A Week With GNOME As My Linux Desktop: What They Get Right & Wrong

[Ericg]

What the hell, Ericg! This is not what I would have expected from you. For one, I agree with Luke_Wolf: the tone was too confrontational. You can point out the flaws without resorting to writing in all caps and language like "confusing, shocking, and non-intuitive", "mentality of apathy", "fsck" and "piece of crap".

Hold on, Emerald. I will grant you that the last two may have stepped over a line. I will grant you that much. But the first two are valid perspectives. To call a design "Confusing", "Shocking", or "non-intuitive" is valid design criticism. In fact that is pretty much standard. You don't just say something is "bad" or "could be better." You have to be specific in your complaints. To call something that doesn't make sense as "confusing" is perfectly logical. To call something that surprises you "shocking", is completely normal. To call something that goes against established convention or usage patterns as "non-intuitive" is pretty much standard practice.

Just the same, referring to a perceived "mentality of apathy" is pretty direct criticism. It's not like the KDE developers can't respond to this, as Sho_ did, and argue one way or the other. If the design comes across as one of "Meh, we just threw a UI together" then that is apathy pure and simple.

It's not even that it's there's strong wording there, but it's mostly because the whole article was just that. It's a rant, because you accented the negatives. Not "it can be improved, look how nice this looks" but rather "this is terrible, fix it".

I accented the negatives because I was trying to avoid sugarcoating things. KDE gets certain things right, definitely. Hardware accel throughout the stack? My battery thanks the Qt and KDE developers. Sticking with standard "desktop" conventions and not jumping ship to the tablet-fad? I'm sure there are many of us out there that are blessing KDE for that. Making sure that things can stay configurable and are up to user's choice? Thank $Deity. But do you notice the pattern? All of those things are 'backend' stuff. I'm honestly having a really hard time right now to think of any KDE application that is just a knockout compared to its Gnome counterpart on usability and user friendliness other than: Dolphin, KTP, and Kdenlive (compared to Pitivi).

Amarok vs Rhythmbox? As far as intuitiveness Amarok definitely wins out there. Totem beats Dragon player though barely, Gnome Disks soundly beats KPartitionmanager, Gnome Software definitely beats Muon, its tied with Apper. KMail vs Evolution is probably a wash, though Evolution seems to have better support for Exchange and Outlook Web Services which is important for me in an enterprise/corporate/university environment.

But the worst part of the article, by far, is the headline. It's more misleading than Michael's article names. What it should have been named is "Visual issues with the KDE Applications 4 to 5 transition", or more aptly, "Overview of KDE VDG to-do list", because let's be honest, that's exactly what you covered here.

I didn't pick the headline. I gave Michael several possibilities, including a straight up "Gnome vs KDE" title. Plus I warned readers in the second paragraph that this would be a KDE heavy article.

Also, you should probably go to the VDG forums and read through the posts there by the VDG staff. And look at their tone. And learn from it. The VDG forums, last I visited them, were the most warm and welcoming place I've been to for a long while. Especially shocking that the VDG, by definition, gets to deal with a lot of controversy. But they manage it really well by actively keeping the atmosphere positive and stopping flames before they start roaring. So an article like this does a real disservice to their goals.

The least you could have done is point out that everyone is welcome (and highly wanted) to submit their mockups to the VDG, and raise any such issue to the VDG, and will definitely get them fixed sooner or later. They don't ignore people and it's a real collaborative process. It's as easy, if not easier, than submitting bug reports!

I'll cover the specific issues from my perspective in a separate post.

The VDG Forums are an amazing place, I love to see the back and forth that is happening there-- on individual applications. It's a lot easier to be nice and supportive when you're only looking at one application and just have to focus on that. I was looking at the KDE landscape across the board and seeing a pattern that wasn't going to get fixed with a "Hey guys, can you please stop doing this? Kay, thank you, heart!" I'm not saying this article wasn't controversial-- I fully expected it to be. All I'm arguing is that what was said needed to be said.

