They'll be surprised to hear this.
Re: Next to last paragraph. "North Carolina Fusion" is a soccer/football team. :-)
Icon because of, well, you know.
33 publicly visible posts • joined 18 Aug 2010
Icon because "what a steaming pile... I'll just be on my way now".
Wow, it's been a long time! Cold War era, first paying job. I was a sysadmin for some folks doing computational fluid dynamics in the Mach 40 region. The big difference in a nutshell is that you really can't entirely model the atmosphere as a gas anymore. You wind up treating it at least partially as a random cloud of various molecules and (violently ionized) atoms. Molecules are being torn up and are recombining into other things, so it's not a constant-molar process across the vehicle. The Navier-Stokes equations start breaking down under those circumstances. Solutions in our group were mostly from something that my naive eye would classify as a Finite Differences simulation. There was a guy doing Monte Carlo simulations - he produced impressively predictive results (I was told) but it was one of those codes that ran for three or four weeks at a time on a Sun 3/280. This was openly published - if I can remember names I could probably find some citations. Cannot remember Monte Carlo Guy's surname for the life of me...
Icon because that's some non-trivial energy levels you got there, buddy. :-)
The waivers the last few days now allow drivers hauling fuel to work for up to 14 hour shifts, instead of the previous 11 hour limit. Some states have also suspended the 80,000 pound weight limit for in-state loads.
The exact impact this will have is not completely clear. While the law is one thing, insurance companies are another, and just because something is suddenly legal doesn't mean your insurance carrier will permit it. Also, tank trailers are made more or less exactly the right size to hit the 80,000 pound combined limit with the smallest, cheapest trailer possible. I don't know how many trailers could hold an extra thousand gallons even if the operator wanted to.
(icon because a fourteen hour driver running grossly heavy could well result in one of those)
It's unusual to give someone 48 hours after conviction to report to prison. Usually it happens on the spot. It's almost as though the judge _wants_ this guy to run. Which was completely unnecessary, because the judge could just delay sentencing for a decade and then dismiss the case if there are no more offenses.
Containers don't care if you're running Slackware. Or anything else. I can tear out YYY Linux and replace it with ZZZ Linux and docker will never know or care.
Heck, docker on windows works. I don't exactly recommend it, but it makes a cool party trick. At least at the kind of parties I go to these days, which tend to be held via zoom and that isn't making them any worse...
Yeah, OK, there are gotchas if the container has to reach down into /dev and do - unsanitary - things. Granted.
I'm curious. How many instances of RedHat are running worldwide, compared to Amazon Linux? I can't say I've ever missed not having pre-built CentOS AMIs. They probably exist but they don't show up in the first few screenfuls of options.
Icon: let me grab my coat and I'll run out to the car and get a different distro for you.
Precisely! This almost certainly reaches the level of criminal negligence. It also quite likely will be considered obstruction of justice, and US courts have a tendency to punish that at a similar level as the original crime - _vis_ Andersen Consulting. Judges _hate_ being insulted like that.
(icon because this is going to be fun to watch from a very, very long way away and the local effects will be similarly devastating)
Short form - you just have to try it both ways and see.
Long form - 64 bit code has access to more registers and that gives the optimizer more chances to do something clever. Fewer memory transfers, which is a nice speedup, can lead to less cache poisoning, which is a really nice speedup. After the 64 bit transition in the early 2000s, it took a while for the compiler writers to catch up but now it's barely even a decision. 64 bit implementations of mmap() can speed up your I/O maybe 5%. On the other hand, yes, if memory is in really short supply then you want to avoid paging and swapping no matter what you have to do.
The Genome Analysis ToolKit (GATK) used (probably still uses) SQLite3 as its database. I started replacing all the sqlite calls with ODBC calls, but before I was done we had moved on to other packages. There were a *lot* of calls into sqlite3.
Are there multi, multi- TB databases in sqlite? Oh yes. Yes there are. :-)
One of the big reasons to go to 64 bit architectures was so that address spaces (not necessarily ram size) could be bigger than 4 GB. The mmap() syscall benefitted highly. I think I remember DB/2 UDB version 6 (???) couldn't have a table over 4GB until you went to the 64 bit version. Or maybe it was DB/2 UDB V7. Anyway, sure, you can lseek() into or mmap() a 100+ TByte file, almost certainly bigger, with no problem.
