The Answer Gang

By Jim Dennis, Jason Creighton, Chris G, Karl-Heinz, and... (meet the Gang) ... the Editors of Linux Gazette... and You!

---

Zdnet: Why Linux isn't ready for desktops

From Jimmy O'Regan

Answered By: Ben Okopnik, Jason Creighton, Kapil Hari Paranjape, Thomas Adam, Mike Orr, Pete Jewell, Jay R. Ashworth, Martin J Hooper, RJ Burke, Adam Engel, Rick Moen, Peter Savage, Heather Stern, Brian Bilbrey

http://zdnet.com.au/news/software/soa/Why_Linux_isn_t_ready_for_desktops/0,2000061733,39202374,00.htm

[Ben] A few of the writer's points make sense, or at least point in a reasonable direction - but what I find annoying... no, scratch that. What I find offensive about Mr. Dotzler's article are his repeated "Linux must do XYZ" statements, which shows an astounding absence of cluons for someone who is supposed to be a member of the Mozilla team. Contrast, for example, this month's article in LG by Dale Raby; despite being fairly new to Linux - or so I have gathered from correspondence with him - nowhere does he say "Linux must" anything. That phrasing says two things:

1. That Linux is some monolithic thing - perhaps a company, or perhaps a ruling Cabal underneath all that Potemkin-village "Open Source" nonsense (come on, do you REALLY believe that people would just work for no money like that? What are you, some kind of Comm-o-nist or sump'n?), and
2. That the writer knows better than the thousands of people who have put in their effort, time, and ability into making Linux what it is. This implication often also carries a stink of "they owe it to me", although nowhere is it actually explained why the person making this statement deserves his assumed entitled status.

Dale Raby's article says, "if some group or company decided to do this, they'd reap serious benefits." Asa Dotzler's article says "Linux must fix this to be worth not turning up my nose at it." Nowhere does it say why the Linux community should not turn up its nose at Asa Dotzler's views on Linux instead... but I suppose that a thing accomplished speaks for itself. Meanwhile, we'll all go on using thousands of dollars worth of excellent software and being grateful to those who have provided it for us without charge or requiring us to agree to freedom-abridging licenses - including Mr. Dotzler when he's doing his part at Mozilla (just as we can enjoy the performance skills of a sports figure while deciding that his political statements are based on nothing short of appalling ignorance.)

As for me, I suppose I'll never figure out why so many people are so deathly allergic to freedom.

[Jason] Yes, raises valid points, but...

http://jcreigh.blogspot.com/2005/08/windows-isnt-ready-for-desktop.html

[Ben] [laugh] You rock, Jason. "Computers are not easy to use" - that's a simple fact of life. The problem is that there seems to be a certain type of people - well represented by Mr. Dotzler - who expect Linux to solve all their problems, including the absence of Tim-Tams at their local supermarkets. What in the hell is that all about? Is it because of Linux advocacy, because they feel that once somebody says "here, have a great OS", they're somehow suddenly owed instant and total magic resolution of their woes? It sounds insane when stated this baldly, but... how many times have we all seen it?

[Jason] nods I hadn't thought about how Linux advocacy probably has a part in it, but it makes sense. I mean, here's people using an OS that they claim is good! Sure, Microsoft says Windows is good, but nobody really believes them.[1]

So then the shock of trying to install this things, that, it turns out, is probably at least at complex at what the user is used to, and on top of that, it's different complexity.

[Ben] Perhaps it's simply that Linux - offered, as it often is, by a person instead of a giant corporation - has a face. There's a person at whom they can aim their computer-related frustrations - and they do. That doesn't excuse the behavior, or perhaps even show anything other than a bare glimmer of the real reasons behind it... but it makes some sense in a highly bewildering situation where nothing else does.

[Jason] Hmm...perhaps people don't understand free software. I mean, they get the part where they don't pay anything, but they don't get the part where they have no right to expect anything.

I saw a nice quote along these lines:

"Notice how people talk about this: "Linux" (ie, the tens of thousands of developers worldwide writing the software that together make up what we call a Linux system.) should have this, "Linux" should have that. Fewer, more powerful distros, tools to do what I want. People who talk likes this are essentially saying to people working for free: "Write my code". If you don't like it, shut up and code."

http://jcreigh.blogspot.com/2005/07/linux-and-greater-good.html

[Ben] I think you're right - if people separate the process into 1) getting the software and 2) using it, then there's going to be a sense of "well, however I got it - store-bought, shareware, stolen, or some weird 'free' thing - now somebody OWES me because I went to the trouble of getting it!" The thing is that I've seen this kind of thing in the world of commercial software, ridiculous as it was: when I was working in tech support, many years ago, we'd have people call us up (you could almost always recognize them by their aggressive-as-hell approach) and tell us how they wanted their DAMNED program fixed RIGHT NOW! That one didn't take long to solve, though.

-- "Sir, do you have your dated receipt and the serial number from the original packaging for that software?"

[ short silence ]

-- "Uh, no. My cousin copied his disks for me... but, uh, he bought them from your store, and it cost real money! So, I want it fixed!"

The conversation after that took a predictable turn.

The problem with handling this, from the Open Source perspective, is that there isn't the obvious implicit knowledge on the user's part of having stolen the software - because they haven't (and no, I'm not saying it would be better if they had - it's just that the attempted theft of effort is glaringly obvious in the commercial case, but is hidden by the free nature of the software in the non-commercial one.) However, what they're trying to steal - via the same brazen attitude (which comes from the same source and exists for the same reason) - hasn't changed: it's other people's effort to which they don't have any right.

[Jason] We're so used to going to a store, buying something, and then expecting to have somebody to yell at when something goes wrong. Then free software comes along, and people unconsciously treat as just another product.

<shrug> I don't know. I know that when something crashes, or doesn't work the way you want, it's awfully hard to think "You know, on the whole, I've saved a lot of time using this software, and I didn't pay a cent for it. Maybe I should drop the author a thank you note. Or better yet, money. Or maybe a gift certificate to some franchised pizza place."

I used to volunteer at the Student's Union when I was in college. At the start of every year we made 'Fresher's Packs', packed with whatever free goodies we could get. We tried to put the same things into every pack, but some things ran out, sometimes people accidentally more or less than others (and sometimes 'we' filled bags with condoms to hand out to the cute girls. Ahem).

