Slashdot Mirror
Open-Source Software and "The Luxury of Ignorance"
Bootsy Collins writes "Using the recent experience of trying to configure
CUPS
on his home network, Eric Raymond has
written an interesting new screed on poor design of user interfaces in general, and configuration interfaces in particular, in open source software, entitled
The Luxury of Ignorance.
A sample quote: 'This kind of fecklessness is endemic in open-source land. And it's what's keeping Microsoft in business -- because by Goddess, they may write crappy insecure overpriced shoddy software, but on this one issue their half-assed semi-competent best is an order of magnitude better than we usually manage.'"
Marketing. People think Macs don't work as well - not as much software, etc. It's just marketing - that's what Bill is good at, after all :)
That's not necessarily true. Mandrake set up CUPS and just about everything else I've needed with no problems at all. It's all about what you're doing. For some programs under some distros you need to be a programmer to install and / or set them up. Under other distros, and with other programs, it can be a breeze. (Just look at how well Knoppix does!)
I don't think its marketing as much as its lack of marketing by Apple. Sure, they are flooding the airwaves with iPod and iTunes commercials, but they have never run a commercial showing what OS X is capable of. Or iLife. Most people I talk to have no idea when it comes to Apple. They are amazed at how well the software is integrated together and that Microsoft Office can run on a Mac and that they can surf the internet as well. I get so tired of doing Apple's job for them. I really should send them an invoice for all of my PR work.
It's true that the OSS community needs to beef up many area of the developpement process.
Software isn't just about the code the same way that a car isn't just about the engine.
For people to want to use it in the first place, to enjoy it once they've started using it and to stay with it, a "product" needs many qualities.
This (often) explains why an inferior design can becomes the norm.
So lets get cracking with artists, GUI/interface designers and and documentation writers!
I will anticipate the "Well, why don't you do something! Where's the patch?" posts and answer:
I'm doing what I can with the talents that I have (often amounts to writing suggestions to developpers, bug-reports, spreading the word on new stuff and donations).
Treehugger? Treehugger... Treehugger!
Either I can take his side and be called an idiot because I'm sure someone will claim to have an easy solution to my problem. That's what someone claimed the last time I mentioned I couldn't get MPlayer working and then of course the suggested solution didn't work. Or, I can stay out of the discussion entirely. I think I'll do the latter instead.
Every year on slashdot people say it'll be Linux's big year. Yes, that means next year people will say it too. It's partly because of this thinking like yours, that you need to be l33t to even touch the machine, that linux's big year hasn't happened yet. You follow "What is a non-technical user doing with Linux anyway?" with "I like to think of Linux as a sort of technical boot camp." So which is it? Is Linux the end-all of nerdom, or is it just an educational experience on the way to... what?
The point is that a better UI isn't something that should be frowned on. Christ, I feel stupid for even having to say that.
If you get nervous, just remember that there are a few billion other people who don't really give a damn.
I don't know a general answer to your qauestion, but here's an anecdote. The first time I used a computer was in 1990, in a computer lab at a university in the US where I studied for some time. I needed to type up an essay, and I had never before so much as touched a computer keyboard.
I entered the lab. To my right, a bank of smaller, friendly-looking Mac Classics (but I didn't know what they were). Menus, icons, mice. To my left, a bank of foreboding but somehow more powerful looking IBM ATs. Green screens with text-mode commands, one of which would launch WordPerfect 5.0. I had to make a choice, and a completely uninformed choice, mind. In really had no idea what was what there.
I picked an IBM. Someone instructed me to press F3 for help and F7 to exit. I took it from there, and loved it. By the time I left, I must have known much of WordPerfect's help system by heart. I did try the Macs once or twice while there, but I went back to the IBMs every time. I wish I knew why, but I don't. Maybe theys looked more serious, more powerful. Maybe they adhered better to my uninformed mental image of what a computer was supposed to be like. Today I can list all sorts of reasons why I prefer one to the other, but it's mere rationalizing after the choice was made. I guess Macs looked too much like toys to me, while those text-mode DOS screens looked inscrutable, and hence they looked fascinating.
"Only the small secrets need to be protected. The big ones are kept secret by public incredulity." - Marshall McLuhan
1 What is a non-technical user doing with Linux anyway? They need to crawl before they can walk.
:)
well, if you LATFA, you see as the second sentence...
It has proved a textbook lesson in why nontechnical people run screaming from Unix.
IOW, if you want to even think of competing with the windows world at the desktop level, you actually have to reduce to the brain-dead level of explanation, support or general UI practice.
Even technical non-unix people struggle (a manager at work, skilled with Novell (stop laughing) is struggling a bit to learn linux.. and deadrat at that). if semi-competent people have some semi-major with what we, the unix-versed, understand (but may still be tasked by on occasion) how can we ever seriously expect Linux to prove its superiority at the joe-schmoe level?