Not sure what all MDM can do, but on the Pi or any other Linux box you should be able to do everything with either (1) Sufficient Chef/Puppet trickery or (2) just network boot the board each time.
With that said, I haven't had good luck with RasPi in moderately demanding applications. I've given up on them in outdoor settings. In hot weather they crash a lot, and if it gets cold enough then it never boots again. I've had good results from Beaglebones and from the devilishly expensive ARM boards from Technologic Systems. I've run the TS-7800 at 78C all day, fanless, no problems, but it's over $200.
Icon because 78C is way too hot to handle without gloves.
In a very real sense, MMX/SSE/SSE2/SSE3/SSE4/AVX/AVX2/AVX-512 does this. AVX-512 could be viewed as x86_512 for that matter. :-)
The big gain in the move to x86_64 was the ability to directly address more than 4G of RAM. At the time ('05?) that was becoming important. IBM had it in POWER and it really made a difference in DB/2. An even larger address space, combined with some smart virtual memory, could make it easier to map very sparse, high-dimensional problems onto a (mostly?) linear address space. Which would be data warehousing and analytics jobs, so yeah... bring on x86_1024!
Wasn't Secret Sharing ("Secret Splitting") designed to solve more or less exactly this problem? Some number N out of a larger pool of M people's knowledge is necessary and sufficient to perform an action. Kerberos Ticket Granting Tickets are often distributed this way to prevent their accidental or deliberate disclosure or loss. Presumably other uses (launch commands?). This at least could be implemented mechanically - N tumblers each raise a pin incrementally on a ratchet until it rises high enough to open the door, not requiring all M. https://en.wikipedia.org/wiki/Secret_sharing
Movement should be easy enough - given the stakes and a phone call or two, diplomatic or visiting military credentials can be issued. US State dept. courier flights are almost certainly still running.
I suspect they'll announce a new aircraft, the 737-1000. There will also be a field-installable upgrade kit to refit existing MAXes and make -1000s out of them. The kit will consist of a thumb drive, a 22,000 page installation manual, and an INOP sticker to go over the blank spot where the AoA light might or might not have been. Oh, and the most important part - new seatback exit cards.
I was thinking just that. The Curiosity wheel reminds me of a "Bar Turf" tire - an "R14" in agricultural terms. Looks like they're going back to sand paddles. As far as unequal spacing, Curiosity's wheels have some lines milled into them that spell "JPL" in morse code. The wheel motors have position encoders, but the lines can be seen with the cameras to help calculate wheel slippage. https://www.nasa.gov/mission_pages/msl/news/msl20120829f.html
There's a good article on the Spirit and Opportunity suspension, motor, steering, and wheel design. It's paywalled (IEEE), but if you can get to it look for "Mars exploration rover mobility development" in google scholar. The wheels are milled from a billet of 7075 aluminum.
I've noticed that when I use my phone as a hotspot and connect to it with my linux laptop, the connection to the outside world tries to go over IPv6 first. Given the chart's spikes on the weekends, that suggests to me it's mobile users. Add SARS-CoV-2 keeping people at home and I suspect mobile usage is much, much higher, which would account for the increase at the right edge.
The glue that holds it all together? The Database. Banks are just databases with some retail branch offices. :-)
I used to work in a (US) FDIC (our bank regulator) regulated environment - we were technically a bank, and when I left we were adding some obviously "retail banking" kinds of things - checking accounts and mortgages, for instance. I got out ahead of the mortgage-backed debacle, but it was pure luck.
We were moving to a sort-of microservices model, but mostly using MQseries Transaction Integrator (formerly NEON, now Who Knows What). Systems were 90+% MVS (IMS and CICS+DB/2)
Also, am I the only one who saw "Cassandra" and immediately thought "Free money for everyone!"?
A quick look at their website and wikipedia suggests they're basically a prepaid debit card business, maybe even just a reseller, who recently added checking accounts. In the US you could outsource all of this to First Data Corp. Many, many small banks do just this - "The Bank of Southeastern Crud County" (over 500 customers!) provides a building, tellers, and a few million in startup capital. First Data does the whole back office, even prints junk mail if you want them to. It's conceivable that Monzo's entire system is just/mostly interfaces to vendors.
So, yeah, if they're 95+% virtual, this would work. FDC (just as an example - I don't endorse them) would provide the database of record and even Cassandra would then be survivable. They already mentioned the big card providers, plus Apple Pay and a few others, so perhaps many of their 1600 systems are marketing and other ancillary activities. Evidently it works, but given how much money they've had to raise I doubt it works cheaply.