[Sluggo] Did you write your phone number on the condom wrapper?

No need. They knew where to find me. (Not that I was actually able to capitalise on it at the time though )

Guess what? We got nothing but complaints for the rest of the week about the contents of those packs. (The best solution I used was to add up the value of the average pack, and offer to refund the value of the missing part(s), once the rest was paid for. That didn't shut them up, but at least they took their grumbling elsewhere).

[Adam] I kind of understand what the zdnet guy is talking about, but not where he's coming from. My first year using linux in 1996 was spent without a GUI because I wanted to learn the command line and also, I had no idea how to configure my machine for X. From what I learned, I was able to get a job with a VAR who put me through six months training to become a "Sun Certified" sysadmin. From Networking, I went on to my present studies of Shell scripting, sed, awk,perl etc. I wasn't a very good sysadmin, but it's amazing to me that a "non-geek" like myself was able to get that far, just as it was amazing for me, after years of fvwm2 and WindowMaker, to see KDE for the first time, which for me, has so many more options than either the mac or Windows desktops. I've turned at least a half dozen people onto Linux where they can easily learn to navigate KDE or GNOME, download OpenOffice and Firefox, and use The Gimp and Ghost View for graphics and pdf files without having to upgrade Photoshop or Adobe's Acrobat for several hundred a year. I'll never be a "professional" programmer, or even a good one, but I have fun for a few hours a night on my Linux box and do all my professional work in OpenOffice/Thunderbird/Firefox,whereas previously I had to buy ( or "borrow") at last MS WORD if not MS OFFICE"S whols 400 dollar package. I never would have learned the stuff I've learned and am still learning and have so much more to learn if I'd just used Windows or Mac to do the typical word=processing/emailing/web-surfing. I also have a Mac running Jaguar. Since I only use the "unix" utilities -- and XDarwin-- I don't have to worry about coming up with $130 to "upgrade" to Tiger and then to Wildcat or Griffin or whatever they'll "offer" next.

"Computers aren't easy." Why should anyone care if Linux "takes over the world" or remains as it is, the labor of love of programmers, networkers, hackers, graphic designers etc.

[Kapil] There is perhaps a good reason why you are not a "professional". Read Paul Graham's article on "What business can learn from Open Source and Blogging". Around line 300 (of this rather rambling article) he discusses why "amateur" is not such a bad word after all!

---

[Kapil] The article that this (ZdNet) article claims to rebut says that one of the things that needs to be done is to have some company come out with good pre-installed systems (with some CD's containing additional software and a support infrastructure thrown in). The OSNews article also says that this is targetted at users who would be willing to learn the new interface if it was "cool enough" and goes on to claim that the current desktop interface is. The target users are also those who are attracted to the MacMini but find it too expensive and/or limited in terms of available software.

The ZDNet article nowhere refers to these points---instead the author talks mainly about how Windows users will find it difficult to migrate to a typical GNU/Linux system. The author then concludes that the way to simplify this is to make a Desktop that looks like Windows but is GNU/Linux inside.

This is a very old flaw---like the British Labour Party turning into a clone of Ma Thatcher in order to come to power. Or the communists turning on its head the philosophy of the withering away of the state. The problem with "World Domination" is that you feel the need to go down such a path---becoming like the rival or perhaps even worse.

[Ben] To draw a parallel that has come up here quite often, it is also the problem with the political process here in the US: no matter how good and decent a man is when he starts out to climb the political ladder, making it to the top will almost inevitably - at least in every case I've seen - make him into a near-copy of the slugs that he's trying to replace. What is bitterly ironic about the whole cycle is that good men choose to climb that ladder because the slugs at the top are so repulsive and need to be replaced.

The salient thing about Open Source, to me, is that the process of reaching the top - however loosely that may be defined - is not destructive in that manner. Trying to adapt it to this other philosophy would make it so, in my opinion - and make it ultimately unethical (no ethics can be called coherent or rational if it requires behavior that is ultimately self-destructive.)

[Kapil] I think the OSNews article makes a very valid point that some company needs to realise that there is a 5% and very visible market share to be captured by having pre-installs of something like Ubuntu/Mandriva or whatever. The folks at Apple and Sun (note the Java Desktop OS) have realised this but have not been able to keep the costs low---perhaps because they are older corporations with hidden costs. So someone can capture this segment first---world domination can wait.

[Ben] Or perhaps that very method is the path to World Domination. :)

---

[Thomas] Not this old polemic again ? How many times is it that this one is put through the mill grind? The result is always the same -- there never can be one. And indeed, the article above doesn't mention anything that hasn't already been discussed.

One thing that the author of the article does depict rather clearly is that he's more naive about operating systems than I initially gave him credit for. His argument is that technology is supposed to be the cure to all of the problems computers suffer from. Hmm, but that's user error in my opinion. And that I don't like. If you suddenly approach it from the fault of the computer's inability to do something, you won't get anywhere.

[Sluggo] That's another aspect of the "both sides are right" problem. We've all seen the "if computers were cars" jokes. My dad used to say in the 80s, "I can get into any car and the steering wheel is in the same place and does the same thing, and so is the brake. Why aren't computers like that?" It still apples today. On the other hand, "If people treated their cars like they do their computers, they would never change the oil, ever, and when the engine falls apart they'd scream at the manufacturer." (I think I heard this one from Ben Okopnik.)

[Sluggo] As to why this difference in perception exists, I don't know. Maybe it's because manufacturers overpromised on the wonders of computers in the 90s (and 80s) to get more boxes off the sales floor and into homes. But I'm not sure if that's the whole story. If it were, it would be the marketing success of the century. But I think there must be other factors too, although I'm not sure what they are.

And now, curiously, a significant number of people are cancelling their broadband and going back to dialup, or putting their computer in the closet, coz it wasn't the panacea they thought it was. Were they duped by the marketers? Or did they not think clearly enough before they bought?

But Thomas Friedman argues (in The World is Flat) that globalization and computerization are actually much further along now than they were in the "heyday" of the 1990s, even though thousands of .com domains have been abandoned. So that's even more curious, that some people are using computers less yet the overall computerization is higher.

