Hacker News Reader

Comment by DCKing | Hacker News Reader

I'm just going to leave a comment to contrast the usual negativity.

What I like about this is:

- The developers are not attempting to reinvent the wheel.

- Material Design is well-regarded on mobile devices, and it's interesting to see what you can do with its ideas on the desktop.

- They clearly pick technologies that are the way forward on the Linux desktop: Wayland and Qt+QML. It's great that these things are maturing to the point that we can start leaving behind X and that it's clear that Qt has won the toolkit war.

- Focused approach of not (yet) supporting different distros, but (like Elementary OS) an attempt to focus on getting it to work well for one system configuration.

I hope these guys can put in the time and effort to make something out of this. It seems they are making a couple of good decisions already.

View on HN

Has Qt one the toolkit war? I thought most applications and desktop environments were still GTK based (though I can't comment how widespread the adoption of GTK3 has been).

I mean, I hope you're right because I really do prefer Qt over GTK (both as a developer and as a user). But my experiences as a user would also suggest that GTK is anything but an underdog.

Qt most definitely won the toolkit war. Multiple applications are switching to it from GTK.

There's also a large repository of software that isn't switching. "Won" implies the war is over; that Qt has earned a monopoly on Linux (or at least the de facto standard) but that's not the case here. At least not yet, maybe in a few years - though I can't see that either personally.

Maybe I'm just being pedantic, but "won" seems a little too optimistic. Particularly on a platform that's famed for it choice / fragmentation (depending on which side of the fence you sit hehe).

> "Won" implies the war is over

It also implies that there was actually a war, and that there is now not. Did gtk+ suddenly stop being developed because Qt won some imaginary war dreamed up by people who don't develop either toolkit?

Wow, I didn't expect people would read so much in my poetic language.

Qt used to be relegated to KDE on Unix but is now also used by Unity and LXDE. People prefer using Qt because it is much more cross platform friendly than GTK+ and has great developer tools. Qt seems to be the future of the Linux desktop at this point. So excusez-les-mots.

It really, really did not. GTK+ these days is superior to Qt in some critical areas, such as HiDPI support. Lots and lots of new applications and platforms are being written in GTK, consciously choosing it over Qt.

> such as HiDPI support.

That is by no means an accurate statement. GTK's HiDPI support involves an integer scaling factor for the whole desktop, and that's it.

Qt/KDE on the other hand allows you to simply specify the size and dpi of your display and everything will scale accurately, aided by the fact that KDE uses SVG icons.

You're conflating a desktop environment and a graphics toolkit. KDE <-> GTK+ ~= apples <-> oranges.

Futhermore, no, GTK's HiDPI support does not in fact force a scaling factor for the whole desktop. You can turn it on app-by-app. The fact that whatever desktop you're using doesn't utilize this capability isn't GTK+'s fault, it's the desktop environment's fault.

Now having said this, you could be entirely correct that for the time being KDE offers a better HiDPI experience than a GTK+-based desktop, but again, that has really nothing to do with the toolkits themselves.

Qt is not even standard C++ . I can't understand how people are double standard'ed and double dealing with problems . just look at what is done by Qt in C++ (whole non-standard Moc's) I am sure if it was from Microsoft (even open source product) whole Internet blaming Microsoft "wtf is this" & "Microsoft every time should disobey standards" and "Microsoft is just crap". how is this work? when we (as open source community) doing something this is good , but If Microsoft done something exactly like this then Microsoft is evil.

Sorry but that's an old, tired argument that only occasionally gets aired these days. Running a pre-processor as part of the build process does not mean that Qt is not standard C++. Also, if you knew anything about the history of Microsoft you'd know they deserve all the stick they get for their abuse of standards - 'embrace and extend' was their philosophy, it sounds innocuous but the effects weren't.

Another thing I forget to tell , Imagine one day whole linux desktop ported to Qt , what about C programmer ? what is Qt's answer for C programmer ? (please don't start C vs C++ , it is personal and technical choice , somebody like me prefer C's simple approach , somebody maybe like you prefer all OOP/C++ approach)

You could probably write linkable libraries to wrap around Qt's C++ bindings - much like how other programming languages do things.

They use a code generator to implement reflection methods on classes marked with the Q_OBJECT macro. The output of that code generator is standard, readable C++. It was, and still is, a very useful, practical way of doing things.

I can't say as to what Microsoft has done or not done, but I don't see anything immoral about the moc.

> Multiple applications are switching to it from GTK.

Which ones? Not disputing you, I don't keep up with Linux desktop these days.

Wireshark, Openshot, Unity, and every app in LXDE has / is switching to Qt from GTK.

Which is a drop in the ocean as far as the FOSS app ecosystem go, and to be honest after all this time I can't see either of GTK or Qt winning the 'toolkit war'.

If either of those would have the capacity to emerge as the de facto standard then it would have happened long time ago in my opinion.

