Linky Friday: Gods and Robots

Robotica:

artificial intelligence photo

Image by Belgapixel’s Linky Friday: Gods and Robots

[Ro1] Minister of AI? Interesting concept!

[Ro2] If we thought we were afraid of AI before, imagine how much we’ll fear them when China rules them.

[Ro3] Sand mounts look like boobs, apparently.

[Ro4] Well, this is a murder mystery waiting to happen.

[Ro5] In the future, you really won’t be able to believe your lyin’ eyes or ears. The fun is just getting started.

[Ro6] Guys, maybe it’s time to pack our things. We’re doomed.

Religion:

devil photo

Image by ChodHound Linky Friday: Gods and Robots

[Re1] What’s up with Mormons and science fiction?

[Re2] Kevin Williamson looks at the Church of Satan.

[Re3] Will the Catholic Church need to resort to ordaining married men?

[Re4] Ahhh, the predictable unpredictable plot twist.

[Re5] Remi Brague writes on the Christian roots of secularism.

[Re6] Matthew Walther explains why he believes in the Devil.

Energy:

solar power photo

Image by langalex Linky Friday: Gods and Robots

[En1] Ramez Naam on the steadiness and transmission of wind power. Naam is ever the optimist on these things (here’s a more cautious view). I am increasingly sold on the general cost of renewables, so storage and transmission are likely the two biggest challenges.

[En2] Amy Harder is skeptical of attempts to sue Big Oil like the cigarette companies. Exxon looks primed to go on the offensive against California.

[En3] Even forgetting the new terrifs, solar panels cost a lot more here than elsewhere. Is regulation to blame?

[En4] The New York Times looks at how climate change skeptics overperform in search engines.

[En5] I don’t always drink socialism, but when I do I drink mineral resource socialism. They can’t easily outsource where the oil is, and also it helps create support for more drilling. {More}

[En6] Never mind the SUV, eco-conscious wealthy people need to start applying social pressure to one another against taking private jets. It’s the holy grail of hypocrisy charges, not the least of which because it sticks.

Science:

beer photo

Image by Lord Jim Linky Friday: Gods and Robots

[Sc1] Good to know, I guess?

[Sc2] They need to work on their disaffected stare.

[Sc3] Even chickens like attractive people.

[Sc4] Andrew Gelman says cientists need to grow a thicker skin.

[Sc5] Chad Orzel argues that we need to make scientists public intellectuals.

Space:

beer photo

Image by phooleo Linky Friday: Gods and Robots

[Sp1] Images from the future of the Space Race?

[Sp2] Bad. Things. That’s what.

[Sp3] Beautiful.

[Sp4] Harrumph.

[Sp5] We must go to Mars.

[Sp6] I would say killjoys but I never want to make contact with aliens so I guess “whew.”

[Sp0]


Editor-in-Chief
Home Page Twitter Google+ Pinterest 

Will Truman is the Editor-in-Chief of Ordinary Times. He is also on Twitter. ...more →

Please do be so kind as to share this post.
Share

95 thoughts on “Linky Friday: Gods and Robots

  1. Ro1-Ro3: Most of the perceived threat to AI is seen in economic terms, a world with unemployment and no wealth redistribution. I’m wondering if the real threat would be to give better tools to authoritarian regimes and illiberal politicians in Western democracy. You can really use AI to enforce stuff.

    Ro4: Isn’t this the plot of the new West World kind of but with self-awareness on the part of the sex robots rather than outside interference?

    Ro6: The double standards that revolve around sex bots are going to be interesting. “When I was a sex bot, its because I’m a kinky cool edgy individual, when you use a sex bot its because your a loser who loses.” I wonder what the fault lines are going to be.

    Sp2: The best case scenario might be like when Columbus discovered the New World with the most technologically advanced side as the Europeans. This is assuming the first extra-terrestrial life we discover has at least human level intelligence. The first aliens we come across could be more like animals than us.

      Quote  Link

    Report

  2. Re3: I have been wondering for a long time how this is going to play out. The Catholic Church has had a priest shortage as long as I can remember, and it has been getting worse. Clerical celibacy is the obvious culprit. The old solution was to look the other way. The American Catholic Church was stocked with Irish priests, who brought their “housekeepers” with them. As a former Catholic priest once told my father, a Lutheran pastor, those guys generally didn’t cause any problems. But this sort of thing is out of vogue nowadays, and not really a selling point for young seminarians anyway. I have known several extremely devout Catholics who considered and rejected the priesthood because of the celibacy requirement.

