Live from Prague!

Live from Prague! Dave Taht, he who sometimes fixes the Internet from my basement (last week was the most recent occasion) asked me to post this “because G+ is dead”:

It GPLs me


    1. Then there’s “enjoying the sun while being old enough not to give a fuck who sees you half-naked or what they think” which is where Dave is at.

      1. >Then there’s “enjoying the sun while being old enough not to give a fuck who sees you half-naked or what they think” which is where Dave is at.

        That’s about right. Dave…really has nothing left to prove to anybody. Not after fixing bufferbloat, then WiFi, and now congestion notification.

          1. I have a cousin who has gotten rich suing people over accidents.

            He is now SALIVATING over the legalized pot because of the product liability lawsuits he thinks will make him a billionaire.

            If he didn’t have children I would shoot him and dump his body in the Atlantic.

            1. Lemme outline the case that you should instead be grateful to the man.

              Fifteen, twenty years back, none of the legalization advocates seemed to have a mental model of legalization that incorporated product liability. I can’t prove it, but am sure that my own mental models incorporated product liability, coupled to certain other factors, at least as far back as ten years ago. I want to say fifteen, and regret that I probably shouldn’t say twenty.

              Information publically distributed about the effects of legalization is effectively under a state of systemic censorship. Media censorship can be shown to exist for other topics, many media people seem both to be pot legalization sympathetic, and independently of the first trait, mostly not smart when it comes to complex mental modeling. The independent information source is rumors filtered through internet analysts, and it isn’t hard to understand how those systems could by accident have blindspots when it comes to this or that important topic.

              If information about pot legalization has become abnormally narrowed in the public eye, the differences between expectations and reality will have consequences for the public reaction. Believable truth excites people less than the rumor that spawns when established sources of truth are discredited. (If the Obama administration had been willing to accept the immediate political cost of transparency on border related crimes, it is likely that the political faction would have paid a lower cost in the opposition that has since spawned.)

              If your friend does make a lot of money from product liability after legalization, that is information about the importance of product liability in the system. His fame is a way of conveying that information to the public. That path of closing the difference between expectation and reality should be for you by far preferable to some of the other paths that I can imagine. You are the sort who would object to mass graves as an outcome, aren’t you?

              1. Let’s start with the most important point:

                that is information about the importance of product liability in the system

                No, it isn’t. There is very, very little resemblance between the results of a jury trial involving science and the real world. There are SEVERAL products that have generated (aggregated) billions of dollars in settlements on science that was so thin you couldn’t repel a dandelion down a cliff. You get a plaintiff that is sufficiently sympathy inducing and a jury that isn’t really so certain that the earth is older than 6000 years and it becomes a contest of lawyers, not science.

                Next, he’s not my friend. He’s a sociopathic leech. He’s convinced that global warming is real but he’s building a 6 thousand square foot house OUT OF CONCRETE in a hurricane zone. Oh, and he’s putting in a room with a separate air handling system because apparently people who buy those kinds of houses want a place they can smoke without polluting the rest of the house.

                He’s also looking for *suitable* plaintiffs to go after the alcohol industry (he informed me WITH A DRINK IN HIS HAND) because it’s “proven” that alcohol can cause throat cancer AND THERE IS NO WARNING LABEL.

                Prior to the “medical” legalization there was scant evidence for the problems with pot because it was almost impossible to do studies on it. As usual the absence of evidence was taken as evidence of absence, and anecdotal stories made for good excuses. Now that evidence is starting to stack up, it’s not “effectively under a state of systemic censorship.” It is simply that the evidence is counter to what we “know”. It takes time for *real* evidence to pile up. You do want to base things on *real* evidence, don’t you?

                There have been articles on the problems stemming from pot legalization everywhere from PJ Media to the Wall Street Journal, so people are starting to notice.

                Another problem with those presenting the evidence is that they come across as moral scolds. They are (generally) the sort of people who also caterwaul about modern promiscuity, pornography and general godlessness. They also tend to be fairly hyperbolic about it.

                This puts WAY too much noise in any signal they are putting out.

                Yes, now that marijuana usage is more in the open we’re finding more an more problems with it. It looks habitual use of pot is as bad as habitual use of alcohol. It’s also looking like it’s about as bad for occasional use as, well, alcohol.

                No one under the age of about 21* should use products containing THC. But then they shouldn’t drink, or take *any* psychoactive drugs. But they have been. For a long time. And yes, legalization makes that easier on several fronts.