-'fester (aix/tru64/hpux/linux geek.. that's in paying order, mind you
-'fester
The kernel configuration system back in 2.4 was crufty and not very user-friendly. So Eric decided to build a new system, CML2. It ended up not going in for a variety of reasons, not the least of which was probably a lot of people don't like him all that much. However, in that case he was practising exactly he is preaching here - making software easier for non-gurus to use.
Any sufficiently advanced technology is indistinguishable from a rigged demo
--Andy Finkel (J. Klass?)
This is especially true if its a non-trivial piece of software. Several times new programmers have come into software packages I've been working on, don't bother to read the structural documentation or even the useful other code that serves as examples for how to improve and extend upon the existing structure.
Instead they try and do things their own way, often end up doing things redundantly or breaking something else and just otherwise fouling more than they contribute.
The best person to improve upon software is the person who designed in the first place! Or someone who's worked on it extensively enough to know the quirks, the reasoning behind non-obvious parts and knows the rest of package throughout.
Telling a user to fix a poor piece of software is incredibly frustrating and lame to those of us who, god forbid, have other things to do in our lives.
-
I've held a fairly obvious view for a long time with regards to interface design (be it computer or otherwise):
;)
Unless you're working under a predefined framework, chances are, your design is going to differ from someone elses when you both attempt an identical solution.
This isn't an answer on how to deal with this issue, as the answer(s) are everywhere, it's more of a thought process that keeps me from going crazy.
How many times have you worked with a piece of software or hardware only to move on to another one that was similar in concept, but totally different in execution? It's gotten to the point that I've stopped trying to become an expert at everything, and simply want things to work (maybe I'm just getting older, and have less time and/or memory).
Maybe that's why companies like Apple have a strong following, with a mantra of "it just works".
The next time that Joe Administrator is getting cocky with "oh, you didn't know how to configure file XYZ for ABC", remember, they're just being programmed to use an arbitrary interface, thought up arbitarily by some designer.
And that folks is why I'm working to get out of System Administration, and into programming
[end rant]
Q: What do you think about American Culture?
A: I think it's a good idea.
(adapted from Gandhi)
A similar diatribe to ESR's could be written on trying to burn a backup DVD under RH9. Gave up; I just FTP my backup over to my Lose2003 box, where the driver worky-worky.
Get thee glass eyes, and, like a scurvy politician, seem to see things thou dost not.--King Lear
Why? Because he's a USER. Not a programmer. Developers have a responsibility to listen to their userbase. If you want market-share, then when your users say "I don't understand X", you DO NOT say "well, FINE, fix it yourself!" That is ENTIRELY the wrong attitude. ESR may be confrontational, but you're even more so.
Why doesn't your approach work? Because they're simply going to walk away. Software is so complex these days that many people, even programmers, couldn't possibly contribute without investing a serious amount of time. Hmm, which is a better use of resources- 12 hours of a user messing around learning your functions, conventions, library calls etc(and probably introducing more bugs than features)- or 15 minutes for you to add the button yourself?
I know -exactly- how he feels. Countless times I've found software that has a super-spiffy web page, touts how damn good it is to anyone who's reading- but you unpack the source and Jeeeeesuschriiiiiist you can't figure out which way is up- and I've been building and compiling unix packages for almost 10 years(when i was yer age, we had to edit makefile library paths ourselves! None of this automake...) Then, if you get it built, you run it and menus have confusing names, there's no help file, there are secret options nobody mentions that are in the ~/.myprogram directory, and so on.
The mldonkey p2p client was an excellent example. The developers continuously worked on all sorts of weird theoretical schemes for this and that, while the userbase clamored for a manual(there was none), a description of what each setting did(ditto- the developers would cheerfully add some oddly-named option and not explain to ANYONE what it did), or for features that were common in other clients. Such as the ability to share a file without having to restart the client(shocking!) But hey, you got three different algorithms to pick from for how it managed sources for files. Yaaaay!
Please help metamoderate.
My recent experience was trying to print to an inkjet connected to a windows machine. Since it was remote, I decided I didn't need a spooler, so I didn't install cups. Instead, I found foomatic, which is supposed to cut through the many layers of drivers in one slice. Through no efforts (reading several confusing and inconsistent tutorials) could I get foomatic to produce a file in my printer's format. Nor did it give me intelligible error messages. I finally posted to the main list at linuxprinting.org (lp.general); but in the weeks I've been subscribed, I've not seen a single useful reply to anyone's question!
Oh, I finally got the printer working. I just have to run gs -DSAFER -sDEVICE=ijs -sIjsServer=ijsgimpprint -sDeviceManufacturer=EPSON -sDeviceModel='escp2-c82' -sOutputFile=out -DNOPAUSE -- file.ps , and send the result with smbclient.