[Sluggo] My answer to my dad was always, "Computers are expected to be more versatile than any other appliance. A car just runs one program that was installed at the factory. But people expect computers to run any old program they got from who knows where, and they expect it to work with any of the thousands of peripherals they might attach to it." How much do people attach peripherals to cars? Maybe a stereo. That's one electric wire, big compatibility deal. Maybe a speaker wire and antenna, but those are already universal.

But my dad has a point too. Why are computers in people's lives?

[Rick] Because they're not -- in general -- smart enough to demand limited-purpose computing appliances, and the industry's too rigid and stultified to offer them.

But, of course, that's not quite right, either: The problem's already solved in any number of such appliances, starting with TiVos and PalmPilots -- which, of course in many cases run Linux. So, the ultimate answer will be that such things will not be perceived as the solution because they'll not be called computers, but nonetheless that's what they'll be.

[Heather] It was brought up at LWE's documentation BOF, that a labelling problem exists, too.

[Rick] If memory serves, it was the guy I shave making that point. ;->

Pardon my hubris, but I think I had some of the best lines, including "Since around 1998, 'desktop' has been functionally defined as anything non-Linux, and/or whatever Ziff-Davis columnists are trolling over, at any given time."

[Heather] Given that a pundit can choose how to define "desktop" before he starts to rant (equally for or against whether $foo is Ready For The Desktop), many seem to be choosing to define "desktop" as whatever Redmond has a lock on these days, "workstation" as anything that's escaped into UNIX' graces but has a GUI (whether Linux or Sun or others, on the premise that Those Engineers Don't Really Use Desktops), and "server" as those scary things with no GUI. If it has no keyboard they summarily declare it not a computer and ignore it. So serial console jugglers and uplink managers with Linux or some BSDish OS inside are "not computers" the same way your microwave or phone is "not a computer" whether a sysadmin could login to them or not. They're "internet appliances" and treated like smart toasters. Pulling out of the almost offtopic dive, appliances are expected to not have or need documentation - their actual manual is more than half filled with legal warnings like "caution: hot" and techie minutiae like what kind of wall power it uses, even though the plug only goes into one country's sort of wall socket.

It leads to a good question about where the NetBSD toaster I saw in the .ORG Pavilion falls. You can login to it as a UNIX shell, they gave it a keyboard and a multi line LCD screen, and it can eject the toast early from the commandline. I asked if it can detect whether it was given toast or bagels. Yes, I took pictures.

[Sluggo] Geeks have them because they like the technology itself. The masses have them because they want to get things done. And in the vast scheme of things, isn't it getting things done that matters?

I once complained to Ben about the high cost and lack of choice in broadband availability. He just said, "The Internet is in its infancy. That will get better in time." It struck me that it's good to think long term like that. Just because things are a problem now doesn't mean they always will be. Perhaps computer usability is the same way, it's just in its infancy.

[Thomas] "The second problem that blocks massive Linux desktop growth is stability." News to me.

[Jimmy] Missed this first time around, but the author was actually talking about API stability, not "the OS is likely to crash" stability.

[Thomas] And this is one thing that annoys me. Of all the articles I have read about whether Linux is ready for the desktop or not, my answer to them has always been the same: "Yes". Linux has always been ready for the desktop, even way back when RedHat were new. The reason why Linux tends to "fail" in some people's eyes is that they make such a tight and stringent comparison to Windows, that a lot of people expect Linux to act like windows; somewhat contrary to the quote, above. In so doing, what the evaluator doesn't therefore do, is appreciate the differences Linux hails over Windows -- even in the GUI world, which is still where the comparisons about whether Linux is ready for the desktop take place. It annoys me.

Yes, of course people are going to want to use GUIs, but they're so determined to make sure it looks and acts like Windows, that they can't see the differences it gives them, either. KDE and GNOME have gone to great pains to make their desktop environment as windowsey as possible. Where's the gain, in doing so, when the comparisons people will draw will be like from like, when comparing it to Windows? It belittles the comparison.

[Sluggo] I'm getting the feeling of people talking past each other.

[Rick] In large part, it's actually Tantrum Theatre (on the part of J. Random User valuing free-of-charge gift-culture offerings at acquisition cost) and Trollfesting (on the part of bored ZDnet columnists).

[Sluggo] Nobody says "FreeBSD isn't ready for the desktop!" or "If FreeBSD wants Soccer Mom to use it, it'd better do this, this, and this, and quickly." Why? Because FreeBSD presents itself as a pure geek distribution. BSD developers/users are rarely interested in converting Windows users; they're more like "If you want it, it's here." But Linux lives in this weird hybrid world that's part geek and part messianic. There are a lot of factors for this. Maybe Linus's "World Domination" joke got out of hand. Maybe some companies like Red Hat and Lindows overpromised.

[Ben] Overpromised what, though? I don't recall ever hearing anybody say "it'll all work perfectly and exactly the way you want it - even if it's not our product." If anybody paid money to RH or Lindows, they know who to complain to; those aren't the people you're going to hear in these venues precisely because they do know who to complain to.

The people doing this "Linux must" crap are those who haven't paid for the privilege of complaining. That's why they're doing it, and why they're screaming so loudly.

[Sluggo] People don't understand that Debian nuts and Lindows management may have different viewpoints. They think there's one official distribution and "Linux" means the whole thing.

[Rick] Yeah, $WHATEVER. When I encounter rantings from the sundry "Linux must do $FOO" people, if sufficiently interested, I inquire: "Which Linux are you speaking of?" If, as usual, the answer is "Uhhh, what?", then I say, "Ah, so you have no idea what you're talking about, and yet you see this as somebody else's problem."

Most people, most of the time, value things acquired at the cost of acquisition. This immediately leads to culture clash, when they encounter gift cultures such as ours. Some guy I shave who maintains the Linux User Group HOWTO has quite a lot to say about that:

http://en.tldp.org/HOWTO/User-Group-HOWTO-4.html#ss4.2 http://en.tldp.org/HOWTO/User-Group-HOWTO-4.html#ss4.4

[Sluggo] When you tell them Linux is just the kernel, it sounds like some irrelevant tekkie abstraction. "Where do I see the kernel on my desktop? Oh, it's just 'there' behind the scenes, invisible and unfeelable. Well, who gives a rip?"

The thing is, you're arguing with this author, but you're both right.

