Posts by bombastic bob

flaws in nano

the only complaint I have about nano is if the Linux distro (say, Raspbian) set up HIDEOUS colorization and I'm trying to edit a file and it's dark grey on black and I can't freaking READ it...

That's editable, fortunately, once you've found out where it is. But I'd like a switch or ~/.something setting to turn it OFF, like I do when I 'unalias ls' and things like that. Colorized text on black background NEVER looks right. It should be easy to turn it ALL off without mucking with 'term' settings in .profile or .bashrc or elsewhere.

color choices aren't just for "I like" vs "I hate". My ~60 year old eyes need less strain, and so I minimize that with off-white backgrounds in the GUI, and white-on-black consoles. Turning off hidesouly dim colors on black background (favoring NO colors) would be a nice start, especially if it's just a single line in a config file, or something I can 'alias'... or 'unalias' as the case may be.

Re: Cult?

"learning curves" are overrated. I just wanna get those 1 minute tasks over with, not spend an hour learning how to do a "1 minutes' worth" of work, and unnecessary additional time looking up how to do otherwise-simple things in documentation EVERY! STINKING! TIME! because "I never use that hideous thing unless there is no other choice" etc..

This is why nano, and its possible predecessor 'ee', exist. Not for n00bs, per se, but for people like me who don't want to "look it up" all of the time, wasting effort that could be spent doing things that are more fun, more interesting, or maybe even just more important.

Re: Cult?

seems all of the Debian-based distros have nano - they used to include 'ee' but for some reason it's not now (abandoned?). FreeBSD still has 'ee' though.

I'm not a fan of 'vi' but I know the handful of commands that let me use it. Or more more importantly, exit without a reboot (or 'killall' from another console/session). Never type 'edit' in a console unless you're prepared for it. Did that the first time I booted a POSIX system (FreeBSD 4.7 in this case). Frustration ensued, followed by the boot button. i was still figuring it all out.

Nano seems to be most useful on an RPi when you have to ssh in a lot and edit things. I'm glad it's there. That probably makes those who use it members of a pretty big 'Cult'.

Re: Kessler effect

In the animated series 'Cowboy Bebop' an accident in space resulted in chunks of the moon clogging earth orbit to the point where it's hazardous to navigate (in space) and large orbital rocks are frequently (and somewhat randomly) falling onto the earth, causing craters, damage, etc..

A worst case sci-fi scenario, yeah, but something to avoid. Let's not have to dodge existing satellites when firing rockets into space, especially as more countries around the world get their own space programs going.

(or will booster rockets have to use particle beams or other space-weaponry to blast obstacles out of their way in order to avoid collisions during launch?)

Another, this time literary, reference to a book I read as a small kid: "If Everybody Did" [lampoons what would happen if everybody were to do certain things, taken to extremes in a funny-when-you-are-six kind of way]

Re: Use of personal data ?

worse, as a "Main In the Middle" they have the capability to inject HTML ads into the traffic, including your e-mails, and ALSO edit out any ads (or content, for that matter) that they do NOT want on their network (or do NOT want YOU to see...). Like with a firewall appliance, they could enforce the use of THEIR SSLcertificates via THEIR trusted authority, and so on.

Using 'https' won't mean doodly squat if they manage to pull THAT off. Not saying they WILL, only that they COULD...

(so to avoid FUD-ishness, I suggest that appropriate regulations may be needed to PREVENT that possibility)

"big tech" --> "not a monopoly". Right...

One of the things done with fuel providers here in the USA is the "divorcement" between oil wells, pipelines, refineries, and fueling stations. You can't own the entire thing from end to end, or else price fixing would/could be used to drive competitors out of the market. So what is it called if Amazon owns the "last mile" in addition to (many) other parts of the internet?

fortunately, with server-generated keys [and plenty of warnings if they ever change], ssh isn't likely to be so easy to "Man in the Middle" successfully. So "a mitigation" exists, at least for some of this. For in theory you could use a server with sshd running "not on their network" as an ad-hoc VPN - kinda. But the vast majority of people relying on an Amazon network like this probably wouldn't have that option available. Commercial VPNs, and even TOR, could be just as "pwned" as SSL in the scenario I described.

