I'm afraid that analyzing those disasters in terms of the _specific_ mechanical failures misses the point. It's possible to spend a project's entire budget, and go profoundly over budget to the point of complete failure, by trying to find and resolve each individual bug as it turns up. I'm afraid that the frequency of space shuttle failures was _amazingly_ low considering the flaws in the overall manufacturing and design process, and I do applaud the individual engineers and inspectors who did their best to keep those craft alive. It seems to have been a constant, hammering refrain of "we have to change this bit to work with that other bit which no one could have foreseen", or in the case of the O rings, "now we cannot launch in cold weather". It's extremely expensive, and demanding, to keep any project alive with that sort of segmented design from conflicting designers and manufacturers, funding turf wars, and scattered manufacture. From my personal systems and international work, it's a _nightmare_.

Pointing out that Dream Chaser would not have that particular failure is irrelevant to the flaws that create these kinds of failure modes of interlocking systems. These failure modes kept playing out for the basic Shuttle design in equipment failures, in launch delays, and in several cases, in the deaths of all astronauts aboard. So I'm afraid that focusing on preventing mechanical disasters like the O ring needs to go further upstream, to the political process.

> Not to mention WHY there is a war in the first place

Why _wouldn't_ there be one? Given the ties between religious wars, ethnic strife, historical conflicts, economic classes, and control of land, water, food, and engergy, conflict is inevitable for any large group. The scale and nature of the conflict is what will be in question. Why would there _not_ be some level of warfare at anty time?

Not at cruising altitude. However, any wifi could be used with Skype and other VOIP technologies, or email or web traffic. So the lack of cell towers is not such a problem. The lack of _bandwidth_ may be an issue. I'm sure, from experience, that any wifi on the planes will be heavily bandwidth limited and proxy limited to avoid carrying video, possibly even voice data.

Also, closer to the ground, cell phones work quite well. For example, they should work well when approaching the landing destination.

> Challenger failed because the failed o-ring between the segments of an SRB caused a jet of flame that impinged on the external tank.

Challenger and Columbia both failed for procedural reasons, not really the mechanical ones that in the end destroyed them.Richard Feynman exposed a lot of them in his biography where he discussed the investigation into the Challenger disaster. Management had one view about the reliability of the craft, and the engineers had another, which they were prevented from saying. That means the engineers are being told to lie, as a matter of course.

Another difficulty was the mandated complexity of the design. In order to accomodate contracts from so many states, it was being designed and built with different components by different teams in different Senatorial districts and being assembled, then shipped, extensive distances for the Shuttle itself and the separate boosters. This is _insane_ for complex designs that have to survive extremely stressful environments that are too expensive to test. I'd be extremely concerned about any software project that did the same thing.

Presidents are politicians. They must keep in contact with the voters to get re-elected, and the accessibility of the current president has been welcome. It helps defuse concerns about his level of education separating him from ordinary citizens, or forgetting what it's like to be black.

Assessing the strange "what if he'd been carrying a weapon of mass destruction" concerns:

                        1) The simplest pony yield atom bombs have to weigh at least 40 pounds for the nuclear material alone, based on rough guidelines for U-235 critical mass published in various magazines during my career. Jumping the fence and sprinting across the White House lawn, carrying something that heavy is difficult and _will_ give the Secret Service personnel more time. Such a device would be more effective _outside_ the White House during a semi-public event where the President is outdoors, such as an inauguration.
                      2) Chemical attacks have similar problems. An aerosol or chemical poison would have to basically flood the air of the White House, which has quite good climate control inside. That means getting past the ventilation system, which would be a _very_ good place to put the sensors and stop the air flow if there were such an attack.
                    3) Bacteriological weapons would, again, have to get from the attacker's entry to the President. Such a biological agent would be more effectively spread by leaking it during a White House tour, not by a run across the White House lawn.

The Secret Service reacted well, with measured restraint. Better staffed guard posts might be useful, but they are _expensive_. If you estimate the presence of 20 more patrolling guards, 24x7, at roughly $100,000/guard/shift covered, that's roughly $6,000,000/year. Which federal budget shall we strip for that funding?