    The stopgap in the US has been a very low clergy-to-laity ratio. The typical American Catholic parish is huge by Protestant standards. The senior priest has near dictatorial powers within the parish (e.g. no congregational meeting voting on the budget) but a bunch of stuff that the pastor would do in a Protestant church is done by deacons or lay employees. You need a priest for the eucharist and a few other sacramental functions, but most everything else can be done by someone else. So you end up with one priest in a large church.

    This only takes you so far. What then? One possible solution is a circuit riding priest. The parish is now run by a “pastoral associate” of some variety or other. But Catholics are very big on regular and frequent mass, and this still requires a priest, right? The leap is that while it does, he need not actually be present. He can consecrate the elements in large batches for later administration by others. Usually this is only done under extraordinary circumstances. So this solution is to normalize the procedure. The parish celebrates the mass as usual, but only sees the priest when it is their turn in his rounds. It would be quite a thrill, I am sure.

    The solution proposed in the linked piece is another approach. There has never been an actual doctrinal objection to married clergy, and there have long been extraordinary exceptions, such as married Anglican priests who crossed the Tiber. On the other hand there are strong cultural and economic obstacles. If you ordain a young man who is married and has or is likely to have kids, the economic calculus changes. Limiting the discussion to “older married men of proven virtue” is a way of minimizing both the cultural and economic obstacles. What makes it interesting is that it could be the camel’s nose in the tent. Once the precedent is set, the limitation may seem less and less defensible.

      Quote  Link

    Report

      • It’s partially true. The Catholic Church, especially in its Medieval version, is not anything like so unified or coherent as many imagine. The idea of clerical celibacy had been around for centuries, at least since Late Antiquity and with antecedents in Paul’s epistles. The essential idea was that sex is icky. The economic argument came later. And even after that, it took a while for the new policy to be consistently enforced.

          Quote  Link

        Report

        • I think the main thrust for celibacy in the Catholic Church was that too much church property was being passed down from father to son unofficially. There were also concerns of spotchy clergy work from priests who inherited and learned the post from their father. One of the things that led to the Great Schism was that the Eastern Orthodox Churches believe that the parish clergy who dealt with the lay followers should be married, long-haired, and bearded while the Catholic Church thought the ideal parish or monastic cleric has short tensured hair, a clean shaving face, and was celibate.

            Quote  Link

          Report

          • LeeEsq: I think the main thrust for celibacy in the Catholic Church was that too much church property was being passed down from father to son unofficially

            Correct, but 400 years later, it nonetheless didn’t stop Jeremy Irons from passing down the papacy to his son.

            (But that’s also how we got Lutherans a few years later)

              Quote  Link

            Report

  3. En3: That explains a lot. I wonder how much that analysis varies from place to place in the US? AFAIK such codes are not federal (although tariffs are, thanks Trump).

      Quote  Link

    Report

  4. Re5 – “Yet, Islam did not develop any secularist or egalitarian tendencies.”

    I don’t know much about this subject at all, but I know this isn’t true with regards to egalitarianism. Half the reason Islam was able to take root so quickly is that it offered a more equal vision of society for many than that provided by the near east Hellenized world, and especially seemed a better deal among the lower castes of the Hindu world around the Indus river valley.

      Quote  Link

    Report

    • One could argue that Islam is egalitarian as between male Muslims, but its not uniquely egalitarian unless compared with something like Hindu caste systems. An Islamic polity never had any problem adopting hierarchical forms of government when it reached a certain scale.

      What sometimes is being described is a romantic notion of pre-modern tribal societies, in which one belongs to a family or kin group that owes complete solidarity to fellow members, including the duty to correct each other’s errors. One cannot row a boat if one of the passengers insists on putting a hole in it.

        Quote  Link

      Report

    • As I’ve said many times, Columbus was wrong and every educated European of his time (all 10,627 of them) knew he was wrong.

