Posts by Lennart Sorensen 332 publicly visible posts • joined 1 Jan 2008
I gave up on vitualbox. The random packet corruptions on the virtual network interface that has been going on for years if reports are anything to go by just doesn't work for me. I need it to work reliably all the time. Having an image fail to install 10% of the time because virtualbox corrupted a packet is not acceptable. KVM on the other hand works great and has more features too.
Well CGA, EGA,Tokenring, atari/commodore/sega/etc joystick users and many others might disagree about what that connector is for. Serial only started using it later on, so I think serial is the one that should have picked something else. I certainly do not assume it is a serial port on the machines I have around.
Oh yes I remember SLS (installed it in 1992 I think it was, back when the kernel version was 0.97patchlevel who knows what) and then later slackware which did improve a bit on SLS but not a hell of a lot. Eventually found redhat and realized you could do package management much better than SLS (and hence slackware) had done and stuck with that until the bugs drove me nuts and I switched to Debian which actually had even better package design and management than redhat, and haven't found anything better yet so that's what I am sticking with for now. I never could understand why anyone would enjoy putting up with the pathetic design of slackware (or any of the BSDs for that matter, that user space and packaging system is just horrible to use).
At least we have choices.
And I do use systemd and want to use it these days because it turns out it really is a much better way to handle init even if it is totally different than what we had before. It actually solved problems the other init systems never could because they were all that sad unfortunately.
Well the Debian package format has always been far better than the mess that is rpm, and Debian's updater certainly has reliably done in place upgrades far var longer than Red Hat has ever managed to do that. That was the main reason I moved to Debian from Red Hat about 25 years ago now. Yum and dnf and whatever else RedHat has added since hasn't done anything to fix that rpm is just not a well designed package format, and the source packages are even worse.
If they really cared they would just drop the artificial hardware requirements in windows 11. My 11 year old thinkpad is running insider builds of windows 11, and has bee for years, and has never had an issue. Microsoft offered to install it back before they added the artificial requirements, and of course it installed fine, and still does. They don't have to continue to support windows 10, they can just let all the machines switch to 11 instead. Not a problem except they decided to make it one.
Still sounds better than the apple program here you have to supply serial numbers and such to even get parts for a device. Certainly under the apple program a repair shop can't buy and stock parts to provide quick service, they would have to get your device, order parts from apple, wait for them, then repair the device. Totally useless, but of course that's how Apple wants it to be. It has to sound like they are doing something while not actually doing it.
Well sure for single user to single user you could use pgp, although you need to handle the secure key exchange and authentication first.
Given many of the chat systems support groups, that gets way more complicated.
Different systems have different features after all. Does everyone have to implement all the features or do services have to drop features that are not universally supported?
And if you change the protocol, you just broke compatibility with all the existing clients on that service. So much for making things more interoperable when you start by breaking everyone.
Sure but that still meant you had to sign up to all the services, it just provided one convenient client to manage all of them at once. That's easy.
What this law wants to do is make users of different services somehow (not sure by which kind of magic it should happen) talk to each other without signing up for all the different services.
Lucky for Apple that they went with NextStep instead or they would not be around anymore. Apple needed a modern OS to move forward, and BeOS wasn't a modern OS. Designing a new OS without security and multiuser in mind was idiotic when it was already well known how to do that. Unix had been doing it for decades, Windows NT was doing it (at least Microsoft realized that designing a new system from scratch with some API compatibility as the way forward for them), there was no excuse for BeOS to do what they did. They only fixed MacOS's lack of SMP and preemptive multitasking, without fixing the other major flaws it had. Adding multimedia support on top of any good OS is no big deal. Designing your OS around it at the expense of security and multiuser flexibility (you know a computer could be used by more than one person in a house) was just dumb.
I remember using the BeBOX, with it's nifty cpu meters on the front for the two PPC603e chips. The GUI was terrible (those title bars that were not the full width of the windows looked dumb, and put buttons in places you didn't expect them to be). Eventually tried to put linux on it instead to do something useful with a dual cpu machine, but it was never very well supported and much faster multicore x86 machines became available not too long after.
Given how horribly obsolete and inefficient the vehicles they were planing to get were, it would have cost a fortune in operational costs. They were a complete joke.
And nothing can compare to the politically motivated taxpayer funded waste that is the military suppliers. I suppose the for profit healthcare system tries to give it a run for it's money though.
First thing I had to do with firefox introduced their built in PDF support was to figure out how to turn it off since it failed miserably at rendering the first PDFs I encountered with it. If it doesn't work 100%, do NOT make it the new default without asking and making it really easy to turn off. I don't care if they add markup and editing support since they apparently can't get basic rendering right.
Well the Debian installer has only been used once for the installation. The root user has only been created once in the passwd file, although clearly with the password updated over time (I would hope). So yeah I think it can be considered the same install. It's not the same physical machine anymore though. I have a 486 around with Debian installed on it and upgraded for years. It was installed in around 1999 or 2000, as far as I remember using potato that was under developement because slink just didn't want to install. Haven't used the machine for about 7 or 8 years now because the poor HD sounds terrible (ball bearings don't last forever) and I don't believe Debian supports running on a 486 anymore since 2015. I think I should put a compact flash IDE disk in it and use it for DOS gaming instead. It used to be good at that before it became a dedicated linux server for many years.
I suspect LVM at the bottom to allow combining different disk types and sizes into pools, then raid on top of those pools to redundancy, and then LVM on top for flexibility in allocation of the disk space.
I have never done that, by just throwing money at the raid layer to keep things identical and hence simpler, but I can see how lvm would allow doing it cheaper and with less initial planning required for handling future upgrades.