Taking a quick look at the gui software (not a whole lot as I typically prefer the commandline) I have on my Linux desktop machine (running bspwm, not a desktop environment):

  Gimp (GTK)
  Krita (Qt)
  VLC (Qt)
  Inkscape (GTK)
  Chromium (GTK)
  Firefox (GTK)
  MyPaint (GTK)
  Handbrake (GTK)
  Galculator (GTK)
  Shotwell (GTK)
  VirtualBox (Qt)
  Brasero (GTK)
  Thunar (GTK)
  File Roller (GTK)
  Deadbeef (GTK)
  Blender (opengl)

Of course this is purely anecdotal, but this notion of some 'winner' in the FOSS toolkit space comes across as fanboyism more than anything based in reality.

I'm not citing Qt software I use, I'm citing GTK software that is now switching toolkits. How many apps are going from Qt to GTK? I could write a list of KDE equivalents to all the GTK programs you are citing - nobody is arguing that there isn't pretty much a program for every use case written against either toolkit. The question is what is software people are using, and what toolkit are people looking to write new software using? Conclusively it seems to be Qt nowadays.

It's irrelevant how many apps are switching, given that Qt have for so many years been a total non-starter for a huge segment of Linux desktop apps. As you say, the question is what software people are using, and you've given nothing real to support your contention that this "conclusively seems to be Qt nowadays".

Did you know OpenSuse still maintains a patchset for Firefox to use it with Qt? It’s amazing.

Subsurface, a program for divers, was another high-profile switch. High profile because Linus started it. If he and a bunch of other pretty smart kernel hackers can't figure out GTK, that says something.

https://www.youtube.com/watch?v=ON0A1dsQOV0

On a related note, GNU Radio has a foot in both Qt and WxWidgets and intends to focus on Qt going forward.

"Subsurface, a program for divers, was another high-profile switch"

I can't see how subsurface, which is an extremely niche application, can be considered as high-profile. And I am a scuba diver myself, so I have a bias going opposite what my opinion here is.

" If he and a bunch ... can't figure out GTK, that says something"

All it says is that they like Qt better. Linus told at a Debian meeting that he does not use Ubuntu or Debian because he tried once to install it and it did not work for his hardware, while some other distro worked right away. He never tried it again since. Does it mean that Debian is crap because if even Linus cannot figure it out "it says something"? No, of course not.

If you use homebrewon Mac. you'll find yourself using --with-qt more and more often. Because fuck X11

Material Design is well-regarded on mobile devices

It is? Wasn't it shipped just last week?

I've enjoyed it for almost a week on my tablet. I find it well-regardable.

On a more serious note, it is well-regarded by at least a large number of well-regarded designers and UI people. Material design has kinda swept the blogosphere in designer/UI circles. And, Google don't play when it comes to usable UIs; they put serious resources into figuring out how people interact with their software, so if they say, "This is good enough to be our UI standard for the next several years in every product we ship." then the odds are, it really is pretty good.

[Google] put serious resources into figuring out how people interact with their software, so if they say, "This is good enough to be our UI standard for the next several years in _every_ product we ship." then the odds are, it really is pretty good.

Not necessarily[1]. Especially in cases like this, where the org is large enough, the cash is endless enough, and the goal is fuzzy enough; that principal-agent problems are inevitably bound to prevail.

Think Pontiac Aztec. Or Lotus Notes. Or, for that matter, the all-new Google Maps(TM).

[1] http://brousseau.info/pdf/cours/grossman_hart_83.pdf

Not necessarily...but, the odds are good. I do find I'm very angry at Google Maps a lot of the time, though, so maybe Google is getting worse at interfaces and Material Design is going to go down in history as a massive failure.

People seem to be tolerating Windows 8 "flat" design, though, and that is pretty much a nightmare for usability as far as I'm able to discern.

In short, I don't now what's right and wrong, but I know what I like. I like my new tablet, and I dislike most other "flat" designs I've used in the build up to Material Design.

> People seem to be tolerating Windows 8 "flat" design, though, and that is pretty much a nightmare for usability as far as I'm able to discern.

FWIW, usability is being able to do something with as little thought or confusion as possible. So there is nothing inherently wrong with a flat UI - its the design choices made using it that matter.

Agreed. I wasn't complaining about the "flatness" of the thing. Just the chaotic nature of it (to me). Though, I suspect that with experience Win8 would become more usable to me. I just don't spend significant time in a Windows environment, so it feels really confusing whenever I find myself in front of one. "Usable" often means "What I'm used to."

I think there is some truth to the concept of material design, but so far in practice I've seen a wide variety of implementations that vary from "subtle yet pleasant" to "what the fuck were they thinking" and "why is this happening?".

I think the implementation is a lot harder than people realize. It's not Bootstrap, you can't just apply it to your site and be done. You have to think about the interaction design in context with the actions in your application.