      The size of the Earth was known since the III century BC; the distance from Western Europe to the East Coast of China was roughly known (at least in camel-days) since Marco Polo at the very latest; and the sailing autonomy of contemporary ships was clearly known too. It was impossible with the technology of his time to sail to China across the Atlantic, not because China wasn’t really at the other side of the sea (everyone knew it was indeed) but because you would die of hunger and thirst months before you arrived.

      Columbus spent three months to cross only about 1/4 of the distance from Western Europe to East Asia by the time he run into the Bahamas. The whole trip would have taken about a year.

      In tech bro parlance, Columbus success was a bug, not a feature :-)

        Quote  Link

      Report

      • I’ve often thought there could be an interesting alternative history novel based on the premise that the New World didn’t exist and Columbus and his crew starved to death somewhere near where Omaha would be today. How would the world have been different if you just couldn’t get there from here?

          Quote  Link

        Report

        • It would have taken quite a long time (perhaps until the invention of steam in early XIX century) for Europeans to attempt to cross the 20,000 miles (30,000 km) of sea they “knew” separated western Europe from China.

          However it is most likely the Portuguese would have drifted sooner rather than later to Eastern Brasil on their way around Africa (which might or might not have been what happened when Pedro Álvares Cabral first arrived to Brazil in 1500)

            Quote  Link

          Report

        • No North and South America? Phew, setting aside the whole climate being massively different the historic changes would be massive. Maybe a much earlier colonization focus on Africa? The Cape route would have become even more important until tech reached the point where the Paxlantic mega-ocean could be spanned.

            Quote  Link

          Report

        • In Columbus defense, he had sailed to Iceland, and had probably heard of Vinland, which he thought was Northerst Asia being reasonably west of Iceland. Hence him accepting the Ptolomeus estimate of the size of the Earth, rather than the correct, Eratosthenes’, one.

          Ptolomeus estimate, abou5 70% of Era5hostenes, would give you roughly 18,000 km from Spain to China. Columbus was still months away from his destination

            Quote  Link

          Report

          • Let’s see: Ptolemaeus thought the circumference was 18,000 miles, the coast of Spain is about 10 degrees W, Shanghai is about 120 degrees east, and both are about at 30 degree N. That would be a trip of 18,000 * (sqrt(3)/2) * 230/360 or almost 10,000 miles. Yup, much further than the 3,000 he actually traveled.

              Quote  Link

            Report

  5. To a certain person, tech bros are criminals so they might have a chance.

    I think that the first trip to Mars will be through a big massive government program than any tech pro private-capitalist one and the first people on Mars like the first people on space will have at least some military background. They will also be deliberately photogenic for any potential movie adaptation.

      Quote  Link

    Report

  6. What would a discussion about AI and robots be, without something about Amazon Go.

    So of course I keyed in on this-

    It uses a combo of machine learning and computer vision to keep track of what shoppers are buying.

    At the risk of being repetitious:
    So, aside from the initial development of software, what value is the Amazon management and shareholders adding here?
    They aren’t anticipating consumer preferences;
    They aren’t managing employees;
    They aren’t managing inventory, supply chains, or finances; these are done by outside vendors.

    So what “work” is being done here, that these detached people are deserving of any of the slice of the wealth created?
    And again, why couldn’t this be placed in the public domain, its bounty of wealth given to all?

      Quote  Link

    Report

    • Setting other morality considerations aside; from a utilitarian perspective at least perhaps we should let them invest their time, capital and talent into developing an truly extensive, ubiquitous and capable system of AI and automation before we arbitrarily appropriate it for the masses. Doing such a thing with Amazon Go strikes me as akin to privatizing the pharmaceutical industry after the first commercial antibiotic was developed.

      And that’s me thinking strictly on a utilitarian basis.

        Quote  Link

      Report

    • So, aside from the initial development of software

      You can’t yada yada yada the added value and then complain that there’s no added value.

      And again, why couldn’t this be placed in the public domain, its bounty of wealth given to all?

      This is a great idea that almost certainly would not have any effect on the incentives to invest in the development of other new technologies. If you want to live in a country whose government operates that way, I suggest Venezuela. I hear it’s working out pretty well for them.

        Quote  Link

      Report

      • Yes, those are very good points.

        So lets consider the development of Amazon software the way we consider lets say, the rights to a song.

        How long should the software be given protection, until it falls into the public domain?