[Ben] See, I have to disagree here. I don't see where demanding unpaid service is ever right, or can ever be right.

[Sluggo] At the same time, you're not getting at each other's core values, you're talking past them to some other faction. Most of The Answer Gang falls into the "geek" category who use Linux because it suits them, not because they're on a jihad against Windows. We'd like to see vendor lock-in eliminated and Linux adopted by the masses, but it's not our #1 priority. Because of this, we get angry when people appear to be putting demands on our donated time. But is that what this author is doing? No, he's merely stating that if Linux thinks it can replace Windows it'd better be more friendly to Windows users. Why is he saying that? After all, he's not saying it about FreeBSD. He's knocking on Linux's door and dissing it because Linux went to him first and told him, "Thou shalt not use Windows, it is an abomination. Thou shalt run Linux, for that is pleasing to God."

[Ben] I take it that Linus came to him in a dream, as a 200'-high penguin on fire? They have dried frog pills these days that'll fix that up pronto.

If some 15-year-old zealot told him a tall tale, then he's 1) way too gullible, 2) needs to learn a bit about filtering for source accuracy, and 3) should demand tech support from that 15-year-old. Making the rest of the Linux community responsible for his... how shall I put this politely?... failings... is nonsensical. No one except him is responsible for his decision to use Linux - just as no one would be responsible if he jumped off a bridge after hearing rumors of treasure in the water below.

[Sluggo] It doesn't matter that we didn't say this. Somebody did (though not so religiously), and that's what he's responding to. Questions about "Who should do the work?" are beside the point. As far as the author is concerned, those who want to increase Linux's market share can do the work. He's just pointing out they're going in a wrong direction for their goal. Which I don't necessarily agree with 100%, but it's a legitimate argument.

[Ben] I hear what you're saying - but again, I have to disagree. The adoption of Linux is proceeding on a large scale, all over the world; no one has held a gun to that author's head (I may be wrong, but it seems to me that there would be a lot of repeated letters due to his teeth chattering if that was the case) and said "ADDRESS THE TOPIC OF LINUX ON THE DESKTOP OR DIE, SCUM!" If he said "if people want Linux to advance in X direction, here are some ideas", I'd see that as valid input. "Linux must" is an imperative, an arrogation of position that the writer has not earned in any manner - and even though a stopped clock is actually right twice a day, that does not make it in any way useful.

[Sluggo] It comes back to usability. [http://linuxgazette.net/116/orr.html] If the goal is to make a system for geeks; it's already done. If it's to convert the masses of Windows desktop users, it has to be familiar to them.

[Rick] Guess what? People get used to unfamiliar software all the time. In the most-common form, it's called "The boss issued me this computer, and I'm expected to use its software to carry out my job, or get fired and take up some other profession elsewhere." A less-common but still popular form is "I wanted application $FOO that I'm used to, but it's no longer available so I learned application $BAR instead."

People whine (e.g., "I demand an easy migration path!") when they imagine that whining's going to get them something -- and few use software they specifically want in the first place: Most just end up using something as the path of least resistance. This has obvious implications, including the people who say "Linux must do $FOO to become popular with J. Random User" suffering from recto-cranial inversion for lack of acknowledgement of that fundamental truth.

Pundits who post this rubbish are wasting our time. Frankly, I suspect everyone on TAG already knew this, going in.

[Sluggo] Many users don't care whether it's "better", they just want it to be familiar.

[Thomas] It's a fair point, and from what I remember of the article you wrote, that more or less sums it up. But one thing I am surprised you did not mention (as it was something that was drummed into my head, when I was studing "User Interface Design" -- "UID") is "Schneidermann's golden rules". Schneidermann lists eight golden rules to good UID. They're generic enough (in my opinion) that if developers stuck to them, interfaces would be good to use.

[Sluggo] I haven't had formal training in this so I don't know all the rulesets out there.

[Thomas] Join the club. :)

[Sluggo] What are his eight golden rules?

[Pete]

http://www.cs.utexas.edu/users/almstrum/cs370/elvisino/rules.html

[Sluggo] And which applications most egragiously violate them?

[Thomas] The link Pete Jewel sent your way is all that's really needed to know in a nut-shell. They're principle guidelines that one can use; both when designing GUI components, and evaluating them. A good heurstic approach, if nothing else.

You can't apply them to any one application. If anything all applications violate the eight golden rules in some way -- and in a sense, it's expected to do so. Really, the rules are about being consistent. Use of colour. Appropriate use of menus, grouping of related components. Making sure people with disabilities (such as not being able to use a mouse) are catered for.... the list is endless.

[Sluggo] The latest discussion at work is whether the Save button should go on the left or right of the Cancel button. My project manager said he heard the Cancel button should go on the left so users don't accidentally do something they didn't intend to.

[Thomas] :) I too had this debate. If you read the windows XP style-guidelines (I don't have the link handy, and I can't be arsed to google for it). Microsoft recommend it to go on the right-hand side. Which, again, if one strives for consistency and familiarity, in the use of GUI apps that have been well-established, then perhaps it is "better" to follow the convention. No? At least, in so doing, you minimise the likelihood of not needing to (re)train people on how to use the end product.

The "problems" arise then, between OSes. Doing the above is fine on Windows, but some GUIs under Linux do not. And I think it's that area that throws a lot of people used to Windows. Windows is, after all, the de facto.

[Sluggo] I usually see it on the right. But we looked at some applications on the Mac and Linux and they did have Cancel-OK instead of OK-Cancel. I tried to do that in a web app but it felt wrong, the Save button goes on the left. Then I pressed Enter in an entry screen and my changes were lost, and I thought, "That's really wrong."

[Ben] Default tab order is a different issue from positioning, but I agree with the implied point: the default action, whether "Submit" or "Cancel", requires careful consideration before implementing.

[Sluggo] Because most modern browsers activate the first submit button when you press Enter, and people don't expect that to be Cancel.

[Ben] You do realize that this is optional behavior, right?

<!-- 'Submit' is displayed first, but 'Reset' is the default action -->
<input type="submit" value="Submit" tabindex="2">
<input type="reset" value="Reset" tabindex="1">

[Sluggo] So maybe web apps have the opposite tradition of desktop apps.

