I'd be curious to know if it actually is a bad thing to have... In the context of a rocket, there isn't exactly a lot of spare mass, spare volume, or engineers just sitting around and wallowing in boredom because the design is trivially simple and every niggling problem has been worked out.

If you skipped the launch escape system, you'd be able to transport more in the same number of launches(or the same amount in fewer) and your craft would be less complex, allowing you to focus on making the remaining systems less likely to need an escape.

Even if you don't fancy a look at our (honestly somewhat curious) level of risk aversion around space activity, it's not clear that adding an escape pod is a better investment, in terms of lives saved, than spending the resources on more extensive testing, improved reliability, and similar for the main systems. It's very much unlike the car scenario, where even 100% perfect engineering doesn't change the fact that other people are going to screw up and crash into you, and that a fair number of your drivers are going to be incompetent, drunk, or distracted; so you fairly quickly run out of improvements to the drive and steering system and have to achieve further survival gains by building in crash resistance. With a rocket, building a launch system that doesn't destroy itself and/or kill the passengers some of the time is quite challenging; but there isn't the same presence of ineradicable external danger, if your system doesn't kill the passengers, they'll survive.(Until you get them into orbit, where the micrometeorites can take them out...)

Is there any reason to suspect that the version of dosbox compiled into javascript is more vulnerable to having programs inside punch their way out than the version of dosbox compiled for your platform of choice is?

I doubt that it is god's gift to efficiency; but it isn't obvious why dosbox would be a more permeable VM in-browser than it is on the desktop(how permeable it is, or isn't, I don't know, it probably gets less testing than the more heavily used and often publicly exposed VMs; but it is also emulating something relativley simple compared to those).

Are there any? negative externalities that people have an obligation not to impose on others; or is it always the other guy's job to suck it up and add whatever systems are necessary to compensate for them?

The place where everyone goes to 'disrupt' things has assured us that business will continue as usual.

NTSB investigators reported the cause of the crash as 'Controlled reboot into terrain".

Man, if only we could afford to use 64 bit values for things. I realize that transistors are simply too expensive right now; but perhaps, in the future, the miracles of science will make this possible...

I certainly have no useful information to add to the speculation about cause; but that is what would worry me. Having to reboot a system every 284 days or less is a nuisance; but not a terribly big one(especially since the system is connected to a giant mass of moving parts governed by comparatively strict regulations concerning maintenance, so it probably gets taken to the shop fairly frequently anyway). However, if there is some value incrementing its way up that eventually causes the system to crash; I'd want to be very sure that there is absolutely nothing else that might modify that value in a way that causes it to grow faster than expected.

Cue the "If they'd chosen Windows, it would be impossible for this bug to occur" jokes...

Those have mostly been unfair since the NT-derived era; but, in the spirit of the joke, there was a bug in win95 and 98 that would cause the system to crash after 49.7 days of uptime. It remained undiscovered for years.

"Really short mission" technology, while it simplifies many problems quite considerably, is pretty limiting...

I don't control the purse strings in this case, unfortunately. Though, if you have any particular favorites(relatively small scale, rack-level stuff, for the most part), I'd be delighted to try to put a bug in the ear of anyone who does. Please, do tell.

Anyone with some legal experience able to clarify this? Given that grooveshark wasn't...exactly...apologetic about their strategy(nor has it changed all that much), my assumption is that the sudden shift to grovelling-apology-mode has much more to do with losing than it does with any change of heart.

Do courts give grovelling apologies enough weight that this 'contrition' is a logical strategy to try to reduce any awards of damages? Are such apologies sometimes added as conditions of a settlement, presumably so that the victor can grind the vanquished further into the dirt? Is there some other advantage to issuing one?

The specs allege "92% round-trip efficiency". I'm not sure how much weasel wording goes into that figure; but if it isn't a lie-for-all-but-legal-purposes, that could easily fall within the peak/off-peak price changes in many markets.

According to the spec sheet, the battery bank itself is a 400VDC unit and does not include an AC inverter(though a locally appropriate inverter, if not already available from a solar install or the like, would presumably be something you'd include in the project if you wished to install one). You are limited by the peak draw that the battery pack allows; but you can do absolutely whatever you feel like, subject to local codes and keeping your insurance, with the output. Unless you live in some really, really, freaky location where inverters that match your grid's flavor of power simply aren't on the market, it should be pretty trivial to spec an inverter suitable to the location.

Also, unless you are in an area without gas service, or in a house built with 100% electric heat, winter tends to be somewhat favorable to electrical demand. Lighting requirements go up, since the days are shorter; but contemporary lighting efficiency has increased by leaps and bounds of late(and anyone buying a Tesla battery pack is probably already running LED lights, or willing to consider it); and the cold reduces air conditioning demand to effectively zero, while heating is often accomplished substantially by on-site combustion heaters, which impose negligible electrical load(typically a trickle to keep the thermostats and other regulatory electronics up; but that's peanuts compared to AC load).

An area with proper winter would make outdoor installs problematic(Li-ion is better than most; but batteries still don't do any better for being below freezing); but otherwise the situation isn't so bad. If anything, places with brutal summers are probably the harder case, since AC is absolutely voracious in its demand, and almost always electrical, and the modest increase in available solar power doesn't necessarily negate the substantial increase in cooling requirements.

I don't know of any FBI-specific issues with DNA work; but various crime labs have had issues with atrociously sloppy practices that tend to go unchallenged, or overtly hidden, for some years. The big FBI story is definitely the "yeah, we basically didn't do a single hair analysis right for two decades; also hair analysis in general is probably bullshit" issue.

In general, DNA-based techniques have the advantage that they are actually 'science', as originally developed by scientists looking for useful research tools and facing some possibility of falsification, embarassing retractions, etc. It requires some skill, and considerable attention to good standards of cleanliness, bench technique, etc.(especially if PCR is involved; that technique is practically black magic it's so good at picking up otherwise impossible to detect DNA; but it is equally good at amplifying your accidental contamination of the sample by a few orders of magnitude)

Much of the rest of forensic 'science', is little better than polygraphs and phrenology. 'Bite mark analysis', in particular, is a tragicomedy.