        How much incentive is needed do we think, to both spur creativity, and also deliver its benefits to society?

        The answer is inevitably arbitrary, a negotiated decision made by society as a whole.

        Which is really what I’m driving at. Private property, especially” non-real” property only exists as a utilitarian device to further the happiness, freedom, and flourishing of all society.

        Amazon doesn’t have some naturally occurring and incontestable right to have exclusive and eternal possession of its software.

          Quote  Link

        Report

          • And with a stroke of the pen, we the citizens could make it one year if we felt like it.

            What stands in our way is this modern day Divine Right of Kings, where we accept passively that property is incontestable and eternal, where Prince Humperdink is entitled to all the wealth in Florin, because, well, shut up that’s why.

              Quote  Link

            Report

            • That assumes too much. Most people have, traditionally, not cared about patent or copyright term lengths, since it was a legal concept beyond their immediate needs. It’s only recently, with the advent of desktop publishing and home invention that the immediacy of such things has become a point of concern for larger and larger demographics.

              Your worm is turning, just slowly.

                Quote  Link

              Report

        • So lets consider the development of Amazon software the way we consider lets say, the rights to a song.

          One difference worth considering is that the song’s creator usually declares it finished at some point and releases it (puts out a recording, sells sheet music, something). The code that Amazon uses to implement its services is never finished — they’re always putting things in and taking them out. At any given point in time, some of the code in that pile is still within the protected-IP window. How often does Amazon have to take a snapshot of their running code base and put it away, so that it can be released a year, or five years, or ten years later? Are small companies excluded? The biggest personal project I’m working on has been in-progress for 4.5 years; would I have been required to issue snapshots periodically, just in case it might be useful to someone?

          Another question is scale and documentation. Way back in the day, when Bell Labs agreed to share early versions of Unix with research universities, a university got a tape with binaries and source code, a pile of technical memorandums that documented the use of all the programs (but not their internal structure), and a note from Dennis Ritchie that said, literally, “Here’s your Unix. Love and kisses, Dennis.” Berkeley had to put gods only know how many students to work taking it apart and figuring it out (and in some cases, reinventing because that was easier than figuring out the existing code). Having possession of 100M lines of source code is one thing — building a working system and understanding it enough to modify it is quite another.

            Quote  Link

          Report

        • What exactly do you think the status quo is? To be honest, I’m not sure myself to what extent the software running these stores is protected by patents, as opposed to just copyright. If it’s protected by patents, the patents will expire in 20 years. If it’s only protected by copyrights, then nobody can just copy the source code, but anybody can duplicate the functionality.

          Even if it is partially protected by patents, historically, software and even hardware patents have not been broad enough to prevent meaningful competition. Amazon’s retail site has competitors. Amazon’s cloud services have competitors. Microsoft’s products all have competitors. Intel has competitors. There are competing video game consoles, competing smartphones, etc. I’m deeply skeptical that the general idea of a grab-and-go store is going to be so well protected that Amazon will have an effective monopoly. The article you mention even says that other companies will be copying it.

          So what, specifically, do you find objectionable about the status quo?

            Quote  Link

          Report

          • That’s a wonderful question!

            What do I find so objectionable about the status quo?

            Or rather, what do about a hundred million of us find objectionable about the status quo?
            Because one of the most striking things about the 2016 election, is that both the 63 million Clinton voters and the 60 million Trump voters were angry, very angry.
            Dispirited, anxious, insecure, believing they are facing a dark and uncertain future.

            Why should we feel this way, in a world marked by technology that is producing such staggering amounts of wealth so effortlessly?

            Because even as consumer items grow ever cheaper, our lives are insecure- the big items in our lives like rent and health care are growing more expensive, farther out of reach than they were for our parents.

            So the wealth that is being produced isn’t actually producing a society marked by greater happiness and security, quite the opposite. The wealth that is being produced is accruing to a tiny and shrinking number of people.

            So I am looking for a way that the wealth that is extracted from the earth which belongs to everyone, and created by robots and machines and algorithms which belong to no one, can be delivered to its rightful owners.

