ended up with two connected computers? It takes minimal tinkering to do it in Windows, and in Mac OS it is very very easy too. In the “not having to know about IP, eth interfaces and stuff” sense.
Last saturday I bought a Mac SE/30 so I could finally try Macintosh networking. And that is really simple. In fact, you only need to connect them using a printer cable and switch on networking.
Not that printing works. But that’s because Epson has not made it’s drivers fit for that.
So, where is the OS where this procedure can be followed to install a network printer:
1. Plug in the Ethernet cable
The printer now automatically appears in the “Print” window, thanks to RendezVous (zeroconf). When needed, you get a Login prompt when you try to print, with a “[X] Save Password” possibility, and a message that you need to have an account at the remote server to use the printer. (Of course the system administrator can also allow free access for an entire IP range, for example 192.168.x.x)
Now that we have the printer access, display access should also work (we have display PDF, right? So if we can remotely print PDF’s, we should also be able to see them remotely, right?) Finally, file sharing should also be that easy. Let’s make a little table:
LOCAL: Screen, Mouse, Keyboard, Local computer
REMOTE: Network disks/printers/scanners
These remote disks can of course contain Applications, so that Remote X is not necessary anymore.
Hmmmm…. I should stop here NOW, as it is not the solution. However, I still think that the entire CUPS configuration wizard should be obsolete. Hook up a local printer and GO!; plug in that ethernet cable and GO on the other computers as well.
It’s called “scratching an itch” OSS is based around it.
Not so sure that this is as true as it was … fame and glory seem to be a lot more involved nowadays
“Last saturday I bought a Mac SE/30 so I could finally try Macintosh networking. And that is really simple. In fact, you only need to connect them using a printer cable and switch on networking. Not that printing works. But that’s because Epson has not made it’s drivers fit for that. ”
And how does the OS handle people who can’t handle plugging in a cable (let alone knowing what an ethernet cable is)?
Don’t laugh, there are such individuals out there. How low do you go when dealing with the public?
“I guess in theory, when it works, its much better, but IMHO, it is a UI disaster in the sense that if it doesn’t work, there is no recourse.”
The funny thing is that your post is “already reviewed”.
How quickly people forget. Remember when Windows would “insist” that you have a particular device? Then it would be “BUT I DON’T HAVE THAT DEVICE!”. Back and forth. The problem with the way autodectection is done is that it takes the human out of the loop, in the name of “protecting the user”. Most of the time one can get away with this assumption, BUT when it gets it wrong and doesn’t allow the human to do what humans do best, then it’s a failure.
Robert Heinlein once wrote that, “Expertise in one field does not carry over into others, but experts often think so. The narrower their field of expertise, the more likely they are to think so.”
Ergonomics, whether hardware or software related, is a specialized branch of knowledge with its own unique set of skills. Technical writing (e.g. help systems) is a separate and uniqe branch of knowledge, with its own skills. Coding is a separate and uniqe branch of knowledge, with its own skills. Design is a separate and unique branch of knowledge with its own skills.
I will never forget once, sitting at the conference table and trying earnestly to explain to a couple of engineers that their tech writing truly sucked boulders. I mean it was abysmally bad. They honestly could not see it. The most astonishing thing to them was my statement that I could merely read the text and tell quickly which one of them had written it. They did not believe it.
Programmers/coders who find nothing unusual in being able to detect a personal “style” in the use of programming language nevertheless have blind spots when it comes to anything outside their field.
The left hand not only dowsn’t know or understand what the right hand is doing. In OSS interface design the left hand frequently seems not to even be aware that the right hand exists.
We need to communicate better. Non-programmer end users, tech writers, we are the communication specialists. WE are the ones who need to concentrate on making sure that the programmers are made aware (in a tactful and respectful manner) of any points where their interface design falls short.
In all justice, you cannot blame someone for being less than excellent in something that is outside their particular field of expertise. It is foolish to expect a non-programmer end user to understand how to write a script. And it is equally foolish to expect a number-oriented coder to be adept at writing explicit text descriptions.
As tech writers, it is our job, more than anyone else’s, to get the message out there about what is confusing/misleading/unintelligible about this software. If, after we do our best to explain it, the prgorammers curse us and refuse to listen the so be it. We will have done our best. But it isn’t really fair to expect a programmer to be good at tech writing, or ergonomic design.
They write programs for a living. We, the tech writers, explain things for a living. It is OUR job to get the point across.