Google Inbox seems to be the best example of good material design components, however, I've yet to see another app/site that uses these principles well. The "material design" frameworks I've seen on github thus far (even the Google Angular one) are sorely lacking.

Agreed. The frameworks provide a veneer, without the behaviors that are actually the important part of the UI/UX.

I'm still reading about it, and our UI is so far behind the curve that anytime I bring elements forward to new paradigms it is a positive gain. Switching to all Material Design would be a vast improvement...but so would switching to all Bootstrap (plus some new interaction guidelines), which is what I'm mostly working on lately. I'll probably experiment with Material Design, as well, but I don't have confidence I could do it well without a lot more research.

> it's interesting to see what you can do with its ideas on the desktop.

I am one of the negative voices but you're right here: it's certainly worth a try.

> Material Design is well-regarded on mobile devices

Says who? Isn't Lollypop all MD based? I've heard quite a bit of 'meh boring' reactions to it

I just got Lollipop on my 2nd generation Nexus 7 and overall it seems like a downgrade instead of an upgrade.

I don't have Android Lollipop and I am really curious , can you explain a little further?

The new keyboard looks awful.

There is some kind of network problem that causes 3 to 10 second latencies when I look up web pages at home with DSL but is not so bad at the gym with cable. I think DNS is involved.

It used to be that a left hand swipe and a right hand swipe would bring up different menus but now I have to swipe multiple times to count through the menus. It's that kind of hamhanded design that make people think bigger devices are unwieldy

On my first edition Nexus 7 it has slowed down everything considerably, it is borderline unbearable to use it for anything now, even just browsing web pages is an uncomfortable experience. The lagginess combined with the double-gestures required for unlocking and getting to settings is especially frustrating. I used to love this device, but now I feel like it's working against me.

While not your biggest issue by far: you can change the colour scheme of the Lollipop keyboard in its settings. "Material Dark" is a far more attractive setting (and I think the "Holo" settings will make it look like the older keyboard).

I have it on 2012 and 2013 N7 along with N4. Really pleased with it. My dad uses the 2012 N7 and he was wondering whether i got him a new device! And i really like the new keyboard. I think my mistakes have been reduced.

I primarily have heard reactions to how buggy the OS is. I haven't heard many complaints about the design concepts.

I tend not to criticize design decisions unless they are just plain asinine... Mostly if it is more of a matter of taste vs comfort level. I won't reject new designs just because they are uncomfortable.

There are only a handful of design changes in Android 5 that piss me off... I don't mind having email and gmail's apps combined. Two irksome things, is that the account switching circles are all but useless, I have 5 email apps, it only shows two bubbles, and there is no way to really tell which one is what. I can use the dropdown, but once you have more than two accounts, it should only show the dropdown picker. The second, is there doesn't seem to be a way to set an email as unread... I often will peak at a mail, and mark it unread to get back to. I know that inbox is working to get around things like this, but inbox really doesn't fit my workflow.

Aside from that, the keyboard changes don't bug me. Even the double swipe for settings really isn't that bad.

I really don't do much on my phone though... I removed facebook for shoving not one but two spyware apps down my throat, I removed most of the heavy battery offenders, and tend not to install apps that ask for permissions they have absolutely no need for... (flashlight apps that want sd and contact access, I get they need camera access for the flash)...

It will really be an interesting few years as phones approach today's desktop capabilities... I think remote displays will be a next thing... car interfaces especially..

So your main complaint is that it is not optimized for users with 5 different email apps? I have 3 on my iPhone, and I thought that is excessive.

No, my complaint is that with more than two email accounts, you can't tell one from another on the menu screen... there's no point in having them displayed that way at all as opposed to just keeping the drop down.

'meh boring' is a lot better than 'oh god help me'. A lot.

But 'meh boring' is not the same as well-regarded.

Isn't the point of a good UI to not get in the way? If that's the case, then whether or not it's been well-regarded might be best determined by the "meh, boring" factor, regardless of how much users/critics/etc are aware of that fact.

In other words, an absence of 'oh god help me' ought to be well-regarded if it isn't already. :)

Android 5.0 breaks the Nook app and Twidere outright, and damages F-Droid. It's also crazy-slow.

Avoid, for now.

Crazy slow on what? My Nexus 5 is as fast as always - possibly even slightly nippier.

The animations feel slower than they did before, but it's apparently a deliberate choice, not a performance issue. Scaling them at .5x in the Dev Options menu make the whole UI feel much snappier on my Nexus 5.

> to contrast the usual negativity

Thank you.

I can't dispute Qt becoming more popular for apps, and even DEs, but I still find that Qt apps AND all the Qt DEs on linux and BSD are horrifically ugly by default. That is harming adoption of Qt more than anything.

It's likely just bad taste and poor theming, but nevertheless, I find that 99% of the apps I use are GTK, and I'd always pick a GTK based DE (Gnome Shell, maybe mate) over KDE or Lumina.

Reply on news.ycombinator.com