The evaluation of an action as 'practical' . . . depends on what it is that one wishes to practice.
... a decent try at best. At first glance it alienates me a _LOT_ less than lprng, which is fully managed with an arcane /etc file that lists configuration directives in no particular order.
But that doesn't mean that CUPS is all peaches and roses. I had to discover what `foomatic' was in order to figure out how to extract a driver for my Epson Stylus C42UX from a large xml file. Its wizard to create the printers was rather friendly, although a belaguering dropdown box full of stuff I didn't have asked me where my printer was. Luckily it identified itself as USB PRINTER #1 (EPSON C42) so I could choose that - but most wouldn't have the slightest idea of what to choose and just stare at the screen glaze-eyed...
Really, all I wanted to do was print a school assignment. I fully agree with esr on this issue. This whole CUPS ordeal should have taken me 10 minutes, not 10 hours (on and off) to get working. And it still doesn't fully work, for example with printing to a SAMBA host.
But CUPS is the best we've got for Unix now. Isn't that sad?
It takes a real man-geek to admit "issues" when installing new software or configuring devices. He loses points for his longish rant though.
/etc/cups/cupsd.conf file.
However, I found myself nodding in affirmative at EVERY single step he took during his trouble shooting. I made a lot of the same assumptions (wrongly). The funniest was when he finally figured out he had to configure the server machine to broadcast, and then he couldn't connect to it. HAHA, it took at least 15 minutes of loud swearing for me to figure out how to configure the &*#&#((#&$&^
You know you're in trouble when the first like in the man page is RTFM.
I swear, if I have to configure another CUPS network, I'll go postal. It works... ssssh, don't touch it, and speak in hushed tones when in the vicinity.
Toddlers are the stormtroopers of the Lord of Entropy.
Being pretty doesn't make it USABLE, and being ugly doesn't make in UNUSABLE. I was messing with fetchmailconf one day and had everything configured rather quickly. I had no previous fetchmail experience going in, and was pretty new to Linux in general. Usable, but ugly.
'Standards' in computing only impress those who are impressed by things like 'standards'.
Sorry, this is one that Eric Raymond should have researched a bit more. Not because the interface he's talking about is any good, but because he's firing at the wrong target, as others have pointed out. I wouldn't expect someone who doesn't know Linux to figure this out, but Eric should have been able to tell the difference between a Red Hat hack and CUPS proper (at least the localhost:631 web interface).
While I haven't used it myself, the number of complaints about it on the linuxprinting.org forums (vs. the lack of complaints about Mandrake, SuSE, etc. in this regard) suggests that there's a problem. From my standpoint this is a real nuisance, since a lot of the people blame Gimp-Print for their problems (reasonably enough from their perspective -- I don't blame them for that). However, ESR should know better, and should be able to pick his targets more accurately.
Recent ASUS motherboard in computer at work. Plug a device into one of the soundcard ports, brings up a dialog box that checks you've plugged when you have into the right hole, and gives you option to work the way that you've done it, or to unplug and reuse the correct hole.
Saved pulling the computer out from under desk as I accidently used the wrong hole (found it by feel) then I knew what I'd done wrong.
All developers miss the point in my experience. However, I think ER is shooting off at a straw man applying this solely to the OSS community.
My recent experience with Mandrake 9.1 and 9.2 on two computers were an exemplorary experience in point and click installs. DVD Player, digital camera, modem, video card office network, printer, you name it it all seemed to go. Email, office tools spell checker (non-US non-German) works.
I am sure I could have gotten into trouble if something had not been recognised or I hit the wrong button or was trying to get a printer server running. In that case I would have done what I used to do when Windows stuffed up - asked an expert.
-- Free software on every PC on every desk
Yes, the fact that the dialogs don't follow an established STANDARD does hurt the usability, but I don't think that they are BAD.
Yes, that is pretty much the definition of a bad GUI program.
I live in a giant bucket.
Please just remember that part...
:
:
I happen to have recently installed a Laserjet on my gf computer, and it's Win2000, and the whole process took me 5 minutes (1 config failure, 5 seconds of intensive neuronal action and then the right click on the right button)
I simply used KDE printing tool that came with the nice Knoppix-Cluster cd, and took 5 seconds before hitting buttons.
Also, please remember
COMPUTER WEREN'T MADE FOR PEOPLE !!! Computers were made for experts in companies, the fact that windows is "easy to use" (damn, it hurts !) or even "intuitive" (I actually wrote that ?) has been the main cause of problems, because the configuration was a "One-Size-Fits-All" solution.