[Ericg]

Isn't that sorta like the pot calling the kettle black? I know the intention of this article was to point out annoying things people will come across during daily usage so that's not a bad thing per say. But if you want to critisize things that can't be fixed then Gnome is by far worse. A lot worse.

The emphasis of the article was on papercuts -- relatively small things that degrade the user experience, but CAN be fixed. "Rip out Gnome Shell" is not a small thing and we won't know what the future holds until Gnome 4.0 plans start getting made. Maybe we'll get another Tablet Experience, maybe we'll go back to the Desktop paradigm. Who knows.

edit: there is no valid way to measure desktop maketshare on linux. The only thing that anyone can do is trend watching. Trends obviously show gnome losing users fast. They probably have the lowest number of users since like 99 or 2000 or so. It -is- possible to say that is a side affect of their being more desktops to choose from. But the facts are most of them are gnome forks. You don't see KDE forks. There is no reason to do that. But every other month there is a new gnome fork. There has to be some valid reason why kde rarely gets forked, but gnome gets forked constantly. It's my opinion that people like gnome better, but hate the implementation.

Notice what does and doesn't get changed. Lots of people ripping out Gnome Shell, but keeping the applications themselves, or at least only modifying them to a much less degree.

[Ericg]

Meh... did the forum just eat my reply?

In short what I wrote was:
I don't believe in objectivity when it comes to design. You can't measure GNOME and Awesome after the same rules and no mater what group of users you'll ask there will always be those who think it's the best/worst experience ever.

Longer posts are being moderated. I don't know what the character cut off is, but if it was a big post then its in the queue to be reviewed before being posted.

[nwaynem]

This was not an unbiased look at where KDE has veered off course. Instead it was a write-up full of passion that felt like a slam against KDE. It was written hurriedly and in improper English. It was an extended rant like those you might find in IRC or some private board somewhere.

With that said, I do agree that Gnome is the more polished product. I try each new version of KDE only to find more documentation on what this does or that does is needed. Each piece of the desktop seems to work against the other. And while some settings might be in the system settings panel others are in the widgets. To my thinking this is very confusing.

Many think that Gnome is oversimplified and that is not true. Gnome builds a desktop that just works. Many complain that Gnome "forces" a workflow. I would ask what that is exactly? For most of my professional life my application windows are always full screen. Gnome works great in this way and gives you a number of different ways to go between these application windows. What more can you need? If anything, KDE could also learn from this "workflow".

All things being equal, the real projects with promise seem to be XFCE and Cinnamon. Both of which focus on bringing together a proven way of getting work done and packaging them into a nice looking experience. I would encourage any KDE user that wants more sensible settings, a fast user experience, and a more cohesive desktop to check out XFCE.

[mgraesslin]

Another KDE dev writing here.

While there are certainly valid points in the criticism, I found it a really hard and frustrating piece to read. This is mostly because of the over-agressive tone of the article. As an example I pick "Then someone got around to coding you and doing your actual UI, and I can only guess they were drunk while they did it."

Valid criticism can be stated in constructive, non insulting ways. Given that I'm rather disappointed by the piece of text and I honestly admit that I had a bad mood this afternoon after reading it due to the way how it is written.

[FLHerne]

As someone who's been using KDE and little else for years, and recently upgraded to Plasma 5, I thought this was a very good article.

Particularly the printer thing - I ran into that last week, and at about the fourth root-pw prompt spontaneously appearing I just thought "WHAT." and went to localhost:631 instead. In the Printers module, clicking any button - even ones that don't change settings - seems to have a 50/50 chance of asking for the root password. Even if you don't have a root password, or entered it on the last page. That sort of completely unexpected behaviour can quite legitimately be described as 'shocking'.

The KWallet thing reminded me a lot of Joel Spolsky's 'choices' essay. Having a lot of options and configurability is great - it's why I use KDE - but it's important to make a distinction between 'normal-user' settings and detailed tweaks that only a few people will ever want. A lot of KDE apps don't, although the new SystemSettings organisation is a massive improvement.