                But guess what? Liberty is ALL ABOUT being able to make stupid fucking choices. If you don’t have the ability to make stupid fucking choices YOU ARE NOT FREE.

                I would like there to be (at least) one place in this world where we can be free. One place that geographically and climatelogically doesn’t suck out loud (for instance Alaska).

                You’re not going to convince pot smokers of that pot is bad for them. At least partially because habitual pot use means you’re spending a lot of your life more stupid than you started out. At least partially because people don’t believe what they don’t want to believe.

                But at least partially because the state of science is SO FUCKED UP right now:

                And because science reporting across the board is so fucked up that you can literally get two articles IN THE SAME PAPER that contradict each other.

                * Brain development is mostly settled by somewhere between sometime between 21 and 25, and that’s when it’s less dangerous to go stomping on your neurons with hobnailed boots. Oddly enough it’s also the time when most people start to slow down on their drug and alcohol intake.

                1. >Liberty is ALL ABOUT being able to make stupid fucking choices. If you don’t have the ability to make stupid fucking choices YOU ARE NOT FREE.

                  But of course it also goes hand in hand with taking responsibility for the consequences of stupid choices: the consequences should as much as possible not affect anyone else, not even indirectly (spending tax money to fix them up etc.) and should only affect them. And there are two aspects of it. The first is the kind that is politically solvable in theory albeit will never happen in practice: legalization / liberty increase should go hand in hand with other, similarly libertarian changes of people not paying for other people’s mistakes and so on. I know you would support that. Every libertarian would. What I am trying to say is that it is more important to protect normal people than people with unusual desires, thus, it makes sense that I would want to see some very serious guarantees that that responsibility-taking will happen first, then legalization / liberty could come second. This is a known and well-working model: for example, at most jobs, one will only be allowed to be work unsupervised once one has demonstrated responsibility.

                  But there are other consequences for others that cannot be legislated away. I mean, think about some drunk or high 17 years old car driver causing an accident killing your child… here the libertarian logic would be to punish drunk / high driving or even just punish causing an accident very drastically and then people will think and make choices.

                  Yes. Works for high-IQ people. Doesn’t work for low-IQ people. Who are children in adult bodies. Who cannot be trusted with liberty. Not because they would harm themselves, who cares, but they would end up behaving in ways that harms others.

                  Now there could be many potential solutions to deal with this, but many of them would require radical social changes. But there is one that is fairly simple. It was accidentally discovered by Irving Kristol when he was studying Victorian Britain. He was surprised to find that those prudes had legal porn theaters, live-action. The only catch was that one could not just buy a ticket, one had to buy a yearly membership. So everybody had to decide if they really want it or not. Had to make a rational, thoughtful decision, not just acting on an urge.

                  So that would work and would be simple. Since more or less everybody who wants to legalize pot would also tax it, because “sin taxes” actually make sense, as they help dealing with the cost of negative externalities, instead of a per-unit tax, just sell one-year permits to use pot for $1000 per year to be paid in advance. It would filter out the people who just act on random urges, and people who are able to think and plan long term would be the ones who would buy it. People who are dumb and think 15 minutes in advance do not buy things like this. People who think a year advance and say yes I want this, would buy it, and those are precisely the people who are able to function like an adult and, say, have things like friends as designated drivers or cab money at hand. This is simple enough, and would work well enough. Of course the media would not accept it and all that, but that is a different story.

                  What you mentioned about one place also makes sense in cases like this. The “what happens in Vegas stays in Vegas” kind of logic, that some places just work as valves to release steam. I think it would make sense to have places like that fully legally, where people can go for a week of holiday, and basically indulge in anything, but most of the second-hand harm of that stays there, and then they come back and become sober normal boring productive people again.

                  (No, not via seasteading. Seasteading is an invitation to get invaded by any fourth-rate military from any rogue state that wants some loot or ransom.)

  1. Hope you like beer and sausages, Dave . And if you stumble upon Daniel Vavra, tell him his games are good and the SJW critics who know nothing about Czech history can fuck right off.

    1. I ran into the “U Sudu wine bar” while here. I HAD to go in due to the pun, and was glad I did – it had a piano, and I’ve played there every night since til dawn. I’ve met and jammed with so many great local musicians, there. It’s 3 stories down, filled with tobacco and pot smoke, with great beer and goulash.

    1. Not quite true. NZ moved Military style semi-auto’s from a Class A (Basic) Firearms licence to a Class E licence (Anything up to and including belt fed full auto).

