Oh I remember those. They were fun. I wonder if the FAA would go after you for tailing a banner add behind them?

Now, now. That would make sense, which requires people in positions of authority to make rational deductions and judgement calls. Have you ever met a middle manager (a corporate "drone" if you will) in any decent-sized organization who feels both empowered to and is capable of taking responsibility and going out on a limb and making a judgement call on something potentially large and consequential? Now remind me again how big of a beaurocracy the federal government is?

It was if you didn't know any better.

In theory, practice in theory are the same. In practice, they aren't. To the point, real science is hard. Damn hard, and always has been. This isn't new. If this were new, we'd be living for a thousand years and taking vacations beyond the far side of the observable universe.

Even things that seem obvious can sometimes break down completely when put in the crucible. And things that you thought before broke down may really not have. No less of an intellectual powerhouse than Feynman famously said that you are the easiest person for yourself to fool. People don't understand this, not really, because it isn't in their daily experience. Even if they're trained scientists or engineers, they learn a pattern of behavior, they ape it, and they say they're doing science or research, but put their work to rude scrutiny, and you often find they've just produced incestuous no-op statements because that subtle bit in their heads that should have told them to scrutinize themselves and justify their own assumptions never flipped.

Nothing was wrong with good hunting grounds and plentiful wild fruits and grains on the east african grasslands either.

At some point it'll also add an atomicclockd to claim authority for all timekeeping.

Someone has to make ntp a part of systemd. It's already half way there with the ntpd name for the daemon.

Your points are valid, and make an existence proof for faster graphics than X11 had at the time the Qt4 team made their decision. You still haven't explained why it's not possible to have a graphics system that's faster than X11 and still has network capability the way X11 does. Yes, I know how 2D and 3D rendering works, and yes bandwidth counts so you can play video over a 56k connection for very fundamental reasons.

Well here's the bigger problem then. New applications are doing less server-side rendering. Now you might ask: why should the server bloat itself up on new different button styles and widgets etc, etc? It shouldn't, because then there's no end. And my answer is that you shouldn't have a system with unlimited button styles and glyphs and widgets, because that complexity and bloat will need to live somewhere, and will end up screwing everyone else over no matter where it lands. So let's ask the critical question: why on earth does anyone need to have all the bells and whistles to do their work? They don't. It's an arts and crafts project.

You can, but it's a callback function that returns back to Xlib code when it's done, and that's where the exit(-1) gets called. I flirted with the idea of manipulating the call stack to avoid that exit(-1) call, I flirted with the idea of passing back faults instead of calling exit, but both options looked way too error prone and would have taken me way more time than what I ended up doing, which is to split logic from display so that network errors wouldn't kill my core processing.

And my reply is always the same: if you make a change, improve the whole system; don't make compromises to core functionality for the sake of cosmetic improvments. Bells and whistles in the window manager are cosmetic. Being able to display output (from a single window, mostly text and line art graphs, not the whole damn desktop) to a different machine across the building is not cosmetic, it is why I use Linux instead of Windows. Maybe X11 doesn't let you do some things that really are better (I wouldn't know, the only annoyance I have ever found with X11 code is that its error handler calls exit(-1) if it panics without letting you deal with it in the calling program, meaning you have to split display and core logic into two processes instead of being able to keep everything in one address space), but network transparency (what else do you call being able to say DISPLAY=whatever:X.Y my_program) is a core capability that people do use, and you won't get them on board if you turn it into an afterthought.

Network transparency is not supported because noone uses it.

Well, I'm certainly not posting from the 24th century. So you've got me there.

Well, here's where you open yourself up. Distance measurements benefit from longer baselines. The biggest one we have now is about 2AU wide. Take a picture now, wait 6 months, take a picture again when the earth is on the opposite side of the sun. If we have a base on mars, we can have a slightly wider baseline with earth and mars on opposite sides of the sun for a simultaneous measurement (can't do that now at all) and two martian orbital radii for non-simulataneous measurements about a year apart.

Now you might say: well why can't we do this with a remote probe? And the answer is, of course we can, but it's damn near impossible to be flexible if you launch a satellite into space where you can't get at it to change instruments or reconfigure it. And big mirrors in space aren't anywhere near as easy as big mirrors on the ground. That's why we still build billion-dollar telescopes on the ground, even though we launch billion-dollar telescopes into space. If you had an actual base on mars, with support staff and manufacturing capacity of the sort needed to maintain a remote outpost, you can now have astronomical instrumentation that costs just as much as a space-based platform or a little more, but is worth dozens of space telescopes because it's manned and has the magical property of reconfigurability, repairability, and flexibility in software as well as hardware. So yes, you can do things with manned systems that you cannot do with unmanned systems, especially when you're doing science.

Well, here's a chance to make some money then, it takes about 50 minutes to set up a fresh install of Deb7 or the like, so train up a couple dozen guys to do this in under an hour, and charge $40-$60 to make housecalls to set up people's new PCs with Linux out of the box and sell them peace of mind for security against all the crap that gets in from outside and that's probably in the box to begin with. Just like Microsoft in the 90's, the secret is marketting, marketting, marketting.