Linux Needs More Haters

Corrupt brings us a ZDNet column by Jeremy Allison, who says Linux could benefit from more "tough love" in order to improve its functionality and popularity. Excerpting: "As Elie Wiesel said, 'the opposite of love is not hate, it is indifference.' LinuxHater really doesn't hate Linux, despite the name. No one takes that much time to point out flaws in a product that they completely loathe and despise. The complaints are really cries of frustration with a system that just doesn't quite do what is desired (albeit well disguised). A friend pointed out to me that the best way to parse LinuxHaters blog is to treat it as a series of bug reports. A perl script could probably parse out the useful information from them and log them as technical bug reports to the projects LinuxHater is writing about. Deep down, I believe LinuxHater really loves Linux, and wants it to succeed."

Comments

OS X

[dan dan the dna man]

OS X is the 'tough love' that Linux needs. I use Linux on the server (although I have a rack of Xserves too) and there's a reason I am happy with it there (unlike OS X).

On the desktop? Well I use a Mac. And I don't think I will ever go back (in the interests of fairness this is being posted from my 'Games and things' XP laptop).

I love the fact Linux is dynamic, and open source. I really do. I don't like the fact that it doesn't seem to 'evolve'. The fragementation of WM's, distro's etc. never actually seems to weed things out. What we never end up with is a 'de facto' solution.

People argue that choice is good. I'm sure it is. But the reason that Windows and OS X still beat Linux on the desktop experience is because they are standardised - there just aren't alternatives. And OS X is a better 'desktop Unix', so as a person who wants that, where else am I meant to go? If nothing else KDE 4 would drive me away... yuck.

I did use Linux on the desktop. For several years. I only tried OS X on a whim.

I don't hate Linux, but I don't think I'm alone. Go to a confernce these days (I'm an academic) and I used to see people booting into myriad versions of Linux as they opened their laptops. These people are now in a minority, as the Apple logo is raised in unison at the beginning of any talk.

Fanboy? Maybe.

Re:OS X

[Anonymous Coward]

Sigh, another heap of bullshit. There are really three simple reasons why Linux have difficulties becoming mainstream in the environments you describe. NONE of those have anything with any of that "standardization" bullshit you seem to believe in.

Reason #1. Linux doesn't generally come pre-installed. If you want it, you've got to install it yourself. Usually after you've already paid for another OS. Don't have time, or can't be bothered? Get your ass into the folder your OEM have decided for you to be in.

Reason #2. All OEMs fall over themselves to get stuff working with Windows - in the case of Apple, obviously they fix it for you. Not so with Linux. In fact there are plenty of OEMs who seems to deliberatly try to make life difficult to use anything but the original, OEM approved OS.

Reason #3. ISVs reluctant to try new markets until it's obvious that they are on the verge of becoming obsoleted. (Don't feed me that "all distros are incompatible" line, it's horse shit). In fact, the way for instance Adobe behaves, one might actually start to wonder if there aren't cheques coming in from certain parties in order to assure that some applications stay off the Linux platform.

These three reasons are basically all there is to it. If Linux had shipped pre-installed, and OEMs didn't put obstacles in the way, I bet we would see a lot more of it. And if some ISVs actually grew some backbone, instead of cowardly assisting with the vendor lock-in, we'd probably see even more of it. It has NOTHING to do with the fact that some people prefer a wm, some gnome and yet others prefer kde on their desktop. Just get your damned libraries in line, and you're home. Anyone with a IQ above that of a log can see that.

Re:OS X

[Magic5Ball]

The main difference I've found between commercially developing for Linux and commercially for OS X is that suggestions for improvement generally get the response "that's a problem with [other part of the stack]", where the [other part of the stack] gives the same form of response, and the pointers rarely dereference to anyone who takes responsibility or ownership of the issue.