=> Most Windows computer are configured almost all the same, default, and so more or less all exposed to the same problems. They work "perfectly" (my hands start shaking) as long as everything is in the "Normal Scope" (everything open and accessible from anywhere, except if you change it, which users don't)
=> Microsoft made 2000 and XP. One is clearly a server Os, where even access to cdwriter for users has to be configured by hand. Many things are accessible, but you have to RTFM a bit and you can get it almost secure (MS notwhistanding)
XP, on the other hand, is a nice "plug-and-play" thingy with lotsa grease and help so that even Aunt Milly can do it herself (or pester her nephew/son/grandson, as in the 99.99% of real life cases)
You want an easy to use OS ? get a playstation.
You want a desktop computer that just works ? get XP.
You want a hard, rugged and stable server ? get linux.
You want a nice Linux desktop easily running in no time ? be ready to lose most of your security, or wait some more time... MS had 20 years to learn how an UI should look, and they do extensive usability tests, have specialists, teams, and so on dedicated to the problem.
It will come in time, but Linux wasn't thought for the desktop, so the transition will take some time. The poor guys making cups did an excellent job as the server works 100% (for me). If you dislike the UI, please follow usual Open source procedure
1 / Email the dev and tell him (gently) what's wrong in your opinion and what should be done. If he has the time, he'll fix it. (99% of real life cases ?)
2 / DO IT YOURSELF AND STOP COMPLAINING FOR CHRIS'SAKE !!! you are a guru Linux wizard, so get emacs runing and do your conf files, or write a better UI.
Ahh ! No point in this post, but I somehow feel better 8)
Linux is about choice and RTFMing : always had, server-side, never will, desktop-side...
If Users knew how to do it, they would be sysadmins...
It takes 40+ muscles to frown, but only four to extend your arm and bitchslap the motherfucker
I too have struggled through a configuration of CUPS, coupled with samba printer sharing for windows users no less. A couple weeks later, when OS X 10.3 came out, I was amazed at what Apple had done for a front end to CUPS. It's extremely intuitive, and a vast improvement to previous OS X printer configuration schemes.
It would be really nice if Apple's config utilities were released back to the open source community.
Hence why a gaggle of volunteers can put together and enterprise-worth OS in their spare time.
Unfortunately, pure R&D is never that clean. You often don't know in what direction a new technology is going to take you. In WWII, the answer both the Axis and the Allies had was to simply fund everything that had a glimmer of a chance, and research everything in parallel. Sure there were a lot of failures, but you also got a lot of radically different and paradigm changing designs. It is the era the brought us Jet powered aircraft, RADAR, cruise missiles, liquid fueled rockets, nuclear weapons, SONAR, and electronic computers. And that's ignoring massive new understanding in industrial production, chemistry, and materials.
When designing something new and unprecidented, you have to play the field and try alternatives. More productive than a complete fork would be to simply try an idea at a time, and fold the best of breed back into a common reference build.
Oh wait, the Linux kernel guys already do that. The wiley hackers!
"Learning is not compulsory... neither is survival."
--Dr.W.Edwards Deming
Yeah, when you have "computer people" who don't want to take the time to mess with Linux cause it is a bigger pain than windows to get to do anything, then you can understand why the average user won't use it.
I'm a pretty computer savvy person. I'm in my 4th year of college in CIS, and I have taken Sysadmin classes using Linux exclusevly. I have bulit my own computers and computers for other people. I've done networking with routers in Internetworking Classes. I've even done some programming.
Compared to the average user, I'm the person they come to when something breaks on their computer.
I loved Redhat Linux for running DNS servers in class, it was great for a mail server or FTP server. It was great for scripting. It was fast, and stable - the multiple user features were head and showlders above home windows offerings in my experiances.
I don't use Linux at home. I've tried, multiple times, multiple distros. It is simply TOO MUCH TROUBLE. I don't want to fight through dialog boxes that don't seem to do anything after I hit apply. I don't want to deal with install issues, like how do I install this today? I recently played around with mandrake 9.2 I believe. This time I didn't want to totally dual boot etc, so I was using VMWare. I don't know if this is a VMware problem, or a linux problem, but let me tell you - Windows 98 virtual machine... click on file in VM ware, and install VMware tools... bam standard windows installer in the virtual machine, and bam, done, installed. I still haven't gotten the linux script to work right. I've given up. I've since heard that maybe I don't need to install that anyway cause newer versions have automatic support for VMWare.
The point of my rant there is that until software vendors and developers come up with a clear consistant UI, with things like install programs that you can double click on in KDE and have work,I don't see linux catching on on the desktop.
The sad thing is I like to play around with linux - to keep up with what's happening, and to stay in *nix mode for servers.
But when I need to get some classwork done, like write a paper or do a spreadsheet, or when I want to play online - I use WinXP. It's just easier.