              Quote  Link

            Report

            • Chip, I haven’t been reading this whole thread, so I may have missed some context, but I noticed this comment and had to reply to it. Consumer goods have gone up in quality and down in price, as you noted. But housing has gone up in quality considerably. Square footage has increased over the years, for example, and wiring has improved, air conditioning is more common, and asbestos and lead paint are things of the past. As I understand it, the percentage of income spent on housing has been fairly steady. And as for health care, yes, it’s more expensive, but people are able to have a quality of life they’d never dreamed of before. You can take a Crestor today where you may have needed open heart surgery ten years ago, and you would have been dead twenty years before that. I can only think of one product that’s gotten more expensive and worse – formal education. Then again, I can watch a lecture on physics or the history of origami with two clicks.

                Quote  Link

              Report

          • How does code/software differ from a food recipe? The latter, to the best of my knowledge, cannot be copyrighted or patented in any way. You can keep aspects of the process secret (I think all you are required to disclose are the ingredients themselves and even the requirement there may be conditional in some way).

            As a layperson, it seems to me that code/software’s closest analogue within our existing structure would be a food recipe, in part because of the reasons Michael Cain above outlines.

              Quote  Link

            Report

            • A recipe almost certainly doesn’t rise to the level of originality needed for a patent, although a list of chemicals and the process for combining them to produce a particular output may, even though it’s superficially no more complicated than the recipe (the phrase ‘a person skilled in the art’ does a lot of heavy lifting). On copyright — and I welcome correction from the actual lawyers here — somewhere between the simple list of ingredients on a scrap of paper and a finished page in a cookbook enough was added to become eligible for copyright.

              Today, all software source code is eligible for copyright. Years ago, advised by corporate lawyers, I acquired the habit of sticking a copyright notice into any piece of code of significant size that I wrote, on the off chance that it would be involved in legal proceedings at some point. (Never happened, but legal took images of my hard disk on a couple of occasions as part of discovery.) Chip is, as I read him, proposing two changes: (1) a much shorter duration for such copyright, perhaps shorter than the the useful life of the code; and (2) a mandatory obligation to distribute the code when the copyright expires.

              “Useful life” can be longer than many people think. I use, daily, a piece of code that I originally wrote more than 30 years ago. I use, regularly but not daily, a piece of code more than 40 years old. In the first case I’m almost certainly the only person in the world using it. In the second, where the code was open-sourced, there may be other people and I just don’t know about them. As part of my cartogram software, I use a piece of code at least 25 years old which is routinely used by hundreds/thousands of people and has been incorporated directly into a variety of current commercial products.

                Quote  Link

              Report

              • I’m not referring so much to legal structure as real world application.

                If I figure out how to make a BigMac, McD’s can’t stop me. They can’t even stop me from selling it (though I couldn’t use their verbiage).

                Why shouldn’t code work similarly?

                  Quote  Link

                Report

                • It does largely work like that. You can’t lift my copyrighted source code verbatim and claim it as your own, but you can decompile my compiled code, and if you are sufficiently clever, you can figure out the general structure and possible algorithms I used. If, however, I have an algorithm in my code that is sufficiently novel as to have patent protection, you may find yourself facing a law suit if you try to implement the algorithm in your code.

                    Quote  Link

                  Report

                  • Any such effort needs to be done very carefully. There are well established legal guidelines for how reverse engineering and clean room development has to work. The people who do the decompiling and study the structure have to write functional requirements, which are sent to developers, said developers never seeing the code being reverse engineered. Writing pure functional requirements, without including any implementation hints, is seriously hard. Anyone involved in the decompiling/study effort is long-term tarnished; they can’t work on the new implementation code for years.

                      Quote  Link

                    Report

                • Copyrighted code works exactly that way today. You can observe my service in great detail. You’re free to write your own code that implements exactly the same functionality. (The less said about the god-awful mess that software patents created the better.) What you can’t do is implement that functionality using my code. Linux was exactly the result of such an effort — an independent recreation of Unix functionality. Complete with — caution, geek gibberish ahead — lawsuits on esoteric subjects like whether Linus’s use of the names for constants in publicly visible C-language header files from Unix was copyright infringement.

                    Quote  Link

                  Report

                  • Thanks, Oscar and Michael.

                    It reminds me of all the food products that make it on to Shark Tank… never anything propietary so they only get an offer if they are sufficiently great, established, or marketable (or some combo therein).