I wonder if, like cell towers, you could privately invest in your own satellite fleet, and just rely on 'roaming charges' to fund it. You could be a member of multiple competing networks, even. Then it's decentralized as far as ownership goes, (potentially) making it less possible to do 'tricky things' like the "Man in the Middle" HTML injection/snooping scenario. Well, it's a thought, anyway...

Or, maybe the temptation to be your OWN "Man in the Middle" content injection/filtering ISP is too great... and others would follow THAT path.

Re: the real life physical dangers inherent in attaching all these home appliances to the internet

It seems obvious that malicious firmware could cause all sorts of very spectacular problems

maybe to El Reg readers, but apparently not to the makers of this particular 3D printer...

(I doubt they did this negligently, most likely they just didn't know what could happen if the firmware were maliciously crafted)

Re: Hydrogen is not the answer

burning diesel in pure O2 results in NO NOx and over 90% efficiency

Although the 90% efficienty claim breaks the 2nd law of thermodynamics (you would need a temperature of more than 2700 C to obtain a Carnot efficiency of 90%, let alone a real world efficiency of 90%, and last I checked most metals would fail to hold integrity at that temperature...) but it's not likely for a number of OTHER reasons as well.

Still the idea of using pure O2 and fossil fuels is not a new one. Rockets. And at very high altitudes, inject O2 into a regular jet turbine engine. The only problem is getting pure O2 in sufficient volume to make this happen. So if you could solve THAT problem [currently cryogenic methods are the most efficient method I know of to get pure O2, and it takes a LOT of electricity to do that] then O2 bottles for your diesel engine might be a good idea. If for no other reason, spike up the O2 to increase efficiency.

Still, I think it'd melt...

NOTE: Rockets avoid the melting problem by using incoming fuel to cool the nozzles, and also spraying a layer of fuel onto the inside of the engine. it's a bit less efficient, but the evaporating fuel would (to some extent) protect the metal of the engine from being melted. Burning with pure O2 tends to make welding-temperature flames and as such the metal just metls.

Re: RFC Electric Greenhouses 1GW = 20 SQMLs

concentrated mirrors, shining on dark boiler piping, can produce steam for electricity. Ufnortunately you need a LOT of unpopulated area for one of these. Solar energy density is generally too low to be practical on a VERY large scale without taking up a LOT of space. A solar panel farm is probably the most space-efficient way to do it, yeah, with current technology.

When a parking area has covered sections that have solar panels on them, it's a creative way to make the parking lot better, AND use some of the space to create solar electricity. Another win-win. Seen this at a Walmart in San Diego, where it's sunny enough most of the time to be a practical location for solar electricity.

Re: Hydrogen is a solution and its own worst enemy

using biofuel production to consume agricultural waste IS a good idea on its own. It is possible to convert bio waste into oil. The process that I read about is actually EXOTHERMIC. You could then use this bio-synthetic oil to add to existing oil production, thus "offsetting" either a) trash production wasted as landfill, or b) whatever environmentalists decide is "helping". Either way it's a win-win so unless there's cost prohibition making it completely impractical, I say "go for it".

Note that SOME of this agricultural waste goes into composting and may still be useful as animal feed. So whatever does best for farmers, probably should do THAT first.

Re: Hydrogen is not the answer

If the hydrogen plant gets its water from a river, there would be no brine to discharge.

an undeniable point. Common sense, WHAT a concept!

Re: Makes sense both ways

checking for updates every 4 hours, though - that's ALREADY ridiculous (nevermind the DDoS'ing 4 minute setting)

How about every 4 weeks? 4 months? 4 YEARS? or, my favorite: when _I_ want to check.

Re: I haven't chuckled so loud in ages

Yeah El Reg as "required reading" for I.T. (and related) staff...

just 'On Call' and BOFH and the occasional IT security-related article would be enough for "basic require dreading", but I'm sure (like any other IT-related person) they'd be spending a significant period of time during the day looking at the rest of the site...

or how about just pointing out the silliness of the anti-vaxer conspiracy theories in general?

I happen to like vaccines very much, I was a pincushion for them when I was in the Navy, and occasionally go for the inexpensive flu shots when available.