Opera, Proxomitron-Grypen,GPG 0x0A1C6EE3
More fun : Interface Hall of Shame
But of course that doesn't sit well with those who have an agenda to get Linux to fight their crusade. Or even worse to get them to not to have to pay Microsoft anymore.
But it is a sign of the time we life in. Give someone dying of a heart attack in the street CPR and they will sue you if you break a rib. Write an excellent printer sharing protocol and people will only bitch about how they need to read the manual.
Opensource doesn't just work with developers on one side and users on the other. If it is going to work then we need manual writers, forum guru's, gui designers, beta testers, patch submittrs.
Users are like customers. MS loves customers because they pay. Opensource is free. What do we care how many customers we have? 1 * $0 is the same as 1000 * $0 but it costs a hell of a lot more to have 1000 people asking stupid questions.
Rant: Old saying is there are no stupid questions only stupid answers. This was true before the invention of the net. Read any forum and you will see time and time again the same question being asked because the asker can't be bothered to first look. Then they will bitch that noone helps them. Obviously their time is more important then everyone elses. Recently saw the worst of all. 9 pages down a ***** said "I am not going to read all those pages give me the answer". ARGH!
End rant.
MMO Quests are like orgasms:
You may solo them, I prefer them in a group.
That's still a lot better than CUPS. At least you didn't have to become root, read files in /etc, guess which options to change, and restart the CUPS server from the command-line. My experiences with CUPS, and Linux printing in general, are the most harrowing I've had with computers.
Don't drop the soap, Tommy!
here's my experience with setting up a HP Color LaserJet 4600:
win 2k/XP - find way to add printer->select local printer->turn off probe for PnP printers->create new port->select standard tcp/ip port->enter printer ip number->click custom device type and then settings->click raw protocol and enter port 9100->enter printer driver info->click a few more next/finish buttons->print test page
linux (RH9 & FC1) - go to system settings->go to printing->enter root password->click forward->enter desired name and description->select networked jetdirect->type in printer hostname->click on printer manufacturer and then model->click finish and then print test page
OS X - go to the printer configuration utility and find the printer already detected, configured, and set to be the default
sure, the linux config could be worded somewhat more intuitively, but windows is a complete disaster for any non-SMB networked printer. the whole having to select 'local printer' to do it is just hysterical. at least linux refers to it as networked.... my only real niggle so far with the RH/fedora printer config tool is that the sharing properties are hidden under the Actions menu and it doesn't let you configure sharing on a per queue basis.
that all said, the rendezvous support in the HP printer is pretty damn sexy. any mac on the network sees it automatically and understands everything it can do. that's the way it's supposed to be. once i enabled the printer's CUPS support, then the linux boxes were almost there, too. poor windows users still need to go through that long drill, though....
tim
hiding in shadows / i hear you coming closer / you will explode soon -- a quake haiku
"If they were really smart (like, say, Mac programmers) they'd leave the impossible choices in but gray them out, signifying..."
Greying out menu items is one area open source can actually surpass Mac OSX and Windows. When I try and use a new desktop app I have never used before I am always puzzled why some menu options are greyed out. Everything else I find intuitive. Greyed out items confuse me.
Why is is greyed out? How do I get to it? Why can't I get to it now?
What would be really nifty is some tool-tip text saying something like "This menu item is only available when you are in xyz mode."
Am I the only one who experiences this difficulty?
OSDL is possibly the one organization that could help various projects test their usability. It would be expensive, and we would really need to squeeze large corporations like IBM, Novell/SuSE and Oracle to provide some serious funding. Perhaps the reason they don't is because the developers could still simply ignore suggestions or demands for improvements. Of course, perhaps the major distributions could choose not to include programs that don't meet some minimal level of usability or conform to one of the guidelines listed above. That might provide some incentive.
I'm a grain farmer. I don't know the first thing about programming anything and didn't have a computer until Windows 95. My tractor was built in 1967. I'm not even close to being a techie kind of guy, but I had zero trouble getting CUPS going and I kind of liked the GUI setup tool.
Actually I've never had any trouble getting printing going on linux and I've been using linux since RH 4.2. I never did upgrade from windows95, don't like it. Might boot windows two, three times a year now max.
We blow this stuff off because we want to make it workable for those smart enough to deserve to enjoy it then quickly move on to the Next Great Thing that Needs to be Made Now. We Peter principle ourselves out of making a real headache for MS, which is something we (ostensibly?) want.
... and I'd imagine many other OSS coders feel the same way.
Speak for yourself. The (open source) code I write is written for first and foremost for myself. I'm open to suggestions and feature requests, and even more so to patches, but I'm not going to go out of my way thinking about how to make it fit to the lowest common denominator of users.
