Log in

No account? Create an account

Previous Entry | Next Entry

FOSS Malaise

This is not a rant; it's lamenting. I got back from IDF at 2:00 in the morning: I'm low on sleep--perhaps this is attributable to my feelings at the moment.

I have been trodding along this path of sticking with FOSS for some seven years now; operating under a principle that things that do not match feature-for-feature proprietary alternatives will soon through some inherently supreme kind of free-market force bringing features where they are lacking in Free Software. When I was younger and more idealistic, I bought in to much that LinuxHater has maligned about Linux fanboys including this precept. Even though I am an atheist, I am like a man that buys in to a particular religion at an early age and doesn't often stop to reevaluate that position: to do so is wrought with intellectually exhausting soul-searching and always tends to prod one back from radical rethinking because to do so carries an implicit admission that you were bamboozled.

Well, I was bamboozled by those retrospectively naive arguments. Like a puritan enduring a trail of the flesh, I find myself today weak-willed: I want to give up on Linux on the Desktop for greener pastures like so many of my colleagues have--for Mac OS X. Having admitted that--however--I still find that there are mature arguments and pragmatic value to be found among the FOSS community of products, most of the time. Sadly, at the moment, the Desktop is not one of those areas. I know--I'm not saying anything that hasn't been said before.

This discourse is an attempt to talk myself out of giving up and switching to OS X.

At the moment, we have an ecosystem of free software developers who--in their free time--manage to sustain the bare necessities required of a desktop environment. I'm not fingering a particular DE here, merely suggesting that no one project innovates in any particular way: we're all stuck on copying Windows 95.

Recently, Joey Hess likened the choice of a window management paradigm to a cult:

Another way to look at it is adopting a new philosophy. Or, in some cases a cult. (In some cases, with crazy cult leaders.) Whether they use Windows or a Mac, or Linux, most computer users are members of a big established religion, with some implicit assumptions, like "thy windows shall be overlapping, like papers on the desktop, and thou shalt move them with thy mouse".

So, changing to a new window manager is a process of being dumped into a different environment, where nothing works like you've come to expect, and trying to construct a mental model that you can use to make sense of it. But it's also a process of modifying that environment to behave the way you like.

I spent a great amount time of my time at IDF looking at radical new interfaces up on display for a new class of devices known as "MID's": Mobile Internet Devices. They're a kind of cross between a laptop and an iPhone: mid-range, low power processor with a 3D accelerator. Each of them (with a few exceptions running XP), are running Linux. Many are based on MobLin. And here's the sad part:

Every single one has evaluated the state of every existing Linux desktop environment and found them wanting: bloated, boring, and more of the same. Even more sad, every company appears to have decided do their own thing, to take existing libraries like Gtk+, Qt, Cairo, Clutter, Popler, Webkit or Gecko and mashing them in to half-baked UI's that are a semblance of the immaculate, integrated love-fest that is the iPhone user interface experience. However, in each case, they come off as either half-baked compositing-based solutions or exercises in poor-performing pixbuf abuse.

In my mind, we're all doing it wrong. And it appears that no company cares to pay to put people on their staff to drive a vision of a radical new desktop paradigm that's not proprietary.

I'm going to look around. Maybe Joey's on to something about this "windows cult" thing. Obviously, people are slathering drool all down the front of their iPhones and no windows are in that user interface. I'll give it a try on the desktop. Even though I'm shopping for a new user interface, the underlying technologies that we rely on to make great products that I use every day are indisposable (ie. most of the libraries that make it in to MID's listed above).

I didn't mean for this to turn in to a "decadence thread" continuation but I guess it did. Sorry. I'm just not a happy Linux desktop user at the moment.


( 15 comments — Leave a comment )
Aug. 23rd, 2008 04:47 am (UTC)
A mistake
My part-time job is supporting Macs in a graphics design lab. Let me tell you, you don't actually want OS X. Apple gets a free pass on design and usability for some reason I cannot fathom, probably because people are simply too invested in Apple's image to honestly reflect on the quality of what they produce.

The iMac is a perfect example of this WTF-ery, perfectly mocked by Penny Arcade.