  2. I can’t think of any better place to ask this, so I’ll do it here:

    Sometimes, when I’m in hack mode and somebody wants my attention, I’ll ask them to leave me be for a few minutes while I “get stuff out of my head”, i.e. finish up whatever’s currently occupying my brain state, or write it down stream-of-consciousness style so I don’t forget it.

    Is there an existing term for this? I don’t recall one from the jargon file, but nobody’s invented search-by-fuzzy-description-of-abstract-concept yet.

    1. It depends on which concept you’re talking about. If you’re referring to having a lot of stuff built up in your head, there’s hack mode itself — which entry describes the very phenomenon you’re talking about — or perhaps juggling eggs.

      If you’re referring to getting the stuff out of your head, the human-memory senses of swap apply.

      1. Ah, yes, thanks. Swapping out is the phrase I was looking for. I could swear I used to know that.

        1. I’ve been known to say “I don’t have that swapped into main memory” when people ask me about past experience that it takes me a while to recall.

    2. >Is there an existing term for this? I don’t recall one from the jargon file, but nobody’s invented search-by-fuzzy-description-of-abstract-concept yet.

      I tell my wife I need to unjuggle some eggs and she understands. This is from the line in The Mote In God’s Eye “We juggle priceless eggs in variable gravity.” I’ve heard other hackers use this. I don’t think it’s in the File, though. It is.

      1. I think by now I have read everything Niven, Pournelle and Heinlein wrote. Are there any other old-timey SF authors who could write something up to the quality of Mote, worthing checking out? I like if there is a general military theme, even if, like in the Mote, it is not primary, but that is not a necessary requirement.

        1. >Are there any other old-timey SF authors who could write something up to the quality of Mote

          Poul Anderson leaps to mind.

        2. I don’t think there’s many writers out there anywhere that can touch Niven or Pournelle.

          I have almost cried myself to sleep at night knowing I’ll never be able to sign on with Alois Hammer. Well, not really cry myself to sleep, but it’s a thing.

          1. It’s hard to find a write of fiction more historically literate, and accurate as only a borderline autist can be, as David Drake.

        3. James P. Hogan has some excellent hard sci-fi. For living authors, John Ringo’s sci-fi is good, his other stuff… not so much.

  3. Those lyrics about Microsoft seem downright quaint in 2019. At least Microsoft was a software company. Today, the browser and OS monopolies are held by an advertising company — and the data they and the other tech giants collect on you and the uses to which they put it make early 2000s Microsoft seem like the Diet Coke of evil.

    1. Honestly, it’s hard to revise a song! I have another one, called cybernation, written in 1994, where I make jokes about irc and second life, and try as I might I haven’t come up with a good rhyme for “facebook”. It’s still a pretty funny song, if you remember what those were

      I played a ton of music on this trip, that one got recorded here:

      (and there’s also a clothed version of “It GPLs me” there:

      Lastly, I have a wonderful song about spacex called “one first landing” that I’ve done a dozen times this trip that I still don’t have a recording of. The chorus is:

      “We tried rogallo wings, and other things
      and nitrogen tetroxide
      and all I dreamed was of rockets landing
      with their engines flaming high!”

  4. On the subject of networking, I’m encountering a problem in my day job, where I work as a network tech, and I’m hoping someone can help, either by writing a program or advising me on an existing program.

    The problem goes something like this:

    Before accepting a circuit and attaching it to a router, it’s SOP to test the circuit’s speed. Usually this is done with a speed-testing website like or . The on-site tech runs 3-5 such tests, communicates the result to the Network Operations Center, usually via screenshots, then the circuit is accepted.

    However… Such solutions are generally only good for circuits up to a hundred megabits per second, though there are a couple websites which seem to give good results up to 250 megabits per second. After that, the fact that you’re running an HTML-5/javascript test inside a bloated web-browser running on top of a bloated operating system… this does not handle higher speeds well.

    To make matters worse, these web pages seem to do their testing by uploading and downloading large files – 2 megs, 4, megs, 8 megs, etc, up to 5 gigs and then timing the uploads and downloads, which once again means javacript/HTML-5 running inside a bloated browser, etc., while a file is being loaded by the OS, and this doesn’t seem like a really good way to test speeds. It’s a bit like tying ten engine-blocks to the roof of your Porsche and then seeing how fast it goes!

