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.
As a long-time resident of Minnesota, you have my abject apology for putting Michele Bachman in Congress. I'm not in her gerrymandered district, but I'm still embarrassed.
It seems to reflect Gnomes efforts to be very selective about their users.
We also had a paper tape based 4K 2Pass Algol compiler that worked, it waited until you reloaded the freshly punched tape of intermediate format to start the next pass and gave you an loadable paper table on the final pass.
Not bad for a machine that had 8 Opcodes.
Or, the another possibility is recording everything that needs to be cleaned up. That introduces more state data that has to be initialized, maintained in a coherent state and cleaned up.
Now suppose that you have to change the order of setup because of a newly discovered hardware restriction. You potentially have to re-order the labels for your gotos, or which cleanup gets done after each label.
If the cleanup is done at the bottom of every block that is executed because of success, it is automatically done in the proper time and place.
So, bailing out leaves dangling threads, leaking memory, and deadlocks that show up the next time the function is entered.
It's the easy way out only once.
The key fact is that