[Thomas] That's a whole different kettle of cod, or haddock, or carp. I prefer haddock. :)

[Jay] I believe the Mac HIG says that the OK button always goes in the bottom right corner.

On a web app, I'm fond of never having a cancel button at all.

You haven't lived

[Ben] ITYM "died". HTH, HAND. :)

[Jay] until you've filled out a four page long webform, then accidentally hit Cancel.

[Ben] [sigh] Or, just as good, hit 'Esc' (with those finely-honed 'vi' reflexes) when you finished typing in a text field - only to see everything you've just typed disappear. Want... strangle... NOW!

[Jay] They want to abandon the form, they can just click Back.

Cheers, jr 'and if they weren't taught about back, that's their problem' a

[Sluggo] I'd never seen a web form with a Cancel button before (as opposed to a Reset button), but it was required in my last application and I found it pretty convenient, so I put it in my current application too. Even though I know I can just leave the page and it's a de facto cancel, it's convenient to use that button at the bottom, and I set a message in the session for the next screen saying "Add record cancelled". Users like that sort of thing, and I found I do too.

[Martin] This is a good point as I am always cursing Gnome apps that I use because I am conditioned to click on the first button - Presumably because I use Windows the most and that has the OK button first.

I think it would be a lot better if all the apps we use used the buttons the same way!

[Thomas] But they don't. Just look at GNOME's save-dialogue window. It's a complete mess. But on the other hand, it is a familiar interface. I watched my step-dad use KDE for the first time, the other day. Being a windows user, he soon picked it up. What tends to allude people are the different widget sets and toolkits -- a problem that is soon solved with familiarity.

[Sluggo] Telling the users what they should want goes nowhere. Presenting them with something that's mostly familiar but with a few obvious improvements does have a chance.

[Thomas] See above. But to my mind, I can see a problem with this. At what point does familiarity across systems become so common as to lose the very uniquess these systems offered in the first place? I know that stance is one that an experienced user might take, but it should be a considered one.

An interesting story that was in the news today, is that a government-invested I.T project for the NHS (National Health Service) here in the UK, is most likely to fail. Why? Because the morons that have "designed" it, somehow neglected to ask the users what they wanted from the system. I ask you. You don't even need to be a software engineer to realise that one.

[Sluggo] Firefox and Thunderbird have taken hold because they offer obvious improvements (95% drop in viruses/spyware) and are easy to learn. Linux as a whole (which we know means, "a particular Linux distribution as a whole") still has too big a gap in that area, too many times you have to call in an expert. Whether the problem is factual, perceptional, or the fault of secretive hardware vendors and the DMCA, it's still a legitimate issue for discussion.

My firsthand experience of this is converting my mom to Linux. Because of her medical conditions, she doesn't learn or remember things as easily as she used to. I show her how to use Thunderbird but she forgets, and then she's afraid she'll screw it up and break something, so she goes back to the 10-year-old Eudora on the 486/66 because she can get its basic functions working. So the next weekend I come over and show her how to use Thunderbird again and she takes notes, but she still gets confused until she's done it several times. Then she asks me, "Why is [computer #2] so slow? Does it have spyware?" (She got it used as a gift to replace her 486.) I just have to say, "I don't know." I don't follow Windows closely enough to know all the different kinds of virii and spyware out there. But I don't want her paying $50 for an anti-spyware program, not when she needs the money for glasses and shoes. The only alternatives are reinstalling Windows (tricky with the incomplete distribution disks computers come with now) or switching to Linux. But I don't have the 50 hours it would take to handle all the incompatibilities that might come up and walk her through using Linux programs and administering her Linux box. But I do want to get out of the "Windows tech support" role,

[Rick] So, do that. The only way to stop being an enabler is to stop being an enabler: http://zgp.org/~dmarti/linuxmanship/#enabler

[Sluggo] and get out of finding license keys buried in documentation, or doing without programs Linux comes with. I've made a Knoppix disk and I'll let her play with that for a while to get used to it. It's great that Knoppix automatically starts DHCP so it's zero-configuration for DSL systems with dynamic IP -- except she has dialup. So I'll have to either edit her PPP chatscript file or find the GUI configurator wherever it is, if it has one. Yippie. One good thing though. I've been able to indoctrinate her against Outlook (vendor lock-in) and the evils of the DMCA. She doesn't like Microsoft's monopolistic practices and she'd be glad to switch to Linux... if it were only more familiar and easier to learn.

Some of the comments in the article appear false. Aren't there Linux distributions that automatically put an icon on the desktop when you insert media, without you having to "mount" it? I know KDE does with hard disk partitions (although I never use them, being a konsole/xterm kind of guy).

There's a FreeDesktop.org project that aims to do Windows/Mac-like hardware autodiscovery (HAL, I think).

[Sluggo] His vision of one Windows-like desktop to replace the KDE/Gnome dichotomy is a bit hilarious. Of course that's unacceptable to geeks, both because of the lack of choice and because the Windows UI is so limited. (Can't type into a window that's not on top. Can't switch desktops.) Users don't have to "choose" between KDE and Gnome; they can take the distribution's default and have a perfectly competent desktop either way.

I'd take his comment to mean that it's confusing to offer the choice of both when there's so little compatibility between the two: you can't embed Gnome apps in KDE apps and vice versa, you can't always drag and drop between Gnome and KDE apps, etc.

[Sluggo] But the fact remains that a significant number of people want a free OS that behaves like Windows as much as possible

ReactOS?

[Sluggo] (minus the bugs),

Oh. :)

[Sluggo] and maybe the companies and volunteers that really want to capture this market share should just give it to them. Hopefully it would have a clear migration path to the "superior" Linux configurations the rest of us use, when the user is ready for it.

[Ben] As I understand it, Evolution provides a complete upgrade path for those who want an 0utlook/IE/etc. migration for that. So do the Alacos Migration Client, Versora, and MoveOver (and, I'm sure, a number of others - that's just off the top.) The methods are there - for those that want to know about them.

Well... I read that part somewhat differently. It is possible to import mail from Outlook Express with Evolution and KMail (KMail even claims to import Outlook/Exchange mail, but I've never tried it), but what I think the guy is trying to say is that the Linux desktops need a 'libwinreg'[1] that can allow various programs to take their configuration straight from Windows without bugging the user with details like the path to their Outlook files etc. (That's what they did with the Windows version of Firefox -- it takes IE's configuration from the registry, automatically imports IE's Favorites, etc.)