As for the MID stuff, it depends. The Ubuntu Mobile and Embedded stuff was apparently Flash driven, and frankly, sucktastic in my experience. Tester hostile. The netbook remix stuff is nicer, but also not very well staffed. I don't know how Canonical is funding it, but it seems like two part time guys who don't dare run development versions. The concept works very nicely for my regular laptop sized tablet, but a) is very new and therefore buggy, and b) exposes some flaws in icons. On the plus side, they do accept patches, if a bit slowly.

The netbook remix seems fairly standard -- a devilspie script to simply maximize windows, and a smarter window picker for the circumstances. It makes sense to not engage GNOME over something they already threw out, at least to me.

One thing I'd really like to see in GNOME panel itself is a layout feature. Have a left applet queue, right applet queue and center applet. That would help with Xrandr, I think, where layouts get messed around because the panel suddenly became constrained and then relaxed.
Aug. 23rd, 2008 02:21 pm (UTC)
Apple is not necessarily a greener pasture
I use OSX a lot at work, and I can tell you honestly that I use Linux and KDE/GNOME because they piss me off less. I admit, I was attracted to the Mac and its OS a while back, but OSX has its own collection of WTFs, like freezing the UI for silly reasons, and I find myself yelling at Macs more than my Linux machines. If you really want a new interface, play with things like awesome, xmonad, and other unusual window managers.
Aug. 23rd, 2008 02:51 pm (UTC)
i hear you ...
i just wanted to comment on the bit about MIDs and your observation of windowing systems approaches for them from the FOSS world. (i won't comment on the rest, even though i think it's similarly interesting =)

it's really hard to do anything new on "the desktop" due to user intertia. describing people's attachment to the current concepts as religious is pretty close to accurate IME =) in plasma we're slowly introducing new concepts bit by bit, but it's a long and painful road.

MIDs and other devices with new physical form factors can skip this problem as people don't have the same preconceived notions about them. hallalujah! =)

the trick, imho, is not to run off and create something completely new and separate from "the desktop" for them, though, as that will just divide up our ecosystem into different buckets of effort. i'd like to see us create something that can provide a productive and productively interesting experience on these devices which can also provide relevant solutions on "the desktop" as well as other devices (the CE world hardly begins and ends with MIDs, after all).

a lot of the design we've undertaken with plasma has been towards this idea, so the stuff we wrote on "the desktop" actually runs already on devices like MIDs ... but most usefully we can provide with minimal efforts a completely different primary user interface to these same components .. without rewriting everything (or pretty much without writing anything).

we've just started work on our first MID-specific primary user interface; it's not at all like "the desktop" plasma shell, but it's using all the same code and components underneath. we will end up writing a few custom widgets and components for it as well, but with scripting languages well supported, 3rd party systems working (mac os, google gadgets, edje components), integrated on-canvas components like web browsers or contacts and the ability to provide an attractive non-windowsy shell around and for the components we will be able to deliver something that's both non-sucky but also still relevant to everyone who isn't interested in MIDs due to the cross-over at the component level.

as for the "everyone appears to have decided to do their own thing", that's why we've worked hard to make plasma a universal canvas so that you don't have to choose between, say, E17 edje files and plasma, or web content centric interfaces and plasma or google's igadget api and plasma or .. and it's all 100% open and Free, and 3rd parties (like Bossa and Google) are contributing hooks for their stuff.

looking to where we want to go beyond just the primary user interface, though, we've been doing huge amounts of work to integrate services and more recently contextual awareness into the underlying frameworks so that we can provide not just something that competes, but which can take the experience of these devices to a new level.

this probably sounds a little bit like a ranterific infomercial of some sort for my pet project, but when i read blog entries like this one it's hard to resist: i get all excited and start jumping up and down in my chair going, "yes, yes, YES! that's exactly it! and we CAN fix it! and here's how some of us are trying to fix it right now!"

as a gnome dev and user, i'm not sure plasma will be of direct interest to you, but i hope at least that the idea that there are people doing something right now about the exact issues you see as discouraging and problematic in the FOSS world might cheer you up at least a little bit. maybe. hopefully.


