QUOTE: You're probably right - most of the reboots in Linux can be avoided while still achieving whatever the goal is by an experienced admin. But this is where Linux makes a decision: Become more like successful, mainstream OS platforms and hold the hand of "Aunt Tilly", or remain an OS with 2% market penetration and solely appeal to uber-propeller-heads like yourself.

I don't see how one has to make a choice. You seem to think that Linux-based systems must abandon the ability to restart core services without rebooting to serve the "mainstream" user.

. . . or maybe you just think that I'm saying something like Ubuntu shouldn't go for the easy way out when dealing with a "mainstream" user. That's not what I'm saying at all. I'm saying that you really don't have to reboot, at least in the vast majority of cases -- technically.

Something tells me this "misunderstanding" of yours was intentional, though, so you could troll me (and other people who recognize that most of what you say is smoke and mirrors) for flames again.

QUOTE: The limitations on Linux adoption are based on difficulty.

The limitations on adoption of Linux-based systems are many and varied. Some difficulties are part of that; they are not the sum total of it and, if the major Linux distributions still offered all the benefits of Debian in 2003, of the popular desktop environments for it in 2005 or 2006 plus a few refinements, and of automated hardware configuration available to the top Linux distribution contenders in that arena in 2010, I think the few difficulties remaining would not be enough to really hold it back. I think the biggest problems with Linux adoption in the past half-dozen years or so have actually been license incompatibilities, image (partially deserved and partially applied maliciously by outsiders), and the fact that for every major step forward in that time period there has been at least one step back -- plus an additional step back if you consider the fact that even its problems haven't been consistent during that time, making it difficult for any documented solutions to have longevity.

QUOTE: In order to be less difficult to deploy and support, Linux must embrace a dumbing-down - and this whole article is about how it can't play BOTH sides of that fence while delivering everything that everyone wants.

It doesn't need dumbing down. It needs easy adoption, which consists of reaching installation parity (that is, widely available preinstalled so that hardware issues disappear), plus easy continuity, which consists of not setting out to break everything on an eighteen month cycle. Some of the reason these things are still problems is the simple fact that people are racing to see who can dumb things down the fastest.

QUOTE: Linux wants to be an "elitist" platform that only appeals to guys who want to learn how to restart a daemon in init.d rather than reboot - that is great - it is awesome.

I can only assume by this statement that you haven't engaged in any activity within the larger Linux community in the last five years. Quite the opposite is the biggest problem there lately. It has become a community where stability doesn't matter any longer as long as every single piece of software in the world is installed; where software that makes your decisions for you trumps decisions that will not actually anger users; and where nobody cares whether an interface is consistently usable as long as it has things that spin, waver, flip, or just generally look different in surprising new ways. In the dwindling cases where parts of the most popular distributions are still consistently usable, where the ability to make good decisions (or even have them made for you with possibility of overriding automated decisions) still exists, and where stability can still be had, it is only because the newfangled BS has not completely replaced the "hard" stuff -- which means that to the extent Linux-based systems still have the benefits that made it a better choice than MS Windows in the first place, it's only because nobody has done anything to make things more accessible to "mainstream" users.

I have always been of the opinion that the proper approach is to offer the fine-grained control more technical users want and (not or: and!) the "easy" front-ends that make "mainstream" users happy. The problem is not that the Linux world isn't trying to cater to the "mainstream" users; it's that the Linux world is catering to the "mainstream" users in some ways that are incompatible with catering to it in other ways and, at the same time, completely screwing over the more technically inclined users.

QUOTE: With Windows if you're running a system built entirely on HCL approved systems, it isn't a matter of "luck" either - if you've gone to the trouble of actually researching and purchasing around those factors.

I don't care how carefully you select hardware: you're still going to run into apparently nondeterministic behavior of the system, built-in security weaknesses, and other problems that guarantee a ceiling on how stable, secure, efficient, and productive a system you can have -- a ceiling that is notably lower for many, many common conditions than other systems provide. When Microsoft does stupid crap like provide a(n undocumented, of course) system call that can actually just switch off privilege separation, that's a noisy friggin' klaxon warning you that architectural privilege separation has still not made it to the world of MS Windows.