The first problem with this dialog is that it's distracting. You are trying to find help in the help file. You do not, at that particular moment, give a hoot whether the database is small, big, customized, or chocolate-covered.
----
Every time you provide an option, you're asking the user to make a decision.

On the other hand, I disagree with

KDE doesn't feel like it has a direction its moving in, it doesn't feel like a full experience. KDE feels like its a bunch of pieces that are moving in a bunch of different directions, that just happen to have a shared toolkit beneath them.

It's not great currently, but with the VDG guidelines every recent change to the UI has been moving toward a common experience. The KF4-KF5 split is quite stark and occasionally just broken, but most things that have been ported got a VDG makeover or are in the queue for it. The new System Settings really does make a distinction between normal-user settings and weird tweaks (although some individual KCMs are still a mess). It's clear where KDE is trying to go, it just hasn't got there yet.

[Ericg]

Another KDE dev writing here.

While there are certainly valid points in the criticism, I found it a really hard and frustrating piece to read. This is mostly because of the over-agressive tone of the article. As an example I pick "Then someone got around to coding you and doing your actual UI, and I can only guess they were drunk while they did it."

Valid criticism can be stated in constructive, non insulting ways. Given that I'm rather disappointed by the piece of text and I honestly admit that I had a bad mood this afternoon after reading it due to the way how it is written.

I'll admit that line likely crossed a line. I thought I had removed that line in particular but the edit may have been lost between versions (juggling three different versions across three computers). Apologies, Martin, for the bad mood, as I highly respect you as a developer-- your work on Kwin has been amazing and incredibly beneficial to KDE as a whole. The rough draft was more aggressive than this, as it was written in a fury in a single day. It was edited down a fair bit, but I admit that a few parts that remained in the published text could have been edited further. That being said, you and Sho_ both admit that there is valid criticism in the piece, and therefore it can't be entirely without merit.

[zanny]

Given that I'm rather disappointed by the piece of text and I honestly admit that I had a bad mood this afternoon after reading it due to the way how it is written.

I'm a contributor, I wouldn't say dev until I'm on the short list of a core projects contributors page, but I get the impression he wants KDE to succeed and is frustrated that great design from the VDG isn't being taken advantage of in implementation.

A lot of KDE apps like, say, filelight are just really old and aren't hot projects for development, so their UI rots a bit after a decade or more sitting around. It is so much worse to actually write code in KDE and realize this platform has magnificent power and capability under the hood but its often under UIs written in one of a half dozen KDE-esque GUI styles (qt widgets? xml forms? qml? pyqt? etc) that could really use an update but there are not enough of us with enough hours to drag the entirety of even KDE Applications (not to mention all the popular kde-apps.org programs that are rotting without maintenance) into a unified interface vision because the VDG makes great mockups but porting KDE to QML / Plasma with much more user friendly interfaces (especially to new users) as a whole is a colossal undertaking we don't have the manpower to bring about right now.

Being frustrated is perfectly valid. I don't see this article as an attack or hostile stab or anything like that, its a Linux fan who wants KDE to succeed and calls out legitimate criticisms we do need to try to address as best as we are able to make KDE a much more general purpose desktop. It is hard to kill Windows when you haven't even matched Gnome in some problem domains like consistent UI metaphors.

I'm working in my own fork repo trying to fix some of the KTP UI, but when I'm done with that I'll put it on my list to go figure out why Muon's UI is so much worse than the VDG mockup.

[anda_skoa]

but its often under UIs written in one of a half dozen KDE-esque GUI styles (qt widgets? xml forms? qml? pyqt? etc)

There are two UI technologies in use: QtWidgets and QtQuick.
To my knowledge there is no such thing as XML Forms and whether one uses Python or C++ to create UIs with QtWidget or QtQuick doesn't change the fact that there are two options.

Cheers,
_

[user82]

Gnome has made it to the future:
Clean, the video player plays video, the audio player plays audio

KDE is the origin:
Apps made by people who use them for people who know how to use them, exactly as they want to.

Both have their target audience, none is objectively worse or better.