Of course, some people are too busy defending what hasn’t even been attacked to pay attention, but that is beside the point.
Well, ESR is doing something about it, but what have you?
Things you can do to improve OSS
1). You can file detailed, descriptive and well written bug reports and follow it up.
2). You can write a patch that corrects the bug or defects and send it to the OSS project in question.
3).If you can’t write a patch, you plead with an acquaintance who knows how to code to write one for you.
4).You can write an article, review or preview about the project in question and submit is to osnews, slashdot or your favorite computing enthusiast site.
5).You can work towards being a hacker in your free time and fork the project.
6). Be patient and respectful to project developers, maintainers and helpers.
Things that won’t solve the problem in OSS
1).Clamoring OSS sucks at slashdot or osnews at the top of your lungs.
2).Comparing OSS projects with commercial and proprietary products.
3).Insulting voluntary developers and helpers, who might not necessarily be UI experts or even understand what it means..
4).Being an armchair critic. (All talk, no action)
5). Wasting your time replying to trolls at the comment sections at osnews and slashdot.
Food for thought
Empty vessels make the loudest noise.
“Apple are the first ones to make a truly functional, user oriented UNIX.”
NextStep.
OpenStep/Rhapsody/Yellow Box IS Mac OS X, and vice versa. But far more realized — UNIX is optional. Call it Open Step 5.
..if linux developers listen to him,then there may be hope for linux on my dekstop.
“OpenStep/Rhapsody/Yellow Box IS Mac OS X…”
Presently yes, however the question is who did it first?
NextsStep was a fully functional, user orientated Unix (not surprising when you consider who the CEO was).
Some people here might even make the argument that SGI were the ones.
Okay, NeXTStep was like the greatest thing since sliced bread, I’d hate to say anything against it
, and IRIX/Indigo is more friendly than other *nixes as well. They’re both in the “friendly” category, but you still can’t use either one without learning UNIX. I think Jobs’ original intent with OpenStep didn’t fully realize until OS X — I’ll still call it OpenStep 5 =D
It’s UNIX, but “UNIX” is entirely optional — It’s barely even there. That’s all I meant by saying “user friendly”. Someone can make use of every feature in the OS without even knowing the terminal exists. I’d say Be kind of did this too, but it left out a lot of UNIX functionality, though could have easily added them as far as I can tell.
‘They’re both in the “friendly” category, but you still can’t use either one without learning UNIX.’
Yes you can. I used NeXTSTEP 3.3 for Intel for few years without ever touching the command line or learning anything about UNIX.
IMO the UNIX complexity was almost as well hidden in NeXTSTEP as it is in Mac OS X. NeXTSTEP is still much easier to use than any Linux distribution and a lot of Linux fans consider them ready for the home desktop.
I have hard time understanding what Eric complains about.
OSS business model is to give freedom to software users, and charge for services when services are needed.
Of course, it would be at least unethical to give user a bad code and charge for bug fixes, or lie about functionality that is supposedly there but not, or about features that supposedly work but do not. But nothing unethical was done to Eric in that specific case!
Lets review:
1. Functionality (network printer) is there.
2. Network printing works as advertised when properly configured.
3. Eric had got Fedora Core he uses for free and can freely distribute it.
See, it is how supposed to be!
Now, Eric (or his aunt) could just get to the nearest Linux User Group and ask qualified person to set up his (her) network printer for a modest fee of $100, paid for a job done. First of all, it would reward people who learned Linux, second it would support OSS business model.
What I see is Eric trying to get services he needs for free. Hello!!! How are all Linux companies supposed to survive, I do not see Eric buying their stock either.
Sorry, Eric- but you are wrong this time. If you need help- be ready to pay for it, support your local LUG. If your friends and relatives need help you can’t provide for free- tell them to find paid help and tell them where (i.e., local LUG). I am sure $100 bill will do wonders even these days.
‘I am sure $100 bill will do wonders even these days.’
Yes, it would help pay for a copy of Windows so that he doesn’t have to put up with so much bullsh*t just to get a printer working.
Yes, it would help pay for a copy of Windows so that he doesn’t have to put up with so much bullsh*t just to get a printer working.
Or he could have just used Mandrake, which would have detected and set up his printer automatically. Or the KDE Printer control panel.
“Yes, it would help pay for a copy of Windows so that he doesn’t have to put up with so much bullsh*t just to get a printer working.”
Does he get a refund if it doesn’t perform as advertised?