I've been peripherally involved in some machine translation projects in the (far) past, and I'm curious. Does anyone have a feeling for how hard it would be to write a Swift-to-Rust translator? Probably have to be a hybrid of a compiler and a macro substitution mess...
Eons ago some of us did a Prolog ruleset that turned microsoft basic (with line numbers!) into Fortran 77. The mapping was really, surprisingly clean, and compiled Fortran on an Alpha totally clobbered Street Basic on a 386. :-)
Sounds like old MacOS version 7-9 code was brought over via the old Carbon API, which kind-of-mostly made OSX look like Old School MacOS. Carbon was 32 bit only and never upgraded. To be fair, it was trying to emulate baggage brought forward from the Motorola 68000 days (pre-mid-nineties) so it might have been a beast of a job. Ints are easy, pointers aren't.
Anyway, Adobe got hit by 32-bit-only Carbon some years ago. And you're right - Carbon to Cocoa is basically a rewrite of the GUI and a good time to see how well you isolated algorithms from UI. :-)
I'm not being sarcastic, snarky, incendiary, whatever (I know, it's The Register, right?)... I'm genuinely curious. Does this machine contain much Apple-designed content? When they said they were going to "make" them in Austin, I figured that was a code phrase for "rebadged Dell", maybe an R740 in a deskside case. Or an emergency white box clone. The clandestine meeting with Gruber _et al_ suggested a desperation move.
In the US, engineering is regulated by statue and licensed at the state level - there is no federal requirement for an engineering license - and there are a ton of state-level exceptions, BTW. This means, and it has repeatedly held up in court, that the design of things "suitable for interstate commerce" is protected under the interstate commerce clause of the constitution. Approximately the only Electrical Engineers who are licensed Professional Engineers are the ones who do electric power generation and distribution - it's hard to ship a gigawatt generation facility across state lines. On the other hand, aerospace engineers are very rarely licensed. It happens, of course, but it's rare.
The case, like the Oregon one, will be tossed once it finally makes it to court, probably when he testifies that he was able to transport the umbrella in question to another state. :-)
In Canada, by contrast, the term "engineer" is protected not by statute but by trademark (!!!) and the rules are (a) different and (b) unfamiliar to me. Other former client states of the empire will probably vary as well.
Icon because I'm checking my jacket pocket to ensure I still have my "50 states and Canada" cellphone in there...
I'm curious about the supposed roughly 1/4 of US-to-Google traffic being IPv6 - is this pretty much just mobile users? Is there only one AT&T Mobile-to-Google link, and it happens to be IPv6? Because short of renting an instance on AWS to play with, actual IPv6 is pretty darn thin on the ground over here (i.e., I've yet to see any in the wild, not on a private LAN in my house or my AWS instance).
I'm sure the stuff exists. Hurricane Electric was advertising it, what, a decade ago? I just think it's funny that I've seen more Space Shuttles than I have IPv6 requests that weren't just me talking to myself.
(Icon because I'm basically clueless)
Turns out I'm buying a tractor in a couple weeks (I have to do it in this half of the year), and Kubota has started making bigger machines. The local K. dealer is too polite and professional to bring it up, but when I asked point blank he said they're seeing more business from soon-ex-JD customers. Their new stuff won't necessarily replace Deere's biggest machines, but they're moving up market. I suspect the plan is to chew JD's toes off until they can swallow the whole body. :-)
This doesn't sound impossible, but definitely hard. It's fundamentally a lightfield camera in reverse - COTS from Lytro, but making an image is harder than recording one in this case. There's a three-way tradeoff between cost, planar resolution, and depth resolution. I'm curious what they've chosen to optimize for. Also, I thought I remembered, and Wikipedia quasi-confirmed, a mash up of a lightfield camera and a lightfield illuminator on a microscope. So the illumination part has been done, albeit I'm pretty sure with a glacial framerate, maybe measured in something other than seconds. The backpack might contain a computer, or it might be something like gigabit ethernet over 60GHz microwave to a killer machine and graphics card(s), or a board full of FPGAs, or who knows what. It'll be interesting to see the price point they hit, and if they have to subsidize the first ones while they continue to design out some cost. This is just tantalizingly plausible enough...