                    But then some Bozo can angle for a huge deal by patenting adding a foil lid to single serve wine cups, something that’s existed in school cafeteria juice cups for decades. So go figure…

                      Quote  Link

                    Report

                  • On top of patents, DRM and rampant use of EULA contracts have also thrown significant monkey wrenches into the legalities of reverse engineering. There are now legal methods of preventing the old right to reverse engineer under fair use of copyright that didn’t exist a generation ago.

                      Quote  Link

                    Report

              • I’ve written here before about the LA City Bikeshare program-
                https://bikeshare.metro.net/
                Where the City runs a bike rental fleet, with bays of bikes stationed around downtown, and its fully automated. Software keep track of the inventory, tracks the fee, and bills your credit card, all with almost no human involvement.

                What is amazing to me is that if 20 years ago I suggested the City run a bike rental business, people would have (quite rightly) objected, on the grounds that running a business efficiently is not what cities are good at.
                Private entrepreneurs are superior at managing people, better at sensing and anticipating consumer need and allocating the business hours of operation to most efficiently match resources to need.

                Yet what I see now is that Google and Amazon and their collective suites of algorithms are performing much of this type of work.

                So with Amazon Go, which clever entrepreneur is sensing market needs for avocados versus eggplant? No one, the algorithm does that. Which bright energetic entrepreneur is expertly managing all the store clerks? No one, because there are none.

                What value is Jeff Bezos adding here? What work is being performed by Amazon managers?

                So how would this be different if it was the Los Angeles Municipal Grocery, run as a nonprofit public utility?

                  Quote  Link

                Report

                • What is stopping LA from doing just that? All they need to do is locate areas with sufficient public interest, acquire the real estate needed for a store, equip the real estate with the infrastructure and technology to operate the store, and assemble the necessary software into a usable system.

                  Or they could just sieze all of that from Amazon, and operate as a public utility, with all the attendant effects of such an act, as well as incurring all the work needed to maintain the stores and the software.

                    Quote  Link

                  Report

                • You see a working algorithm and ask, “What value does the corporation add once the algorithm is up and running?” I see a working algorithm and ask, “How large an organization, developing how many wrong algorithms, over how much time, did it take to get to the one that worked?” How do you propose getting people to take the (sometimes very large) risks associated with research and development?

                    Quote  Link

                  Report

                  • Not to mention that the core algorithms (the valuable IP is rarely a single algorithm) is typically useless without everything that goes around it, which, while probably not something that is high value IP, is still not trivial to build or maintain. And all of that ignores the regular work that will need to be done to the core IP as I mentioned above.

                      Quote  Link

                    Report

                    • So the value of lets say, Amazon Go or Uber or GrubHub or even Robotic Logging isn’t “Selling groceries”, “Taxi”, “Delivering food” or “extracting lumber”, but instead the value being added is in software research and development.

                      Which is kind of where my imagination is taking me. Machines have replaced physical labor, and are now replacing managerial expertise, leaving imagination and open ended research.

                      There just isn’t a lot of value in saying, “I can manage a large team of workers” when there isn’t such a thing, or saying, “I can track inventory” when software can do that better.

                      In the People’s Republic of Chipatopia, the actual doing of things like converting natural resources into finished consumer goods can be essentially automated and socialized and software R&D outsourced.

                        Quote  Link

                      Report

                      • For all of this, you have to think about the value added chain. Extracting raw materials from the earth, or from trash, adds value. Transporting those materials adds value. Shaping those materials into machinery parts adds value, as does assembling them into machines/robots. Designing machinery and writing software creates value out of thin air (as does writing books or screenplays). Maintaining machinery and software adds value, or at least extends the valuable lifetime of these things.

                        Every step of adding value that requires a human investment is something that needs compensation. Bringing all those dispersed pieces together into something useful also adds value, potentially quite a lot. Boeing designs aircraft, and then assembles the complex supply chain in order to build an airplane. Amazon does something similar. So does Uber, to an extent.

                        The question is not, are these companies creating value, the question is, are the people in the C-suite at these companies being overcompensated relative to the value they create? That is an argument you can make, but making that argument requires understanding what, exactly, those people are doing to add value, and can you quantify that added value sufficiently to be able to argue that it’s not comparable to the compensation offered?

                        It’s not about software.