Having had the "pleasure" of using BeOS on an actual BeBOX, I guess to me BeOS never had a chance because it was a bad design. It wasn't designed with multiuser in mind from the start, and other annoying issues that were inexcusable at that time. Sure it did some multimedia, but so did my Amiga years earlier. Of course I was already not convinced object oriented was a good solution at that time, so the design didn't win any points there either. Apple certainly made the right choice by switching to nextstep rather than considering BeOS as a MacOS replacement.
Using GPS makes sense when you have a GPS navigation system already and you want accurate time without the user having to keep fixing it. Of course Honda has managed to show that they are not good enough to actually use GPS correctly for this. My Prius which does have a GPS system does not use if for the clock. It has a couple of buttons for me to set it. Maybe the Toyota engineers were on to something when they kept it simple, even though I thought it might have made sense to let it at least have an option to set it by GPS. Or perhaps they thought since not every trim level had a GPS, the clock might as well work the same in every trim level.
And at least in the case of linux, they have already been working on the unix epoch issue. 64 bit systems no longer have any problem, and 32 bit ones are close to fixed, so at least they seem to be working on it with about 16 years to spare, which is a lot better than the Y2K work that was rather last minute.
GPS is trying to fit a lot in a small message. So they went with 10 bits for week number and 19 bits for seconds of the week. So 29 bits handles the entire date + time for a 20 year period. Newer GPS signal formats did extend the week to 13 bits making the entire message 32 bit and handling 157 years with second level precision. A rather efficient format really. Unfortunately only the newest GPS satellites use the 13 bit week format, most are still 10 bit, and almost no receivers are taking advantage of the 13 bit option yet, probably since most of the signals they receive wouldn't contain it yet.
Keeping crap like day of week, length of each month in days and such out of the format simplifies things a lot. Converting to human format can be done in software from the raw signal afterwards.
So GPS managed 20 year cycles with 29 bits (less than half of your proposal) while the newest GPS does 157 years in 32 bits (exactly half your proposal). I think the designers were rather clever.
Well I could see for very infrequent use, your click in a window method might be convenient, but compared to the convenience and flexibility of wincompose, it really doesn't seem like a match. And of course hard to compete with free and open source in the case of wincompose. But using a compose key certainly does have a higher learning curve for sure.
The relay disconnecting the hybrid battery stays off unless you turn on the ignition and the computer checks everything is OK. They don't want that done unless you are actually going to start the car since depleting the hybrid battery is a much much bigger hassle than dealing with a drained 12V aux battery. I have only seen it drained when the lights were left on in the car.
Or you could just: unlock driver's door with physical key, pop hood, connect charger to the terminals for boosting the battery, hit power button to turn on car, and the car computer will fire up and start charging the aux batter from the main battery instantly. Totally trivial. Only had to do it twice when someone left the lights on in the car (would be nice of toyota had made them turn off after a while like VW has been doing for multiple decades).
My toyotas both beep at you if you take the key fob out of the car while it is on. If both people have their keys then it doesn't of course as long as one of them is still in the car, but if only one key is with you and you try to leave, you will know. This is not a problem.
Well apparently to replace the keyboard on a macbook requires buying an assembly containing the entire top of the case (case, touchpad, keyboard, battery, etc). Apple won't sell you just a keyboard. SO that makes your keyboard repair $400 rather than $50 that most other laptops would be.
I suspect most people would rather go to a repair shop that knows what they are doing (which means one that isn't authorized by apple of course) and pay $100 than pay apple $400 for an assembly to do it themselves. Of course Apple has no interest in allowing that. They only want to pretend they are allowing repairs.
How is spacex intel? spacex at least tries to innovate and compete. intel looks a lot more like the old dominant players in the business happy to keep selling people more of the same until someone comes along and starts to eat their business (like AMD sometimes does). And intel has made plenty of flops in terms of design (pentium 4, iAPX 432, itanium, etc, not to mention failing to get new process nodes to work while their competitors catch up). intel reminds me of boeing these days. They think they know everything better than everyone else and then things go wrong, but they manage to fix it eventually and claw their way back.
So no spacex is not like intel.
Oh like the Dlink DGE-530T where revision A and B were 3Com/Marvell Yukon chips and then revision C was a DGE-528T (which no one wanted) using a realtek 8169. Far inferior network card and of course totally incompatible with drivers, and since dlink didn't want people to know it was a realtek it had it's own PCI ID so linux didn't even know it was a supported chip until someone (me in this case) added the PCI ID for it to the kernel. We of course told IT to stop buying those cards anymore.
Well they came out pretty much the same time.
Of course given the A12 gives half the storage, half the ram, slower video, LCD rather then AMOLED screen, at lower resolution and refresh rate, weights more, is bigger, has much lower quality cameras, runs an older version of android, no IP67 water resistance, etc, I would hope it was only half as much. But if none of those differences are things you would use, then half the price is great. In my case I bought the A52 (not 5G) since I like the features, but have no use for 5G and saved some money that way, even though Samsung doesn't even offer the non 5G version in Canada, so I bought it on ebay from the US instead, where it appears it was aimed at the puerto rico market (it defaulted to spanish).
16:9 is outdated? Well I remember 4:3 and then 16:10 (1920x1200 monitors were great) before 16:9 took over because everything wanted to match HDTV resolution. So 16:10 is going back to a better resolution in my opinion, but it certainly isn't because 16:9 is outdated, rather 16:10 seems to be the one that became outdated.
So the board regrets it but not enough to undo it? I guess any board that would let RMS rejoin will be too stubborn to admit they made a mistake, and given RMS said he isn't going to resign, I guess the only way he is leaving is that someone kicks him out, and clearly that won't be the current board planing to do that.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