[Ben] The problem with that, as I see it, is that 'libwinreg' will have an arbitrary dependency: Wind0ws. Whoever maintains it will have to spend a hell of a lot of time learning and relearning Wind0ws every time a new version comes out - which is, I submit, not what people generally want out of their Linux experience.

Sure: anyone who intends to interoperate with any other system has to expect that. The advantage here is that there are so many other projects that depend on knowing the format of the Windows registry: Wine, Rewind, ReactOS, chntpw, etc., so at least it's a burden shared.

[Ben] Good point. In fact, it seems like a project like that could be a small, low-effort spin-off from one of the above.

Well, I pointed out the code from Rewind that does this already, and should have a simple tool to dump the registry as text in the next couple of days. There are a couple of things I need to figure out before I can make a useable library (things like "is it reasonable to demand that the calling application specifies where the windows partition is mounted"), but it shouldn't take too long either.

Hmm. Turns out that Samba already has a better version of what I was thinking of: http://websvn.samba.org/cgi-bin/viewcvs.cgi/branches/SAMBA_4_0/source/lib/registry (I think Samba 3 has it too). This has the added benefit of being able to access the registry of remote windows machines.

Further investigation turns up: RegLookup (http://www.sentinelchicken.org/projects/reglookup), based on the Samba stuff. NT/XP, read only Regutils (http://www.cs.mun.ca/~michael/regutils) Win 9x, read/write Readreg95 (http://www.indoktrination.at/projects/readreg95) Win 9x, read only

Users of this as yet imaginary library have an easier task when it comes to supporting new versions & programs: install the software in Wine and read Wine's registry (which is stored as plain text).

[Ben] Yeah... but using Wine as an additional layer of complexity for new users - in effect, having it as a dependency for everything they do - is probably NOT a good usage pattern to establish.

[Thomas] This is one of those examples where software shows its true colours as being supreme over its users, and other software dependant on it. New users might look at having to use Wine as complex; to me it is. OTOH, if they reslise its a prerequisite to having to run an invaluable piece of software they can't do with out, they might overlook the fact its complex and try and use it anyway. Either way, the use of Wine is out of the user's control, alas.

Whoops. When I said 'users of this ... library' I meant developers who are creating software with it. For them, it'd be easy, yadda yadda. For general users of software using the library... well, they can just bitch at whoever they'd normally bitch at if a new version of $RANDOM_WINDOWS_APP doesn't get migrated right.

Thinking along those lines, I can imagine it'd be nifty to have things like a PAM module that authenticates against the NT SAM (so you can have a version of Knoppix, say, that accepts the same username & password as Windows, then automatically sets up a 'My Documents' link etc), a tool to dump Windows' dial-up settings as chat-scripts, etc. etc.

[1] Like a standalone version of this:

http://cvs.sourceforge.net/viewcvs.py/rewind/rewind/misc/registry.c?rev=1.106&view=auto

[Sluggo] Geeks don't need to be offended when people ask for this. Just tell them politely that you personally have other priorities for your volunteer time, and give them a glimpse of the variety of viewpoints in the Linux world so they don't think we're all monolithic. Then, if you can, point them toward people that are trying to create what they're asking for.

[Ben] I'd call that excellent advice.

[Sluggo] As to why this difference in perception exists, I don't know. Maybe it's because manufacturers overpromised on the wonders of computers in the 90s (and 80s) to get more boxes off the sales floor and into homes. But I'm not sure if that's the whole story. If it were, it would be the marketing success of the century. But I think there must be other factors too, although I'm not sure what they are.

And now, curiously, a significant number of people are cancelling their broadband and going back to dialup, or putting their computer in the closet, coz it wasn't the panacea they thought it was. Were they duped by the marketers? Or did they not think clearly enough before they bought?

But Thomas Friedman argues (in The World is Flat) that globalization and computerization are actually much further along now than they were in the "heyday" of the 1990s, even though thousands of .com domains have been abandoned. So that's even more curious, that some people are using computers less yet the overall computerization is higher.

---

[Reilly] It's one thing to have the application settings imported to Linux, but there's also the FILE TYPES. For us, the deal-breaker has always been CorelDraw. There's no way to use our 15,000 CorelDraw files on Linux, and none of the Linux applications can open a CorelDraw file.

Most enterprises have this same type of make-or-break issue, and a perfect Linux desktop still won't be good enough. BTW, we use Linux in those places where there's no graphics, and it works just fine on all other file types.

[Ben] Well, RJ - the question under discussion wasn't "why don't companies adopt Linux more often" - large companies, government offices, and entire countries are switching over in droves; you may have even noticed that the pundits who used to loudly proclaim that "Linux isn't ready for the Real World!" have quietly vanished away - or have switched over to "Linux isn't ready for the desktop!" I'll admit that I'm not certain why you're introducing it as though it was relevant to the current "Linux on the desktop" discussion, although I'd be interested to hear your reasoning.

Having said that - I've worked for many different companies as a consultant, and I disagree with your estimate of "most enterprises" having a make-or-break issue based on filetypes; in fact, I'd say that applies to a relatively small percentage of those who want to switch over but decide that they can't, since many, many common filetypes are handled within Linux apps.

There's the thing: common file types. Corel's (cdr) files aren't common file types. Support from other Windows applications is almost non-existant.

[Ben] Good point - it's not a direct connection. Corel runs under Wind0ws, and it supports .cdr files; neither Wind0ws nor Linux support them directly. However likely that point is to be elided in this sort of a discussion, it's not really about Linux or Wind0ws - it's about Corel's software, and its availability under a given OS.

Common file types are, in general, well supported.

It is possible to use Corel's files on Linux; a simple, two-step process:

1. Install Wine
2. Install CorelDraw

Then "Linux" will have just as much support for CorelDraw files as Windows does.

[Ben] Yeah, but are the Blue Screens of Death and crashes included? Or are they a separate, extra-cost option?

No, but it's being worked on.

(That is, there is interest in integrating Windows screensavers with XScreensaver, so they'll get the BSOD for free :)

[Ben] If you've installed XScreensaver, you've already got it.

