Mark Shuttleworth on “GNOME vs Canonical vs KDE” | Hacker News Reader

Mark Shuttleworth on “GNOME vs Canonical vs KDE”

markshuttleworth.com

53 pointsrbanffy30 comments

To be honest, it's a little bit embarassing to see such a... well, how would you call it? Flamewar, maybe.

I've been active in a irrelevant project of GNOME (ie. marketing), and my initial opininon still seems to hold: It's not about the desktop; it's about the apps. Neither GNOME's Shell nor Canonical's Unity solves any interesting problems for users.

And the lack of apps is basically just a matter of a missing business models for Open Source Desktop Applications.

To see them fighting over stuff that probably makes up -- at most -- 0.01 percent of the users' time, let's me wonder whether Open Source is ever going to be any major force on the desktop. I've once believed it would, but this belief dwindles.

Not sure it's a flamewar (the flames are basically going on one direction!), but let's go with it... I'm participating in that flamewar, but agree with you 100%. That's part of where the conclusion of my series goes. :-)

Uh,

How is a debate a flame war in circumstances where name-call is not happening? What we have here is a heated debate. I might accept that Shuttleworth is totally wrong for the sake of argument. Even, I can't see ad hominem attacks, innuendo, comparisons to Hitler or any of what I'd see as the characteristics of the flame wars I know about.

The idea that the app Launcher doesn't matter seems implausible to me - first impressions are very important for example. But maybe you're right and it doesn't matter. We still have to argue what is or isn't important. I don't see anything embarrassing about that.

I'm writing a series of posts about the issue, largely as a result of extraordinary comments (attacking GNOME and individual participants) made by Mark over the last week. It is an attempt to correct the record and set a path forward.

http://bethesignal.org/blog/2011/03/12/relationship-between-...

Post it to HN and link the HN link here so we can find and upvote it.

Will probably do that when it's more complete. :-)

So is your argument that canonical developed their stuff in private and only wanted it approved once it was finished whereas the appropriate approach would have been to develop it in public (or within the Gnome process) so you could see what they were doing, comment and tweak it so it fit the gnome vision?

They did that, but they also didn't even try to upstream a whole bunch of stuff. Then in 2010, after not laying any groundwork, they were annoyed that the very first thing they proposed which was (even only partly) related to their UX work was rejected. Then Mark accuses GNOME or people within GNOME of playing politics.

Well, how could you imagine they might make up for this now?

This is something of a dilemma in any team-based development. Everyone thinks they have the right to go off and do their own thing. And everyone does have this right.

But once someone has created something, how do they work that into the "mainstream"?

Unity? I don't think they want to upstream it, let alone upstream wanting to take it... and that's not such a bad thing. All kinds of downstream users of GNOME have made totally different environments with the platform.

But what about stuff they want to upstream? I just hope they learn from this experience and do it better in future, like others have learned before them -- like Red Hat and Novell in GNOME, lots of Linux kernel contributors...

That's the technical side of things. The social side of it all is much harder for them.

To try to steer the discussion to more constructive tracks, I wrote some guidelines on making cross-project collaboration work:

http://bergie.iki.fi/blog/on_cross-project_collaboration/

This is based on working several years on getting both GNOME and KDE to adopt some common ideas and libraries.

OP (by Shuttleworth) is very hard for someone not intimately familiar with the history of who said what and who did what to follow.

True, but the overall impression I got was that there is a design by a committee in the land of Gnome and MS is having hard time reconciling that with more structured development practices.

There are 2-3 designers working on Gnome Shell, but describing it as 'design by committee' is neither fair nor accurate.

It's also worth noting that Gnome and Canonical have completely different and separate collaboration systems, Gnome uses their mailing lists and git, Canonical uses Launchpad and bazaar.

Between the technical barriers, development priorities, and needless missteps (Canonical requires copyright assignment for /all projects under their umbrella/), the two communities are pretty well alienated.