Aaron Seigo
Aug. 23rd, 2008 03:07 pm (UTC)
Not everything it's cracked up to be
I love Macs and OS X, and for certain use cases they are fantastic. But if you are used to Linux, there are going to be things that bug you (and it will probably take a while for you to notice).

1) It's going to bother you that modifying your system in significant ways is difficult, undocumented, dangerous, etc. You won't care at first because so much of OS X seems just fine, but eventually you will run into things that don't work the way you need them to. This also leads me to...

2) It's going to bother you that most pieces of OS X and most popular apps are not open source. Again, for a while you won't care, but then you'll run into a bug and you won't know what to do! You certainly can't fix it yourself. There may not be any meaningful way for you to report the bug, and even if there is there may be no way for you to follow up on its status short of trying out every subsequent release.

3) OS X is always behind on important system packages (Python used to be really far behind, for example) that you take for granted, forcing you to begin the process of setting up your own environment, which leads me to...

4) Fink and DarwinPorts and all of those are great, but they just never work as well as package management on Linux. Eventually they will let you down when it matters.

That being said, OS X is quite lovely and there is an active community of open source hackers, so you wouldn't find yourself to be alone. But I've lived extensively on both sides. You already know the benefits (a nice desktop, a lot of "shit just works", consistency, nice hardware, etc), but make sure you know the drawbacks before you make any decision. To me, it's worth having Linux and OS X prominent in my household.
Aug. 23rd, 2008 06:14 pm (UTC)
Re: Not everything it's cracked up to be
This is consistent with my experience of OS X as well. I tried to switch, found it too slow and frustrating and Not Debian.
Aug. 23rd, 2008 03:22 pm (UTC)
yes, but..
I think that if you look back over the last 30-odd years of UI development since XEROX's PARC did it's thing, the brutally obvious conclusion is that users really just aren't interested in vastly new UI paradigms. People are interested in working out that paradigm one time on a given type of hardware (this case being - i'm not exactly sure on this - but maybe computers with keyboards?) and then not thinking about it any more. What you've noticed about the iPhone is probably that Apple has been the one to determine what the rough UI of small devices will be, the sort of XEROX PARC of that arena.

That said, it makes sense, and isn't even all horrible news, though to people like you and I naively underestimated how difficult it is to convince people to want something new. Though we often forget about it in the depths of philosophical conversations about desktop interface paradigms, it is really easy to forget that a computer is a tool for getting things done. People want to be able to sit down and use their computer to do something in particular (I don't always do this, and I'm starting to think it's an unhealthy addiction thing). Sure, current desktop metaphors have their drawbacks: files end up disorganized and hard to find, windows use screen space inefficiently, etc. etc. But these problems typically cause a very small amount of extra time and brain power on the part of the user. Learning to be truly efficient and comfortable in a new OS or desktop environment, however, is a pretty big time investment, and really probably isn't worth it to a person who is not interested in computers and just wants to get things done.

I dunno. There's a lot of handwaving up there. I guess you can take it or leave it :)

Aug. 23rd, 2008 05:20 pm (UTC)
How much ...
... of this[1] do you think is possible with the current FOSS technologies that are available? I think the focus on frameworks is what is lacking when it come to consistency and innovation. You should not really worry about the many different paradigms that are out there as long as it is possible to accomplish them within the available UI frame works ... based on the attainability, accessibility and ease of use companies and developers will use the development method that works for the widest range of target platforms.

Aug. 23rd, 2008 06:07 pm (UTC)
I did it
Hi Jason,

I did the switch, after being a long-time GNOME and kernel module contributor and maintainer. I used (and later developed) Linux out of frustration with Windows and idealism with FOSS. The switch to OSX came for the same reason as you: the never-reached feature parity and the feeling that new features come in sooner than "we" can implement the old ones.

Linux will not reach feature parity at any time in the near future. It's inevitable that it will eventually, but that is only when companies such as Microsoft and Apple release their new proprietary products as a cross-platform, Linux-compatible product. Until that point, we will have to reverse-engineer their latest inventions and will therefore inevitably trail and fail.