QUOTE: Most people who have a system, especially a DIY Linux box - aren't paying attention to HCL lists when they build their system. If they meet those requirements, it *is* a matter of luck.

You're talking about a different set of circumstances than I was talking about, so this doesn't actually answer my point at all -- but I'll address it anyway.

When you can get open source Unix-like systems that have not been utterly hosed up by their own developers (I'm looking at you, major desktoppy Linux distributions, when I say such unflattering things) preinstalled by reputable vendors with the same level of support provided by those vendors for MS Windows, then you'll see that there's zero luck involved in what benefits there are in MS Windows' supposed benefits in this arena. In fact, given the parameters I've specified here, what you'll see is MS Windows looking like tawdry trash beside the alternative.

QUOTE: Your last comment makes me think we're arguing apples and oranges here, though - because being able to replace your GUI interface isn't even part of the equation as I'm discussing this aspect.

You said "Or maybe you're just lucky. Maybe you've got a system built on ideal hardware and you're running an ideal set of apps that maximize the stability of your distribution of Linux for you." You brought up applications. How the heck is me mentioning applications not relevant to what you said?

QUOTE: non-sequitur with bonus straw-man is non-addressable.

Poppycock. You use these terms without any evident understanding of what they mean. You said "It is absolutely *insane* to claim that the OS platform itself has any consistently and inherent superiority over another in terms of stability and reliability in this regard." What about "absolutely *insane* to claim . . . any inherent superiority" of one technical implementation of an OS over another is not addressed by my observation that you tend to do things like notice a moderate hardware related effect on stability then blow it way the hell out of proportion so you can call people "absolutely *insane*" for having the audacity to notice things like the fact that open source Unix-like OSes automatically clean up after memory leaks in ways MS Windows does not, have architectural privilege separation that DRM software cannot blithely bypass, and enjoy a more active and responsive vulnerability fixing "workforce" than MS Windows (to pluck three items at random out of the bucket)? How is that possibly not an excellent example of how you note some moderately affecting circumstance, turn the molehill into a mountain, then use that to make literally absolutist proclamations of extreme judgment on people for noticing you ignored important factors that really do show specific areas of inherent benefit to one system over another?

That's no "non-sequitur", and it's no "straw-man" -- and, by the way, a straw man is equivalent to a non sequitur (both are off the point, the only difference being whether you refer to the off-point comments as a whole or just the part someone addresses in place of your actual argument); repating yourself in different words does not add up to there being more problems you've (mis)identified in what I said.

QUOTE: Your argument would have been stronger without this paragraph - it comes across as a troll.

Freud would have called this "projection". I think it's just a cigar.

QUOTE: Weird, after all that, you're basically *actually* saying I'm right.

I have never said Linux-based systems are perfect. I've just said that all your posturing about how "Linux" will destroy the world, and all your completely off-base mischaracterizations of things -- making up problems to complain about because, apparently, you are wholly unaware of real problems -- are exactly that: posturing and off-base mischaracterization. Saying that I'm actually saying you're right is like saying that two people arguing about whether Pluto was declared "not a planet" was due to a rational set of rules for identifying planets or due to jealousy from Eris are actually in agreement on the crux of the issue just because neither of them wants to declare Pluto a planet.

In fact, in this case, I was specifically responding to the fact that you said Jack Wallen has moved toward the "center" on Linux-related issues: I explained that it's easy to do so when Linux-based systems have gotten worse over the last few years, not that you're right about MS Windows being better than it actually is, Linux-based systems being "bad" because they're open source (and thus prone to mutating into weaponized flu viruses), or some other cockamamie nonsense you've spouted in the last few years.

QUOTE: If Linux is going to be more accessible, it is going to embrace the GOOD of Windows and OS X - which are things you see as black and white bad.

This is like when you claim I want you to be censored by jackbooted gestapo agents. No. That is not what I am saying. You should really put a sock in it when you feel the urge to mischaracterize my statements as the kind of thing you'd say (but on the other side of some dividing line or other), just for the sake of not looking like an arrogant ignoramus who doesn't bother to read what I say before responding to it. Note that I'm not saying anyone should draw and quarter you to prevent you from voicing an opinion, before you go accusing me of being a censoring bully again; I'm just saying that if you exercise in a little discretion before saying something stupid, people will think you're less stupid.

