SoylentNews is people
SoylentNews
The PC market may have stabilized and could see its recent declines reverse:
The PC market is forecast to return to growth next year according to Gartner, as buyers come to the end of their evaluation periods for Windows 10.
Worldwide PC shipments are expected to hit 267 million units in 2018, a 1.9 per cent increase on 2017, when shipments are forecast to reach 262 million. By 2019, shipments are pegged to hit 272 million units.
This year's PC sales are however expected to fall yet again for the sixth consecutive year, with shipments dropping three per cent when compared with 2016.
[...] Elsewhere, smartphone shipments will also continue to grow at a healthy rate, the market watcher claims. Shipments are expected to grow 5 per cent year on year to nearly 1.6 billion units in 2017. Gartner claims that the market is experiencing a shift away from low-cost "utility" phones, towards higher-priced "basic" and "premium" smartphone devices.
PC components are becoming slightly more expensive at the same time.
Also at EE Times.
(Score: 2) by JoeMerchant on Friday July 07 2017, @04:08PM (6 children)
Those are design decisions... if the designers choose to bloat with Chromium in a simple desktop app... that seems like a really good case for shared libraries and an integrated suite of desktop apps ala KDE.
Whatever happened to KDE? it had such potential for greatness.
🌻🌻 [google.com]
(Score: 2) by Grishnakh on Friday July 07 2017, @04:52PM (5 children)
that seems like a really good case for shared libraries and an integrated suite of desktop apps ala KDE.
Yeah, that's the whole idea of shared libraries and an integrated suite of desktop apps: consistency, sharing memory, etc.
Whatever happened to KDE? it had such potential for greatness.
AFAICT, three big things:
1) Thanks to GNOME stealing its fire, it just never got that much traction. There was a giant fuss raised about the licensing of the Qt libraries it relied on (leading to Miguel de Icaza and co. creating Gnome), and by the time they finally relicensed Qt to GPL it was too late as Gnome had too much mindshare. For some odd reason, most of the distros also jumped on the Gnome bandwagon, and continue to do so today as seen with Ubuntu switching to Gnome3. It doesn't help that Gtk+ is a miserable set of libraries for other developers to use as the Gnome team constantly changes APIs and deprecates things, plus it's all in C which is pretty lousy for writing desktop apps.
2) The KDE team lost focus and went after too many crazy ideas like with the Nepomuk and Strigi search/indexing systems which absolutely killed performance, and their "Activities" feature, all of which almost no one really cared about or used. So instead of making a stable, reliable DE and quashing bugs, they pursued "advanced" features of dubious usefulness.
3) The whole KDE4.0 debacle: they released KDE4 too early with tons of bugs and missing features, the distros that used KDE adopted it wholesale without having a fallback for KDE3, users got royally pissed because KDE4.0 was such a disaster to use and so broken, so tons of them abandoned KDE for other DEs (namely Gnome, though Gnome did a lot of the same stuff too).
It'd be really interesting if you could go back in time to 1997 or whenever, meet with the Qt people and show them all these events to come, convince them to license Qt as LGPL, and avoid Gnome even being created. Would Linux on the desktop be any farther ahead in 2017?
(Score: 2) by JoeMerchant on Friday July 07 2017, @07:30PM (4 children)
Qt has had so much unrealized potential, my big disappointment was that Nokia never built (real, mass marketed) phones with it.
It's not perfection, and maybe it's not even the best thing out there (depending on your perspective), but it was the first really good cross-platform toolset I ever saw. That was one of the amusing things about KDE being based on Qt: KDEWin.
🌻🌻 [google.com]
(Score: 2) by Grishnakh on Friday July 07 2017, @08:01PM (3 children)
AFAICT, it really is the best thing out there, for what it does. All the alternatives have giant negatives:
1. Gtk+ - C-based, controlled by Gnome devs, very unstable, generally low quality
2. .NET/C# - Microsoft-run, not FOSS, not really meant to be cross-platform in a meaningful way, slower performance than C++
2a. Mono - Not 100% compatible with #2, never really adopted by Linux distros that much
3. Java - Slow performance, doesn't really succeed in being "write once, run anywhere" as claimed, never looks like native applications
4. Boost - Doesn't provide cross-platform UI toolset, much narrower scope than Qt. Also rather ugly.
5. Python - Terrible performance compared to C++, really horrible at multi-threading which is pretty critical for UI code, library management is ridiculous
(Score: 2) by JoeMerchant on Friday July 07 2017, @09:35PM
Yep, I don't claim to know everything about everything, but your list mirrors my experience. The only time I've seen other toolsets decisively "win" a selection process is when the developers have previous experience with them and the scope of the project is limited to a specific OS. Or, I've also seen a case of "management loves Python, so we're doing it in Python because the last project got so moribund before it reached requirements coverage that we really learned a lot and will do better this time...." I actually saw a group go in for three tries with Python, each time making a bigger more tangled unextendable unmaintainable mess than the previous. That's something I've been pleasantly surprised with by Qt, even when you make implementation mistakes (a mainwindow.ui with 1000+ widgets in it, for instance), it's pretty easy to refactor and correct.
🌻🌻 [google.com]
(Score: 2) by Wootery on Wednesday July 12 2017, @02:09PM (1 child)
Modern Java and Java FX aren't too bad, are they?
About Boost: parts of it may be ugly/bloated [catnapgames.com]/whatever, but for some things it's great. It's not a complete multi-platform GUI development environment, but it's not intended to be.
(Score: 2) by Grishnakh on Wednesday July 12 2017, @02:52PM
Modern Java is still Java: verbose. No thanks. And it doesn't look native, at least not that I've seen. That makes it a deal-breaker.
Boost is useless: as you said, it's not a GUI development environment, so it's just as useful for creating GUI applications as any random non-GUI library. The whole subject here is what can you use to make GUI applications, which Qt is excellent for? Something which doesn't do GUI development doesn't compete with Qt.