MacOSX has its quirks, but it's wonderful to be able to watch the NBC olympics live online on my TV out and see it "just work" on my TV - and it doesn't have the quirks that moved me away from Windows. It's so simple in words, but this stuff just doesn't work on Linux, and whatever technique they use to broadcast the olympics 8 years from now will not work on Linux either, unless Microsoft blesses it or the IOC stops using that technology, neither of which is likely to happen...
Aug. 23rd, 2008 06:21 pm (UTC)
You seem to be conflating two separate issues here.

One is the success of free desktops in the world at large. This is obvious and true for all the reasons you suggest. Ultimately there's just no way for a free enthusiast product to have the integration of a locked-down commercial product. I wouldn't inflict Linux on my mom, and that's fine.

The other issue is what to use for yourself. For the same reasons OS X works for so many I find it doesn't work for me, because I am not a normal computer user. The one that killed me was window management, but what killed me more than that was that not only was there no way to change it (that's fine enough, Gnome doesn't have many options either), was the dawning realization that there was no recourse for me: I could either take what they had provided or buy a new computer, with nothing in between.

Finally, there is the religious argument. I use free software in part for moral, non-pragmatic reasons, and that doesn't bother me at all; I also carry my litter to trash cans rather than on the side of the highway, even if I'm in a place I'll never return to, because of the same sorts of moral-but-not-pragmatic decisions. I don't buy an iPhone because I can't morally support giving money to that kind of user-hostile lock-in.
Aug. 23rd, 2008 06:22 pm (UTC)
(PS: I'm a happy user of a tiling window manager. I also thing the overlapping window thing is overblown.)
Aug. 23rd, 2008 09:02 pm (UTC)
Use a Mac at work
I have to agree with what the earlier anonymous commenter wrote, because I'm in the exact same position. When I got my Mac at work, I was totally expecting to grudgingly concede that it was nice. I was very unpleasantly surprised to find out just how unintuitive it is in a lot of cases. No using "alt" keys for keyboard-based menu navigation. Weird symbol for the "option" key, which is the exclusive symbol for that key in menu shortcuts, but isn't printed on the keyboard. Fink and DarwinPorts are nice, but neither is as good as even a middle-of-the-road Linux distribution repository at this point. Prepare for finger retraining on things you may take for granted like shift-Home, Shift-Pgup/PgDown, etc. And don't get me started on the lack of a right mouse button. As for hardware quality, I don't think my laptop has aged well, and prepare to have a doctor's office-like wait to talk to a "genius" at Apple's Genius Bar.

Many will point out that Linux has just as many and bigger deficiencies. I'd argue it's more of an apples/oranges comparison (pardon the pun). The things that piss me off about Linux are things I've come to accept, and can at least feel hopeful about getting better, even if it's slower than I personally prefer. There's no good reason to believe that the things that piss me off about Macs will be any better five years down the road.
Aug. 23rd, 2008 09:42 pm (UTC)
I agree with the sentiment, but for me it's not the paradigms that I have a problem with. What I have a problem with is the lack of really quality applications that are a pleasure to use. Sure, it would be great if FOSS innovated and created something super new and cool (Plasma seems to be heading in this direction), but it seems we're always creating super new and cool frameworks, and the applications never appear around them.

I've been wondering lately if FOSS applications need to stop trying to match proprietary counterparts feature for feature, to stop trying to innovate, and just create applications that do the basics really really well. More importantly, to create applications that for 99% of users are a pleasure to use.

Easier said than done I guess, but I have great sympathy for you and your post was timely for me: this very week I've just reinstalled OS X on my MacBook for the first time since I bought it feeling "malaise" over FOSS. I couldn't put my finger on what was bothering me until I spent this week with OS X, but now I know I think I'm ready to ditch OS X again and start working towards realising what I described above. (Mac OS X may stay on a partition so I can watch the occasional movie from iTunes).

My decision not to hang around on this platform has also been aided by me thinking a lot about "digital rights" issues etc. I think FOSS has a significant role to play in securing a future where we control our technology, and not the other way around.