                          Quote  Link

                        Report

                    • Absolutely.

                      Governments come as close as any organization to treating software as if it’s a static thing. Which is how we get into the situation we’re in with air traffic control. Or one of my favorites, nuclear reactor control software. The certification process is so onerous that we have reactor operators scouring the country for replacement boards for 40-year-old mini computers, and looking for people who can code in assembler for those machines, because certifying a new computer and a new software suite would be so expensive.

                        Quote  Link

                      Report

                          • Boeing still (as of 8 years ago) has 20+ year old machines that are kept alive because those machines are needed to run engineering software (that was written for that hardware) that is needed just in case an older airframe like the 767 needs a change, because that was the software the plane was certified under, and upgrading the software would require re-certifying the entire line.

                              Quote  Link

                            Report

                            • Would running the programs on a software emulation of the old machines be okay? My understanding of the nuke plant problem is that the plant certification included specific hardware. In the DoD cases, at least a couple of the systems involve custom unibus boards for controlling peripherals.

                              There’s a scary thought: “This system depends on hand-wire-wrapped boards that are almost 40 years old.” Granted, here in the closet I have a hand-wrapped board that’s 30 years old, but I don’t put power on it and expect it to work.

                                Quote  Link

                              Report

                              • IIRC, you would have to emulate the math coprocessor, or whatever the equivalent was on that hardware, since the analysis can shift a bit if the processor instructions set is different. We have this issue even today, where the same simulation would have two slightly different results on Intel versus AMD processors. Not a big deal when you are just doing engineering, but if it’s for regulatory certification, they usually want things to be the same.

                                  Quote  Link

                                Report

                                • Such exist, at least for some processors. My understanding is that the best of the PDP-11 software emulators, as an example, has been tested against an enormous range of software and is considered bit-perfect, including floating point. I would think that at some point someone at Boeing would have run test cases on an emulator and compared the results.

                                    Quote  Link

                                  Report

                                    • I’m old, and remember too many power supply capacitors failing catastrophically and taking the motherboard out with them. I’d at least pay to know if the emulator produced an obviously wrong result — that’s a pretty trivial expense.

                                      Isn’t one of the corollaries to Murphy’s Law that the $500 chip will, at the worst possible time, burn out in order to protect the $0.50 fuse?

                                        Quote  Link

                                      Report

                                      • If Boeing’s IT hardware acquisition practices 20+ years ago resembled what they were when I worked there, Boeing had a pallet or two of the machines in question sitting unopened in a warehouse somewhere. Chances were pretty good they could cobble together a functioning machine or three if needs be.

                                          Quote  Link

                                        Report

                            • Our software is one used in flight certification, which means we have to be able to load old problems (generated as far back as 20 years ago) and be able to document why the answer is different.

                              People get shirty, especially if they see a drop in lifespan. (Although admittedly, you generally get a longer lifespan under newer models. As with anyone doing serious engineering work, everything erred on the side of conservatism, if you weren’t 100% sure or having to do estimates, fudge factors, etc).

                                Quote  Link

                              Report

  7. Sc5 [thin-skinned scientists]: That wasn’t the article I was expecting it to be,* but it was interesting nonetheless. I see something like that dynamic in other professions, like history. Once in a while, a journal will publish a “forum” on some author’s book. The format usually has four parts:

    1. A lengthy review/critique of the novel.
    2. A middle-length review/critique.
    3. A very short review/critique.
    4. A medium-length response from the book author in which the author acknowledges some good points, but in the (admittedly small number of) forum issues I’ve read, the author never really reconsiders anything substantive.

    The Sc5 article also reminded me of how I respond to critical comments to a post I’ve written. I put a lot of time into most posts and write several drafts, revising repeatedly. I get testy when people bring up the “what about x’s,” etc., even though those critiques usually have merit and are not often even meant as a criticism.

    Of course, the Sc5 traces scientists’ reactions to peer review, which I assume operates differently for the historian-author in the journal forums. And of course, at OT, there’s not really peer review. Or rather, the comments section is the peer review, at least that’s how I see the comments at least sometimes.

    /two cents

    *Not a wrong link, but just different from what I thought it would be.

      Quote  Link

    Report

Leave a Reply

Your email address will not be published. Required fields are marked *