/usr/X11R6/lib/xscreensaver/bsod

[Sluggo] Now you just need a routine to activate the screensaver at random.

[Ben] But ask, and ye shall crash. Say... I wonder if this is the ultimate qualification that Indu Stree Pundit meant by "ready for the desktop"?

# Add this to your crontab; ~1% chance every minute...
*     *       *       *       *       [ "$RANDOM" -lt 327 ] && /usr/X11R6/lib/xscreensaver/bsod

[Sluggo] Brilliant. You have renewed your credentials as Dr Evil.

[Jason] Nice! But when cron runs this, won't $DISPLAY be unset?

[Ben] I don't know why he'd be upset; it's not like we didn't invite him to the party or forget to send him a birthday card...

Oh. "unset", right. Good point. So, forget "crontab" - life doesn't have to be nearly that complicated.

(
while :
do sleep 60
[ "$RANDOM" -lt 327 ] && /usr/X11R6/lib/xscreensaver/bsod
done
) &

[Sluggo] You could put DISPLAY=:0 in the command line and hope for the best.

env DISPLAY=:0 /usr/X11R6/lib/xscreensaver/bsod

[Brian] Sure, but if you wanted REAL evil, then instead of running the bsod screensaver, you'd set the image on the root window, and put a listener on the event interface that kills init.

[Ben] Hmm... if we're going for precision, that image would disappear almost instantly, so it actually wouldn't be all that MS-like. OTOH, I can't think of anything that you could do to make the system totally non-responsive while continuing to display that image (maybe "perl -e'fork while 1'"? :), so it seems that Wind0ws can do something that Linux can't. Oh, The Horror. Perhaps we'll have to write a kernel module, or fork the kernel (kernel-source-2.6.11-BSOD?)

[Ben] The major issues in adoption, from what I've seen, are 1) employee retraining and 2) administrator retraining/replacement - both of which are heavily political issues based on much that is unrelated to technical merit. This may well relate to the question of desktops; there are certainly non-technical considerations related to why people are going on with this "Linux isn't ready for XYZ" noise... although I, for one, would be hard-put to define the exact kit of aberrations involved. Perhaps it's actually very simple:

Shy and unready men are great betrayers of secrets; for there are few wants more urgent for the moment than the want of something to say. -- Sir Henry Taylor

---

[Adam] While I agree with just about everything that's been said about this article on TAG, except for the positive comments, I think that going off on logical arguments regarding technical feasability and work-arounds makes the fundamental mistake of taking this obnoxious writer and his petty article too seriously. I think the meat of the matter comes down to this statement:

'For "regular people" to adopt Linux (which usually means leaving Windows), it is going to need a serious migration plan. The OS will need to install on machines next to Windows, leaving that completely intact and easy to return to, and carry over all or nearly all of the user's data and settings.'

See, I think he has a point there about importing settings from Windows. It's a pity he had to bury it in a list of things that are already done.

[Adam] The writer is, I assume, a super hacker a la Stallman or Torvalds, not a 'regular person.'

I assume the author is a Mozilla developer, having been billed as such.

[Ben] Just to correct a small misconception:

Asa Dotzler (IPA: e??s?? d??tzl??), born in Tennessee on June 5, 1974, is best known for his work on Mozilla. In addition to being the QA lead, including community quality advocate for the project, Dotzler is also the Firefox and Thunderbird product release manager, and Mozilla's community marketing coordinator.

Not to take any deserved credit from him, but he's not a developer (in fact, he says so explicitly in an interview with LinuxQuestions.org). Perhaps that article would not have come into being if he was; that gratuitous "oh, somebody will get it done" attitude does not often come from those who are usually responsible for getting it done.

[Adam] Regular people aren't supposed to be smart enough to learn a new operating system, even though learning to use KDE on top of Linux is comparable to learning Windows 3.0 on top of Unix's retarded younger brother, DOS, although much more straightforward, logical and fun.

Yeah, but, as the author was saying, it could be easier. That's why there's a 'Simple KDE' project (http://simplekde.org, I think). (Some of) the KDE developers already knew that.

[Adam] Regular people aren't even supposed to use computers.

Yeah, but despite the constant LARTings they keep coming back ;)

[Adam] Who would have foreseen a world of Desktops connected to millions of web sites via high speed ethernet/fat pipe conduits when "the net" was a failed DARPA project 30 years ago?

Eh? Where did that come from? (And failed?)

[Adam] Just the name of the article, "Why Linux isn't ready for desktops" is ridiculous.

I agree. That title is nothing but sensationalism.

[Adam] What was Linux originally developed for, UNIVACS?

[Sluggo] Uni-what? Linux was originally developed because Linus wanted to experiment with programming a 386 chip.

[Ben] I think that was irony, Mike. For a guy whose nickname used to be "Iron", you sure whiffed that one. :)

[Sluggo] He wanted something like Minix that took advantage of the 386's power and was free. (Minix was built for the 8086 and cost money.) I think the 386BSD project was already going but he didn't know about it. 386BSD eventually folded and FreeBSD emerged.

It was designed for the 80386, which was the first processor in Intel's x86 line with an MMU, and therefore the first in that line capable of powering a Unix workstation. (386 powered PCs didn't really spring to mind when you thought of 'home computers' at the time).

[Adam] Linux/GNU offered a Unix-like system to "regular people" to run for free (or a relatively small fee for CD distros) in 1991.

There's a big difference between what you could expect of 'regular people' in 1991 and 2005.

[Sluggo] "Regular people" meaning those who could not afford a $1000 license fee for Xenix or BSDI. But anybody who was interested in Unix at the time was not a regular person. "Regular people" or "desktop users", as the term is commonly used, means those who cannot program or do sysadmin, and don't want to learn. That's precisely why Windows and NT became so popular, because Microsoft gave the illusion that it didn't require configuration or troubleshooting, you could just use the friendly dialogs and wizards and avoid "programming". Of course, that's a loaded term in itself. Writing an Excel macro or setting the myriad dialog options comes pretty close to de facto programming.

[Adam] Okay, I see your point. I guess I'm a "regular person" in terms of automobiles, which I do not understand, could not fix if one broke down, and don't plan to learn. But I just figured if you can use the three "basics" -- office suite; email; and web browser -- on an operating system, it's "ready for the desk top." Open Office or the KDE Suite, Firefox, and Thunderbird (or the GNOME equivalent -- Simian?)