Note: it's not because I'm trying to specifically exclude stupid users, it's just that it takes a hell of a lot more work to create a dumbed-down interface, and that these type of interfaces often make things slower
Mind you, I should also add that I have never had the aim of "making a real headache for MS" when programming, and I think that that is a terrible reason for writing code.
I have some quibbles with this. Mostly it is correct, but I think that it needs fleshing out. First, Command line. Half the time I need to poke around in the man pages to find out what the command I want is called. And when the program spits out an error, like 'you must specify --sign for raw data type' (happened to me today, flac, never used it before, piping data to it from sox) you need to poke into the man pages to learn that --sign is either signed or unsigned , not yes or no etc. Printing all of the options in the error can make error messages way to big. I think the answer here is: command line programs are not for the non-technical, and they are not 'esoteric' Opinions welcome.
Second, frequently, things are too complecated for the interface to be the full documentation. CUPS is actually a good example of this. CUPS does a lot of stuff, and has tons of options. The interface would look more like a man page than an interface if you took this too far. What the interface should do, and this was mentioned in the rant, is lead the user to the parts of the documentation that he needs to read to do what he wants. And it should be clear from the start just what the user needs to do to sart doing to do the job. CUPS miserably fails at this as the rant points out. Flac did not fail. I had to read the man page, but it told me specifically what I needed to look at. (does the author consider the help screens that CUPS gave to be part of the interface, or the documentation? assuming they had helped of course!)
On the whole, I agree with this article, but this rule needs some qualifiers, because as stated, it will make GUI interfaces unusable for anyone, including the non-techie people.
I ran into the same issue with CUPS, and never got it to work.I gave up and just print to a .ps file, and use smbclient to actually print it! Maby with these pointers I can get it to work.....
"We blow this stuff off because we want to make it workable for those smart enough to deserve to enjoy it then quickly move on to the Next Great Thing that Needs to be Made Now."
This has to do with psycological personality types. google for INTP and INTJ (the two most common geek types) INTP in particular gets bored with something as soon as they have it figured out. ( I am an INTP BTW)
Laws are horrible moral guides, moral guides make even worse laws.
The truth is, that despite the intensity of the point, the point still holds truth. I love linux, I have used it for almost 10 years now, and have done everything from kernel hacking, to my own C programming etc.
However, what began as an enthusiasts project, became an essential part of my work, has now become to some degree tiresome, and laboured. It's simply because binary distribution and configuration designs between OS's varies so much, that it becomes difficult to release software that easily integrates into ANY environment. Permeatations on OS's means many more for the software.
However, it is up to successful programmers to fix this, and trust me, if it can happen, it will happen with linux, and open source, if not demonstrated by the current wave of self booting, nice looking Linux distro's, the installation menu's these days, etc....sure it needs more work, but it will have it shortly.
Just think of the next wave of Linux Distro's in 12 months time, how much easier even still they will be to use, install or download software.
Now imagine 24 months.
Now compare that to Longhorn?
Microsoft knows it's coming....
It would be really nice if Apple's config utilities were released back to the open source community.
If they just gave it back, then what would be the point of owning an Apple computer? By creating the extra value, they are able to charge a pretty penny for it and justify their existence. For that part of it, at least, they are fulfilling the promise of open-source: a level playing field for everyone that they add their own particular brand of value to.
To celebrate the occasion of my 1000th post, I will post no more forever on Slashdot. Goodbye.
A similar diatribe to ESR's could be written on trying to burn a backup DVD under RH9. Gave up; I just FTP my backup over to my Lose2003 box, where the driver worky-worky.
No, no! The driver works *perfectly*, it's just that it requires correct entry of hardware parameters in one of the assembly language sources! Yeesh! Don't blame the hard-working open-source developer for your MCSE-like lack of computer knowledge!
Seriously, though, I'm so glad to see ESR ranting about the state of userland GUI stuff. I've been doing it for a while, but it's often dismissed as a FUD campaign by people who don't like what I'm saying.
Fire and Meat. Yummy.
He was obsessed with the Macintosh being a work of art. He was so picky about the look of the damn calculator app, the designer got tired of revising it and made a calculator interface designer for him. The final design Steve made stayed with MacOS up into the 90s. He even had the Mac designers sign their names on the inside of the mold for the casing. That's a mentality I like--the connection between emotion and computing. The creation of a computer that blendds into someone's life as a useful tool and portal to computing.
What happened to that melding of art and computing? OS X still has it, but without support for x86, it's not exploding like it should. That leaves Linux--and Linux is completely missing the ball here because it's been written by developers for developers, and still is. It's massively technical and powerful for dev-heads, but the other front--the one that Windows lacks--is the intuitive, artistic side.
But, I fully expect everyone to stick with crappy XFree86 for another 10 years and espouse how great their poorly designed "KDE" and "GNOME" interfaces are. Five years after Longhorn comes out, KDE will finally get around to attempting hardware acceleration and also speeding up the horribly shit-slow app-loading.
