A first look at Ubuntu desktop metrics
blog.ubuntu.com
blog.ubuntu.com
There is a bit of a chicken-and-egg problem here, though: Ubuntu (or rather, Linux in general) doesn't have a great reputation for consistent, high-quality HiDPI support, so there's less incentive for its users to buy HiDPI displays.
It’s the fractional scaling required for all the other resolutions in between which is troublesome, and I believe that applies to Macs too.
In this space, I think the only contender doing it really well is Microsoft Windows.
That said, I prefer Linux. For me having a few warts here and there is an perfectly acceptable price to pay for having a full computing environment I can fully trust and adapt to my needs.
Not if many apps don't have, and will not soon get, double-density assets.
(I'm a high dpi screen user)
Was responding to the argument that it works just as good as in MacOs though -- where companies were swift to change assets.
Not if you have a secondary monitor that doesn't need doubled, e.g. laptop built-in 4k with a 1920x1080 external. As far as I know this is an X problem.
Windows has been getting better but it's still not brilliant, even with resolution independent frameworks like WPF there are a ton of issues.
Ubuntu/Gnome was completely broken for me when I tried it last year. KDE was the best of the desktops but still not as good as MacOS.
So for example a MacBook Pro with a native resolution of 2880x1800 and native effective 2x of 1440x900 can also run at effective 1680x1050 or effective 1920x1200. This works on external displays as well.
... or if there were like more than 2 monitors on the market that were meant to be run at 200% scaling, i.e. 24" @ 4K and 27" @ 5K.
As an example, latest well supported desktop hardware that I have is a Haswell-based system from 5 years ago. A few of my more recent systems have too many problems.
I find it really convenient that I can run a CPU-intensive task using lots of storage on my desktop and no-one (especially me) notices any fan noise. A few months ago I took on a task to adapt a local data processing pipeline to run some analysis on the Hadoop cluster. Instead I ended up running it on the desktop over the weekend: total development time about an hour, EC2 cost $0.
The systems are from Dell, a previous model of the 5820: http://www.dell.com/developers (no affiliation, this is the Google search result link).
Or in other words: (K)Ubuntu users aren't all poor.
I was also surprised by the amount of people choosing automatic login. I thought there was a downside to that regarding the 'keychain' thing or has that been solved over the last decade or so?
Also
Why spend 600+ on a board that supports 2 processors when you can buy a uni-processor board that supports a cpu that can support 2 - 4 or 8 real cores, not including HT.
The Intel Mac Pro’s that followed had options for multiple Xeons, but the multi core nature of everything from the C2D forward put a major damper on the demand for multi CPU configs.
The single stat metrics look a lot like grafana single stat panels with sparklines and a light theme. The bar charts and the geographical view look a lot like kibana 5/6 though.
Either of these could point to an elasticsearch based storage? That would make sense, because linux installations with all of their annotated metrics are pretty event-based.