Aug. 23rd, 2008 09:43 pm (UTC)
I don't understand this-- is OS X "more of the same"? If it is, how can you both be wanting to switch to it because Free software is "more of the same"? If it isn't, how isn't it? I'm not sure I've seen anything in the OS X UI which hasn't been cloned somewhere in one of the Free desktops.
Aug. 23rd, 2008 09:52 pm (UTC)
eg openmoko
Your post reminded me of a review I read when the openmoko free runner was launched. It went something like: "Phone calls work most of the time."

The 3g iPhone has had lots of problems, but none as dire as those wrought upon the world by the first free phone ever made.

To make matters worse, the software openmoko stack seems to be a clusterfuck of competing and overlapping technologies.

Oh and has anyone noticed how ugly it is?

Linux phones have been one big epic fail, and I'm guessing mostly forgotten once Android debuts.
Aug. 24th, 2008 06:40 am (UTC)
Random Thoughts
(Three separate points are made, so if you don't like, say, the KDE love bit, just skip that paragraph.)

To anyone suggesting Linux has all of the features of OSX, it's not just just having the same checklist of features. It's also about integration and the entire cohesive package. I have to say, OSX is a very compelling package depending on your needs, and I've been tempted to switch from GNOME to it many times.

Speaking of switching, I have felt the same way described in the post about the FOSS desktop for a while, fueling the Mac-lust even more, but in the end I ended up switching to and following KDE 4 closely, as I feel it represents the best possible future for the FOSS desktop, and I say this as someone who used GNOME from 2.4 or 2.6 or so all the way through 2.22, and hated KDE 3 to probably as unhealthy of a level as is possible without being compelled to be a dick about it online (I just silently hated it, thankfully). It's not "there" yet, but 4.1 is a pretty solid platform, and 4.2 is incredibly promising sounding to me. Most of the gaps that I perceive are easily fillable I think, once a bit more work is put into various bits of glue-tech (Plasma Python bindings, I'm looking at you >.>).

In the meantime I'm left without a comfortable desktop for actually getting work done though, and it brought to my attention one of Linux's best strengths: flexibility and ease of customization. This is especially apparent with some of the different distributions available, such as Arch Linux. I've used Ubuntu since Warty, and have enjoyed it immensely, and my current KDE desktop is Kubuntu, but I recently acquired an abandoned 300 MHz laptop with 64 MB of RAM, and that actually provides a really good work station for me for a lot of my needs, and there's no way I could have put it together on top of anything but Linux, and Arch in particular turned out to be REALLY good for it's flexibility in building exactly the desktop I wanted. It's a combination of Openbox, Conky, remind, todo.txt shell scripts (todotxt.org), vim, and bmpanel, among other things. It's turned into a combination PIM, word processor, and programming environment (mostly Django with Vim and Midori for testing). I don't think the setup could have been done with any other OS (well, Windows, that is, or some other distros like Ubuntu) without a lot of unnecessary other stuff on top of the bits that I used. So flexibility is potentially a really strong selling point for FOSS.

To summarize: Even if you hated KDE 3, and have been a long-time GNOME user, KDE 4 might be worth checking out. If you feel that OSX can fit all of your needs without modification, it is a really compelling package, and I was a really happy user of 10.3, and only stopped using it when my beige G3 box died and I couldn't resurrect it, because it was oddly "hacked" to be able to run 10.3 in the first place (which brings up an interesting point: it was fully capable of running 10.3 if I was willing to accept some performance tradeoffs, but Apple didn't think I should be able to do that, so would NOT have been installable without some hackish coaxing--if you go the Apple route you have to be willing to accept their control over some aspects of your experience, or be prepared to work around them in ways that they will not support at all, such as the whole jailbreaking iPhones fad). I think I may have had another point or two to make, but I think it's already turned into a large enough mess of a ramble as it is (really wish the comment box was bigger, but then it is supposed to be "comment" and not "mini-essay"), so, hope it helps in some way or another.
( 15 comments — Leave a comment )


color, uphair, smile
Jason D. Clinton

Latest Month

September 2011


Powered by LiveJournal.com
Designed by Tiffany Chow