Nobody's artistic about computing anymore, except Apple. We should be too. Obviously, that means rethinking the way people are writing their apps/environments, which ain't gonna happen.
I ran into this recently in Debian too. It seems that the open source software just doesn't quite cut it for my TDK drive. I could write cds fine with cdrecord, but it would fail with an "unrecognized media" or some such error when I tried to burn a dvd. I was using dvdrecord (based on cdrecord), but it doesn't work with a lot of drives. The really frustrating thing is that cdrecord supports writing dvds with my drive, just not in the opensource version. You can get the "pro" version here it is free for non-commercial use and it works well (there is a readme file with more info).
By not writing software for non-technical users the so called 'digital divide' widens so much that we are no longer 'in danger' of creating a digital underclass, but we are guarenteed of creating it. An interface to a piece of software should be elegant, simple and intuitive. If the ATM's we all use had been more complicated than remembering a 4-digit pin and pushing a plastic card into a slot, then the every-day consumer wouldn't use them.
Fine, then don't bitch about "M$" being the dominant monoculture when you're to lazy to bother making your software usable for other people. If you only want to scratch YOUR itch, keep your software on your private network and don't let major distros pick it up. Understood, Mr. I'm-the-poor-unpaid-volunteer-developer?
Guess what? Users don't care either. They'll drop your shit like a bad habit and go back to what works.
I installed a GNOME desktop for my wive. She has never used a computer before and just wants to use EMail and WWW.
The GNOME desktop works perfect for her. There are no obscure things that could mess up the system and she immediately was able to work with the system. I had explain almost nothing to her.
Before I installed GNOME for her, I had a look and KDE and found it terrific - from the UI point of view. Menus loaded with items rarely used, bloated toolbars and the like.
Afterwards I had once installed WINDOWS XP on a square patition to run a specific program which only runs on WINDOWS. She had a look at it and - immediately wanted GNOME back. She could not stand the "Lunar" style of XP and also complained about this tiny task bar at the bottom.
I find that currently there is nothing better than GNOME for people who just want to do some specific tasks with the computer and who do not want to bother with configuration and tweaking.
While his style is, as usual, not quite professional, the points he makes are right on target. Usability is sorely lacking in most Unix/Linux setups.
But instead of pointing to various short user-friendliness rants and mini-howtos, I suggest reading a few books, to see what the current state of the art is.
I suggest the following two, which I am using for my thesis work on this subject as well:
Donald A. Norman: The Design of Everyday Things
This book focuses on everday gadgets and appliances instead of computer interfaces, but the advice Norman gives is perfectly applicable to our field of work. Highly recommended.
Alan Cooper, Robert Reimann: About Face 2.0: The Essentials of Interaction Design
Now this book is pure gold. Excellent advice on user research, goal-oriented design and lots of insight on GUI design as well. Yes, Microsoft gets some praise for parts of their efforts - where they deserve it. They also are criticized properly - just like everybody else - where they failed. If developers would apply at least a little of this stuff, we would have vastly better software.
There was earlier article that compared *nix and Windows programmers...
*nix programmers write programs for other programs to use (hence command line arguements that are easy to parse/create etc). Ie they do the guts first, then bolt on an interface later.
windows programmers write programs for users. ie they write the interface first, then the guts.
Would be interesting to see how the Mac guys concentrate their efforts.
Check this, a fairly academic review of the usability of open source sofware.
Ah, I see you have discovered KDE's design guidelines.
That's not funny, it's sad... In a fresh kde 3.2 on my dell desktop, I have an extended "Sony Vaio Laptop Hardware configuration" panel in control center. Wtf? It is not only completely unusable, but KDE is even aware of that, as every option is grayed out and there's a complaint about a missing driver.
Maybe OSS needs a central repository of OSS development "best practice".
A collection of technical howtos (subversion/cvs, patches etc), articles on UI design, documentation writing, managing distributed volunteer teams, handling users. Things like "Dos and Donts", articles from experienced OSS developers and users - maybe a little less inflammatory than ESR's, though. A Wiki maybe?
All this information must be out there, distributed in mailing lists, forums and developers' memories. Surely it would improve OSS quality if new developers sent a few hours reading through that sort of material before starting to contribute.
While I certainly agree with you that following standards may not lead to a good user interface for all applications, I would submit that (at least for end user applications on mainstream PCs) it is usually better than not following the standards, and that most attempts to "innovate" are usability failures. To wit:
This is true. And as a professional developer using Visual Studio .Net, I'd like to thank Microsoft personally for giving us:
and all the other "innovations" that cost me several minutes of my valuable time every day.