Evolution. The company that developed it was Ximian (now part of Novell)

[Adam] can run on any Linux distro, I'm pretty sure.

[Sluggo] Yes, all those programs are adequate. Where people get into trouble are the thousand things that can go wrong or require computer-specific configuration -- things we don't even think about but stop a "desktop user" cold. E.g., X doesn't start, your mouse pointer doesn't move, your printer produces gibberish (unconverted Postscript or double-escaped Postscript), you have to mount a disk, you have to learn and run a shell command (which implies learning the Linux directory structure, etc), you have to edit a config file (better not use OpenOffice), etc. We all learned these because we had a strong desire to learn Unix and we picked up the skills gradually over years. But sit Aunt Tilley or a secretary before a Linux box without tech support and it's a whole different story. I thought that's what everybody meant by "Linux isn't (or is) ready for the desktop". The need to set your preferences and bookmarks in Firefox is minor compared to this. I certainly didn't mean to suggest that OpenOffice, Firefox, Thunderbird, KDE, GIMP, etc, aren't full-featured stable substitutes for their Windows counterparts. The only thing they're not adequate for is when you need precise compatibility with complex Office documents, like Word documents with tables or Excel spreadsheets with form macros.

[Adam] From there, it was up to "regular people" to do what they wanted with it. Considering that nearly any type of application, from Word-processing suites, to Photoshop-type image processors, (Gimp), Vector Graphics, heavy-gui email and web browsers, in addition to the "standard" GNU and other Unix-like scripts, languages, filesystems etc. all for free or a low price and a lifetime of free upgrades are offered on the PC desktops and notebooks of "regular people" the world over, I'd say the writer of this article has very little idea of what he's talking about; merely a pre-conceived notion of what type of Windows monopoly "regular people" might be willing to accept.

No, I think he does have an idea of what he's talking about: "We learned this lesson in the Mozilla world. It wasn't until we implemented a very capable migration system in Firefox, which carried over the user's IE favourites, cookies, history, passwords, etc, that regular people started moving over in serious numbers -- and staying (and bringing others over)."

---

[Peter Savage] I've read some of the responces to this "massive" topic,

[Thomas] It would have been nice to have had some context, such as a quote from whichever reply you had read. No matter.

[Peter] Well, I was really trying to say I'm just adding my 2 cents worth ;)

[Peter] but would like to add that sometimes peoples problems seem to lie in them "requiring" a system to work in a way they are used to.

[Thomas] Sure. But there's a difference in requiring and demanding, which is what the author of the article bordered on. Requiring something often has certain levels of flexibility associated with it. The software engineering process, when developing software undergoes two things: validation and verification. That's where requirements are met, and their limitations taken into account with the people whom will ultimately end up using the software.

[Peter] My point was that some of the time the "requirement" is not actually a requirment, just a "desire" of the user. In some cases that I have seen it's been shear narrowmindedness to the point of,

"Does it do it this way?"
"No, it does it a different way"
"Oh, I don't want to know then!"

[Heather] There are plenty of people who feel a need to do business presentations, but find the present offerings of presentation software somewhat scary. Why? because they are complex - they didn't start that way, but they got that way over time, because the power users who thought up interesting things to do were also the ones vocal enough to Fuel Demand (in other words, to tell companies that they would part with hard earned cash, if the app did $feature).

This sort of creeping featurism seems to apply to many varieties of app. In the open source world, the burden of a complex app drives some developers to stomp off and write a "better" one that meets their desire for a simpler app. So we've a splash of email clients at varying levels of either complexity or resemblance to a UI popular elsewhere, because "simpler" is in the mind of the beholder. The same thing happens in the commercial world too, leading to a pile of apps named $brandname Express or Lite, and trial editions containing the most popular 2 tasks out of 5 things you might really want to do with it.

Searching for the app that only does what you want and no more is just another complexity. In order to even attempt it, you have to want something.

Power users of a given app have learned its language, as much as Ben and I have learned perl whilst Jim learned Awk and Mike groks what parts of Python actually contain more magic. http://www.jargon.net/jargonfile/a/AStoryAboutMagic.html

We're spoiled though - in our techie culture, while it's a burden to take some time to learn a new language, we sometimes take it as an adventure, and do it, because we want the special abilities offered when it's done. A less adventurous soul looked twice at MS Word and clung firmly with both fists to the Word Perfect he already had - and thinks exactly the same of OpenOffice. They no longer care about simple - they care about not losing features, and not losing too much time during the learning curve. They find reviewers who aim at "is it simple enough" just as frustrating as we do, because it's not simplicity that dogs their learning curve, it's familiarity and that can only be earned with some time behind the wheel. What he wants is to know that he's not wasting his time.

Is the car ready for the road? Depends on the road and the driver.

[Peter] Whilst in a business environment time is money and time spent learning a new system will cost the company money, it seems as if people are not open-minded enough to see that there may be a better solution to a problem.

[Thomas] But it comes back to the same question: "which interface is right?". There is no knowing what's familiar to any groups of users, let alone intuitive (although it could be argued that intuitiveness is not a necessitate of something unfamiliar in this context.) Read Mike's article and my comments about Schneidermann to realise that it's a huge field. Personally, though, a business will always have some training skills involved with software.

[Peter] I take on board your exact point here about which interface is right, but firmly believe that if the interface is flexible enough then the problem is less of an issue.

[Peter] Many problems that I hear people complaining about today, cause me to mutter silently "wouldn't happen if you had a linux box."

[Thomas] Well, that depends on the solution. There are quite a few things in Linux that are more cumbersome to do, than if I had to do it in Windows.

[Peter] Again, I agree here that both os's have advantages and disadvantages, I tend to use which ever I need for whatever task I'm trying to achieve, but it is frustrating to hear people complaining about an operating system day in, day out, when they are not ready, "mentally", to consider an alternative.

[Peter] Certain differences between the linux and doZe os, have firmly rooted in my mind which side of the fence I'm on, not to mention the ethical and moral standings. I hope people can see what I'm getting at here

---

This page edited and maintained by the Editors of Linux Gazette
HTML script maintained by Heather Stern of Starshine Technical Services, http://www.starshine.org/

---