That well may be, but it would be a real loss for the Linux desktop if the two groups can't figure out how to work together. Between Ubuntu, Fedora and openSUSE, I suspect GNOME provides the UI foundation for most Linux desktops currently in use. Ubuntu is the most prominent Linux distribution. It will be very unfortunate if they can't reconcile.

The bickering in all of these blog posts makes me sad. I had hoped at least some of the project leaders could rise above the frivolous nit-picking. When will the conversation move on to constructive topics like "where to go from here"?

Projects inherently involve disagreements.

This post seems to be written in the most diplomatic language possible.

I've only looked at and used a few of the standards of Freedesktop.org but they certainly seemed problematic.

Disagreements are fine. What is not fine is the tit-for-tat he-said-she-said bickering. It is not professional and is counter productive.

This post seems to be written in the most diplomatic language possible.

You should see his previous post. :-)

Everyone wants to forgive and forget, but I don't think human beings in general are capable of forgiveness unless there is acceptance of what went wrong and some sense of resolution, so trust can be rebuilt.

The issue of "what went wrong" is under dispute. Dealing with that is the first step to "where to go from here".

From my outsider perspective "what went wrong" is clearly just a case of poor communication by everyone involved.

From what Shuttleworth is replying to: Mark argues that GNOME should be a place where we have internal competition. But his idea of internal competition seems to be competition between different end-user experiences.

So is the idea then that Gnome is oriented toward imposing a single user experience? Like Twitter and Apple and so-forth????

I deeply disagree with this. A good portion of open source user experiences indeed suck but competition seems like a far superior approach to creating a good user experience than accepting a single committee of expert.

What does that even mean? How could GNOME do what Apple does if all the code is open source? Anyone can come, change whatever they want and ship it as their own. This is what Ubuntu and others have been doing for years. The kind of competition you consider "ideal" is already happening, and has happened for years. The only problem is that some people seem to expect that because you hack on free software you should not be allowed to have strong opinions about the UX of the code you are doing, and must submit to the will of all users and drive-by designers on blogs (which routinely ask for an innumerable set of mutually contradicting features, threatening to leave the project if they are not listened to).

What I consider ideal is a multi-layered structure where a "graphic shell" (Gnome-shell or Unity) sits on top of "agnostic" windows manager code.

The current version of Gnome actually seems "good enough" to do this. My main concern is that it stay that way. To consider an opposite pole, KDE is certainly open source too. But, for example, to use the KDE code to parse the gnome/kde panels start menu you need to import the entire KDE virtual file system. IE, KDE has a monolithic structure in that regard. Uh, it's open and you could rip out this piece of the code but it would be a mess. Gnome isn't that bad... yet.

There's a tension between making software fast, reliable and attractive and making it completely configurable. The current culture in GNOME, as far as I understand it, is that the primary goal is to provide a great experience for your users, even if extreme configurability needs to suffer in some cases. Others can disagree and try to do different things, but if you want GNOME to agree with you the only route you have is to join the project and change things from within.

As I hope the post you replied to shows, I'm looking for configurability at the programmatic level, not at the end, end user level. Are you saying that is going to suffer too?

Considering Canonical is currently committed to bolting Unity on top of some version of Gnome, that version is going to have to be configurable enough for such purposes. Hopefully that will be the main line of Gnome.

Incidentally, I have recently (re)discovered fluxbox. It's a breath of fresh air compared to how bloated both Gnome and KDE have become.

Me too, but replace Fluxbox with XFCE. XFCE 4.8 is actually quite usable and pretty. I'm impressed by the developers' attention to details, like the fact that you can alter settings of a panel no matter whether you right-click on the panel itself or on a panel widget, or the fact that the taskbar correctly rotates 90 degrees when the panel is docked to the left or right side.

That is fine for you. But not for the average user out there.

With this and this [http://news.ycombinator.com/item?id=2319375] it seems like today is nerd thunderdome day.

Reply on news.ycombinator.com