To their credit, Microsoft's developers (at least those I've talked to) do seem to have a genuine interest in improving this, and their hearts are in the right place. Some of the nasty context-sensitive stuff can be disabled in the 2003 version, for example. But a lot of these "usability innovations" gain me nothing, while slowing me down and/or wasting valuable screen real estate.
If you disagree, post your argument. (-1, Overrated) isn't your personal censorship tool for views you don't like.
Software designers need to realize that the clueless users aren't the only ones who have trouble with the software; plenty of intelligent programmers and sysadmins are just as screwed when it comes to configuring complicated stuff like printing, sendmail, and Apache. The problem is that there are just too many options. Sure we can read the documentation, when we have the time, but there is rarely time to read the hundreds of pages of documentation that go along with a lot of really complicated software packages. It's not that I don't appreciate the flexibility that all the nifty features provided by large software packages, it's just that I rarely use most of them, and don't have time to sift through the documentation.
If you want to add cool features to the software, go for it. We'll love you for it. But if you want us to actually use the software, stop every time you add a feature, and make sure that you are providing simple, straightforward, easy-to-find documentation, or create a nice GUI. Otherwise I'm left with the options of sifting through my book collection and google results or just using Windows software with a nice automatic setup wizard.
Some projects still do not get it. There is an unconscious assumption that merely wrapping commands in a menu/GUI is making an app user friendly.
The biggest culprits seem to be projects that port command line tools to a GUI like emacs, x-cdburn, and oracle's sql interface ( not oss ).
One of the ways a GUI makes an app user friendly is if the GUI takes away some of the need for knowing how to do something in the app.
X-CDBURN ( name? ) is a good example of this. It is GUI, but the user still needs to know how to use the command line tool commands in order to burn a CD.
What is the point in wrapping the command-line tools in a GUI then? Those sequence of commands could just as easily be typed into a shell without the overhead of the GUI.
In contrast there is K3b where a user can burn a CD without having to read a HowTo to learn the theory/practice of making CDs.
Not to pick on X-CDROAST, other apps do this as well.
If you are not going to design a GUI that eliminates some of needing to know how to do a task it is not worth porting an app to a GUI.
If I have to know a string of commands and how to use them Xterm tastes great and is less filling.
Steve
What if there was a project whose sole job is to look for this kind of problems. There are projects for documention , drivers, devices, hardware for linux.
Why __not__ for this. This would requred mostly non-technical people and an efficient feedback system to the project maintainers. And the Linus's "enough eyeball's" Law which applies to the Software bugs would work extremly well in this case, since the core project developers can't do this alone. And at the same token, depending on non-technical end-users to look for these & report would be a wait for ever.
My two cents.
The problem with having the "worker nerds" do their thing first is that the very architecture of their system may preclude (or make difficult) some necessary newbie functionality. To paraphrase Alan Cooper, code is to design like concrete is to architecture: once the concrete is poured, it's REALLY hard to change it, no matter what changes you make on the pretty blue paper.
Ideally, you let the design nerds do some user research before you start coding at all. Who is the target audience? What design metaphors are the already used to using? How much (usually, how little) experience can we assume?
Then you prototype. Prototyping isn't much different from coding: prototype your designs (on paper for starters), find out where they crash (i.e. where people get "hung"), debug, rinse, repeat. You won't work all the bugs out with a paper prototype, but you can nail an awful lot of them.
THEN you start coding. And you test and refine as you go, since some things (scrolling, for example) can be hard to simulate with paper. But you can get so much information if you just take a couple of weeks at the beginning and put some thought into your design, and then find some people who are representative of your target audience, and say "You have a printer attached to a different computer on your home network. You want to be able to print from this computer to the printer on the other machine. Here is the first screen..."
(Spoken, by the way, as someone with a foot in both worlds -- a design nerd who has also co-written a C compiler).
I am impressed it only took him one night. I lost three days trying to get cups to work. I now have my own vodoo ritual for getting a printer on another machine to print. I sort of understand his complaint but, If you want something clean, packaged, and perfect buy a completly integrated network and peripherals from someone like IBM. I still do not understand why any individual feels like this stuff has to conform to their standards. It is what it is, no more...no less. I get frustrated daily, I am constantly hosing boxes sometimes I can recover sometimes I can't. So what do I do? I reinstall a lot. If you don't like it abandon it. Life is too short. I for one am grateful that all of this capability with it guts hanging out is available. To me this is an adventure not a task. If you demand professional go and pay for it, don't scream and holler at those who were brave enough to stick their work out there for all to see. If you can't afford it then shut up and take what you can get.My first computer had one floppy disk and 64 kilobytes of memory and displayed 32x16 graphics on a green television screen. It costed thousands and was useless. I'd say the present state of affairs is pretty good. If you want to be critical and demanding then open your wallet not your mouth.For the record I don't care who uses it, I only care that I do.