https://en.wikipedia.org/wiki/Power_to_gas has more information.
Thank you for the information, but I haven't been inspired to do any work to improve UI behavior over WAN connections. Years and years ago the worst UI behavior I saw was for a connection that was running over a 26.5 dial up connection. Strangely enough, some X apps were actually still usable. But, most weren't, and some took an hour to paint the first window. The worst ones seemed to be behaving like gimp did, causing a continuous stream of unnecessary synchronous calls to the X-Server.
I've been using X since it was X10, that includes all levels of programming from bare Xlib up. I've never written a server extension, but I have worked on the Matrox mga Linux kernel module just enough to make it work on IBM RS-6000 systems (you can grep for my last name in the kernel source if you care to check). I've also done a bit of messing about with the Doom3 sources to make it work better with Xinerama on a multi-screen setup.
By the way, I've still haven't seen a window layout object that works as well as the Motif Form Widget.
X-Windows can be doing either remote drawing command or sending blobs of pixels, it's all within the protocol. So, the only way to tell is to get something like xscope into the picture and see how the application is handling the data. One of the better ways to handle general window drawing (not pictures or video or other random pixels) is to send the drawing commands to the server to draw into an off-screen pixmap in the X-server, then have the X-server do a blit from the off-screen area to the viewable screen buffer.
So, yes, I can know its not falling back to some non-X compatibility layer.
Mostly these days I don't bother with xscope, because performance is pretty good. The last time I used xscope was when I thought Gnome was pretty neat, but wondered why remote performance was so poor. When I ran gimp through xscope I saw what seemed to be the toolkit asking the X-Server thousands of times how big its window was. Every one of those calls had to be synchronous and had to make a round trip to the X-Server and back. I never saw if they fixed that, but the proper way to deal with window sizes in X-Windows is to track the window size in local variables that are updated by the X-Windows event notifications the come every time the window changes.
Then there's the cube-squared law, for a given VOLUME, increasing the surface AREA by decreasing the depth, increases the heat transfer capability. So, thinner ice over a larger area can give off or absorb heat more quickly. If the average temp is below freezing, you'll observe more heat given off to the surroundings and more ice forming, if the average temp is above freezing you'll observer more heat absorbed from the surroundings and more ice melting.
QED is apparently mistranslated in this case.
What does your authoritative source have for the measurements of increasing Antarctic Ice mass? Especially in the interior of Antarctica (which is officially a desert climate)?
Most of the interior is a desert, no more than a few centimeters of ice crystal precipitation per year.
I would classify avoiding having glacial ice sheets covering large tracts of the northern hemisphere as a good thing, yes. As in almost everything, moderation is best.