A fine example currently is asking for modal dialogs to be dismissable on 400 px tall screens of UMPCs ("Please let me scroll the dialog, or locate the OK/Cancel/Apply/Help buttons on screen where I can see and click them"). The echo chamber of "talk to (app | xorg | GTK | Intel | Java | Nvidia | distribution | libc | vendor | etc.)" really turns me off of wanting to help any of them resolve the issue. As a developer on Linux (but with OSI Layer 2/3 stuff, rather than GUI things), I could probably spend a week or two to figure out how all of those pieces interact (without deeply understanding the design philosophy or project plans) and make patches that would work on my current setup, but that would generate significantly more regression testing and QA load than would be required if the patches came from in-house where the developers are already intimately familiar with their own code. Also, as one of the advantages of a package management system is supposed to be that it all gets taken care of for me, I have no interest in maintaining my own versions of app, X, GTK, etc, nor do I want to spend a day each understanding the 32 to 200 KB spec files that build those packages, nor am I interested in waiting 3 months to year before the fix makes it into the non-beta parts of the distributions I might use.

Now, ask grandma to change the screen resolution back to 800x400 (using a dialog the entirety of which she can't see or access) after she experiments with plugging a regular monitor into her new eeePC (or whatever UMPC the banks are giving away these days) and it stays mirrored at the new resolution after the experiment. It's unreasonable to expect that random non-technical user would want or need to understand that the entire stack around the problem even exists, let alone attempt to fix it.

By contrast, if I encounter an issue manifesting in CoreFoo, Cocoa, some kext or library or wherever else on OS X, Apple will offer to take ownership of it even if it isn't directly their problem (and then work on it in the background), other vendors/developers in the stack will at least acknowledge if not fully investigate the problem in the test case(s) submitted, and random other developers in the same space will be thankful for the new knowledge instead of responding with RTFM or its analogues. Granted, fixes in OS X still take weeks to a couple of months to widely roll out, but that's still faster than many distributions update their stable packages.

And then there are (the comparatively few) great OSS people like Tim Waugh, who knows the (printing) stack up and down, and responds with a reasonable fix or workaround within 48 hours, even though the problem is not in his part of the stack, regardless of who's customer you are. Plus, you'll usually get some insights by responding to his "I'm curious about what you're doing with this" follow-up.

The Pleasure of Hating.

[delire]

If Jeremy is correct, then the author of Linux Haters has chosen what is possibly the least likely route to garnering interest from Linux developers. Which linux developer would consciously choose to read a blog that refers to them as a 'luser' incessantly from paragraph to paragraph.

The 'benchmark' OS he seems to use as the basis of the bulk of his criticisms is OSX, an OS I find really frustrating to use (and I use it fairly often these days). If I were to start an OSX Haters on this basis should I expect the Aqua and XCode authors to read it daily in the interests of improving all the braindead things about both those aspects of OSX? Didn't think so.. Maybe the guy just has a crippling case of Internet Rabies induced by deep boredom and Jeremy's simply being a little generous..

There are, afterall, blogs featuring meticulously prepared images of meals that people hated eating. Perhaps this blog is simply in the same vein; just another masochist whiling away the hours in public.

Must be a slow news day.

You have to care to hate

[Maxo-Texas]

This is important to learn in life. When you reject someone that loves you, then they hate you. As long as they hate you, they still love you.

Once they don't care any more then it's over.

It discovered this all on my own when going through a bad breakup so that part of the comment particularly leapt out from the page to me.

Re:Don't Turn Blind Eye To Complaints

[dargaud]

Of course, linked to that is the really annoying challenge to "Just fix it yourself! You've got the source!" That's an absurd claim.

Yup. I write, among other things, device drivers under Linux for a living. But each time I take a Linux graphical app and try to make some changes to it, it fails. Wrong compiler setup. Wrong libraries. Wrong rpm. Wrong system config. Wrong wrongness.

It's to the point that there are only 3 types of Linux progs that work: the one that comes with the system (and its updates), the simple "./configure; make; make install" and the kind I write myself. Any of that "fix it yourself" is crap.