It's often difficult to solve the "social engineering" problem if they're overwhelmed by immediate issues. Reducing the first problem to manageable levels helps establish confidence in your advice, and helps provide time for the retraining. And there's often a lot of "low hanging fruit", relatively simple changes that will help a great deal.

The "we can't fix the whole problem, so we'll just ignore you" approach is one that's all too common among frustrated support personnel. Their frustration is understandable, but doesn't actually help the people who pay their salaries. I've spent quite a lot of time with such personnel helping them get those "low hanging fruit" in place so that they can concentrate better on the remaining issues.

GMail supports POP3 and IMAP. IMAP can be vastly, vastly more efficient, especially because it can pre-filter the email into usable folders that are visible from multiple email clients, and only download mail when you want to read it.

I don't have one in hand. Can a Chromebook use a USB/modem device? I've had good success on Linux laptops with those, when people broke their modem jacks.

If it's Cuba, I suspect that the KGB also left behind a lot of monitoring assets. The military and espionage history of Cuba is fascinating material, and I'd bet it would be a fascinating junk yard visit to see what the Soviets left behind, even if they mostly withdrew in the early 1990's. With Cuba's shortage of international cash, and excellent location for monitoring US and Latin American radio and satellite communications, I'd suspect the KGB maintains some significant monitoring assets.

Personally, I'd _love_ to see how Cuban technical people see the open source and freeware software movements, and how they manage network operations on a shoe string, especially with the security issues they face.

> Once a local maximum is reached, further generations have no impact.

Except when they do, of course. Cross breeding can profoundly remix different 'local maxima' and even produce new breeds. Environmental changes, or changes in other species, can also interact profoundly with inherited characteristics, and some genes are even activated or de-activated by environment. And canine behavior isn't merely genetic, it's trained by the limited culture in dog families, and by the humans they interact with.

Dogs are likely to have a _different_ set of criteria for judging people. And people are under enormous personal and evolutionary pressure to lie effectively, espefcially to other people. So I'm unsurprised when a colleague's dog has a very different sense of whom they like when visiting the office. I have noticed that the dog's opinions about new hires seem to match my long term opinions, and it's not even my dog.

> There is no cure for cancer, heart disease, athsma, arthritis, or even the common cold.

Your definition of "cure" seems quite odd. "cancer" can include quite modest, benign lymphomas, easily and safely excised. Some types of heart disease, such as pericarditis, are very effectively cured with antibiotics. "asthma" can include environmental asthma, easily handled by getting _out_ of cities and moving to a beach can handle it quite effectively. Many cases of mild Type 2 diabetes are effectively cured by weight loss, exercise, and diet. Arthritis is harder, but the common cold is any of hunderds or thousands of minor infections, and almost all of those are cured by supportive care and time.

There's no "apply this medicine and it is gone" for most of them, but for some, well defined treatment can cure many cases.

Tied to a video display, which would take much more sophisticated development, it might be easier. It's a fascinating idea for people like Stephen Hawking, or as a fall back device for people whose more sophisticated tools may need repair.

I hope that this youngster talks to Lady Ada, over at http://www.adafruit.com/, about publishing a do-it-yourself kit for this.

I'll beg to differ: the real problem for day to day use is the freezing. It's been unusable for at least the past year.

The only reason that "anaconda's base supports all" is because anaconda, and kickstart tools, have the ability to support '%pre" scripts that allow manual use of hte command line partitioning tools to tune the partitioning as desired, and completely skip anaconda partition. Anaconda has never, and from all signes will never, be able to support all disk management and partitioning tools.

Since it's a Python based wrapper for the actual system tools used, features can be added. But there will be inevitable mismatches between configurations manageable through anaconda, and configurations manageable through command line tools for new disk and filesystem tools. And anaconda's use in system critical critical tools like kickstart mean that it _must_ be thoroughly tested before updates. This will slow feature addition in a way that gparted, or other tools, need not support.

> The sane choice of course would be for the facebook app to limit mobile data usage by culling data-heavy features as video autoplay

Don't you mean "The sane choice would be to drop FaceBook" ? The service, and its encouragement of careless video and image uploading, is extraordinarily and unnecessarily bandwidth heavy, especially with the constant pre-caching of both advertising and facebook poster content one has no _intention_ of ever actually selecting, but which winds up pre-cached because you opened someone's web page to read their post or check a social announcement.