QUOTE: You can't have both mass appeal and hardcore legitimacy.

It's difficult, but possible. It's not impossible as you imply. It just requires refinement, rather than syphilitic promiscuity and gold lam. (Just in case that doesn't print properly on TR, that's supposed to be "lame" but with an acute accent on the E.)

QUOTE: You, on the other hand are at the point where you're going, "Halen *sucks* without David Lee Roth. It isn't the same band at all. It is half the band it used to be. They better change things around and find what they used to have, because if they keep going the way they're going, just some commercial bubble gum pop-rock band - then I might just be done with 'em".

1. Van Halen was never all that great, but at least they were fun with Roth.

2. Van Halen was much more successful with Roth than with Hagar.

3. No, seriously, your OS doesn't have to be afflicted with bovine spongiform encephalopathy to appeal to the "mainstream".

4. My reasons for having actually given up on Linux-based systems in 2005/2006 have nothing to do with the fact it was trying to be more widely appealing, and everything to do with the fact that people did it by attaching parachutes, mid-1980s boomboxes, and pink taffeta to a Ducati Monster, removing half the engine with a plasma torch in the process, rather than by actually making it more usable for less-technical users.

5. (Sorry, I'm getting tired of typing.)

Seriously -- do you think that trading "difficult to configure without technical knowledge, but stable and useful after that" for "self-configuring, but the configuration is wrong 30% of the time, even when right it breaks itself another 50% of the time, and in no case do you have any reasonable recourse to fix the problem" is an improvement?

QUOTE: You've never made a single post before that was a *better* example of illustrating *exactly* why my claims about Linux aren't exaggerations or misinterpretations, but grounded, realistic observations about the challenges and limitations that Linux faces going forward.

It's nice you think we can agree on something. Too bad you're wrong. Almost none of the commentary I've seen you make about problems with Linux-based systems agree with anything I've said about problems with Linux-based systems. Two key points about where we disagree are the fact you blame many of the problems with popularity of open source Unix-like OSes on the fact they're open source and the fact you generally refuse to acknowledge that there are actual technical differences between MS Windows and various Unix-like OSes that have real, substantive, significant effects on the various benefits and detriments of working with different OSes.

QUOTE: Mass appeal is a measure of success and of value and of suitability.

It's not always a good measure, though -- just as many microbenchmarks are truly awful measures of the real-world performance of a given programming language implementation for people writing real code for real production use.

QUOTE: That applies against my own biases too - where I would like to see Android do better against iOS in the tablet space, but it still lags behind.

Apple actually put some work into ensuring iOS would have what it needed to be a strong contender on tablets, and leveraged its fanbase's (err, "userbase's", I mean) credulity (uhh, "aesthetic sophistication") and idolatry (sorry, that should have said "loyalty to a company that has earned their respect") to get quick uptake among not only users but also developers. Meanwhile, the vendors selling Android tablets basically just threw a smartphone OS on a tablet and figured "Branding will do the rest," thus sabotaging their own efforts. They've finally started making a real effort to do something meaningful with it, but I wonder if the tablet market will even be viable long enough for them to make the long, hard slog past the damage done by their own insipid gaffe.

QUOTE: The Linux community still needs to improve in this regard, if they ever want to meet their challenges in a meaningful way.

I've arrived at the opinion the Linux community, while it includes a lot of very smart and reasonable people, is a lost cause because it also includes a lot of stupid and unreasonable people, and the latter are increasingly running the show. I kinda hope the smart, reasonable refugees will come to various BSD Unix communities and work on broadly "user friendly" stuff in a more measured, intelligent manner, and pick up the scent where the Linux community at large lost it a long time ago. The only people married to Linux as the basis of an OS these days who are doing good work are (as far as I've seen) drowning in a sea of stupidity, engaging in misguided efforts to build something good but not broadly "user friendly" on top of the mess, or actually just working on the kernel and drivers without bothering to get their hands dirty with the user-facing stuff at all. Examples include a few good people fighting the tide at Canonical (I suspect they had something to do with the initial development of the HUD feature), the Arch Linux project (I'm guessing, though I don't have enough experience with it to be sure that fits), and of course some of the people in the core Linux kernel team.