    Once you’ve given up on browsers you can download a speed-testing app, which takes the javascript/HTML-5/browser out of the picture, and these apps seem to be just barely capable of testing that a 500 Mps connection is good, but I suspect most of them are running on python, ruby, or some other interpreted language, not to mention that these apps still function by up/downloading large files and measuring the timing.

    Which meant that last week’s attempt to verify that a recently installed circuit was getting one full gigabit per second was very difficult and we never quite managed to prove the circuit speeds, which caused major problems for everyone.

    Does anyone have a good solution for any of these issues? I’d specify that a “good solution” would involve a fast, compiled language, probably one with excellent networking and math libraries, the ability to run it on a Linux, Mac or Windoze command line, better methods of speed measurement than downloading/uploading large files, and the ability, after the measurements are run to generate a nice, management-ready graphical representation of tests done and results found. For bonus points, run the whole thing on the network card.

    Any thoughts?

    1. Why exactly are large files undesirable? It seems they’d be more effective at amortising out any fixed costs associated with initiating the transfer.

      I used to set netcat up on each end of the link I want to test and pipe /dev/zero in one machine and out the other to /dev/null using dd. When dd exits it prints total bytes seen, elapsed time, and average speed (lack of disk I/O lets it saturate the network link). If the powers that be wanted pretty graphs I let ping run for a while and made a histogram of the round trip times.

      Both dd and netcat are written in C.

      1. That’s really brilliant. The problem is that I don’t have any access to the machine on the other end, and also must use a Windows machine for work, as certain applications preferred by my various bosses only run on Windows, not to mention that the techs on the other end would probably panic if I said, “I’ve got the ssh daemon running, and you know my IP address. Have fun!”

        My objection is not to transferring large files as a means of measurement; file transfer, in certain circumstances, should be a reasonable proxy for network speeds. But I can’t help thinking that there must be a better method of measuring the link speed.

    2. I don’t know if windows has sendfile, I see reference to TransmitFile, which is supposedly similar. That said, an interpreted language is plenty fast enough to do link-speed testing if there is a sendfile equivalent (sendfile on mac and linux).

      That said, be aware that TCP is slow-to-start, which can affect the link-speed measurement. UDP is not slow-to-start, so if you want to measure actual throughput, listening (via netcat or similar) on UDP, and then trying to send a gig of 0s through is a decent low-complexity way to measure the link speed.

      1. sendfile(2) was added to Linux to clone TransmitFile, because Windows was trouncing it in HTTP server performance.

    3. I also have a weird network problem. Wifi on desktop PC gets super slow. Try everything possible with software to no avail. Meanwhile the same wifi is okay from the phone or from another computer which is closer to the hotspot. Finally, move the hotspot closer (someone moved it a while ago) and it works. For a while. It still gets intermittently very slow, but only for that PC. And we are talking same room here, albeit with things like concrete columns here and there. Makes no sense. Is it possible that the antennae on that network card somehow got screwed? I don’t want to run an Ethernet to it. I find that using wireless things as much as possible and avoiding “chicken guts” (cables) helps with wife acceptance factor, (future) dog chewing factor, small child and all her friends not yanking on things connected to fragile items factor, me not tripping over crap after X beers factor and so on.

  5. It looks like iperf tested the upload speed just fine. I typed: “iperf -c”

    and got:

    Client connecting to, TCP port 5001
    TCP window size: 85.0 KByte (default)
    [ 3] local port 41882 connected with port 5001
    [ ID] Interval Transfer Bandwidth
    [ 3] 0.0-10.2 sec 15.6 MBytes 12.8 Mbits/sec

    Which is probably a good reading for my upload speed. But I’m on an ADSL circuit, and nothing I typed would get me a download speed, which should probably be around 30 Mbits/sec.

      1. I tried that and got nothing back from the server. I’m not sure whether that’s because I’m behind my own (consumer grade) router and the ISP’s router/modem, or for some other reason. I did notice that the outgoing port wasn’t Iperf’s port and either Iperf or the OS didn’t let me force a port change. I’ll try it on my Windoze machine at work tomorrow while I do speed testing on a new circuit and see what happens when I have a private IP address.

    1. >OT: Eric, have you run across the PiDP-11? It’s a simh PDP-11 instance running on a Rasperry Pi driving a replica front panel through the GPIO pins.

      Yes. I have. That front panel is so nifty that if I needed another time-sink in my life I’d make one. I don’t need another time sink in my life

Leave a comment

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