So yeah SOME real dangers from vaccines, but you just have to have the smarts to watch for them. Flu shots CAN give you the flu, or a rash, or other allergic reaction. But they usually don't. So you either get a doctor to monitor it, or show some common sense and read the warnings on the label...

(/me not a doctor, so YMMV)

Re: Secure Bag

the general practice with classified material is that it doesn't leave the location where it's stored. That was how it was back in the 80's, and I doubt it has changed. If you must transport it, you're supposed to do so in an approved manner. Although at times this may have been bent/gray, it was still "the rule".

In any case, working from home with a government contract, where security is concerned, is generally NOT allowed, EVAR. "Lose your security clearance" is just the tip of the iceberg.

Re: why aren't postal votes considered a fraud risk in the US?

"Without an iota of evidence."

you mean like THIS EVIDENCE ??

others also exist - this was one of the top results in a google search for "mail vote fraud evidence US", that wasn't a "snopes" or biased "fact check" site...

(you're welcome)

Re: I almost wonder if his politicization of coronavirus was deliberate

I don't know which seems more politically biased (in my bombastic opinion), the comments, or the article itself...

(yeah I'm a Trump voter but you knew that already)

[I need a bigger thumb-down icon]

Re: Best wishes

agreed. the entire idea of privatizing space (so it becomes less prohibitively expensive to explore it) ought to include other nations' space programs as well as private companies in the USA, etc..

I wonder when India will send something...

/me wonders how long before a Russian company launches rockets from Baikonur... (or have I missed something?)

Re: Why...? Just Why?

Any kind of forced "phone home" updates _ARE_ highly overrated... and the IOT is _NO_ exception!

It was a lesson in "why the Micros~1 way is BAD". Meanwhile, many bricked boxes later, they were ALSO caught revealing your private viewing secrets by regularly 'phoning them home'.

*BAD* Sammy. No biscuit!

Re: Let's look at a few Rust facts ...

"But for the common use case space really isn't a big deal "

Except... when it comes to KERNEL programming! It is a completely different thing. Tiny and hyper-efficient. That's what KERNEL programming requires.

And for THAT reason, I question the "fit for purpose" of Rust as a kernel programming lingo.

Re: Let's look at a few Rust facts ...

" they call write(2) to print an identical character sequence to the stdout file descriptor."

uh, wouldn't '2' be stderr? Not to pick nits or anything. icon for humor.

But yeah, this program shoudl be trivially small, only startup code for libc and call to 'main()', plus 'write()' call.

On FreeBSD, clang output was a 23128 byte binary. gcc8 however was 14160 bytes. I thought this was kind of interesting...

#include <unistd.h>

int main()

{

write(2, "hello world\n", 12);

}

Re: Let's look at a few Rust facts ...

I like your analysis! This reflects on what it *MIGHT* do to a kernel's footprint, too...

Re: Rust pushes developers away

general observation: your more in-depth analysis seems to make a lot of sense, and is not inconsistent with my initial (gut) reaction to all of this.

"When I try to point out the above facts and how Rust at no point would pass the Steelman requirements, it quickly gets hammered home that there is no intellectual debate possible"

like another 'religious war' in the programming world? yeah, we NEEDED THAT, *snark*.

I should look for examples of THIS kind of thing on my own, to see if I can get info about more than one person's experience with this... and if it's common, it's worth bellowing about.

Re: Compiler Version

"I can't remember the last time I actually built a kernel"

Dynamic modules have made kernel building almost unnecessary, especially if you're not using a device with limited hardware resources where a custom kernel would greatly limit the footprint, etc.. [that used to be a somewhat geek-popular thing, make your kernel as tiny as possible].

I've recently built FreeBSD kernels, most recently a version 13 kernel for an RPi 3 - I wanted to see how well it would work (It worked very well, so I left it running for several weeks with a PostgreSQL database server running on it to test database performance on an RPi for a customer). But if you're doing kernel dev on a module, you need to re-build a lot [but don't clean all first, or it'll take hours - must build incrementally].

Seriously, though, if you're working on a kernel module you do kernel builds a LOT. But otherwise, no real need these days.

but if you're doing 'make menuconfig' every time in Linux - consider 'make oldconfig' - save time. You can hand-edit the config files, too, which I believe is easier once you have something close to what you want. Then edit the config file(s) and 'make oldconfig' etc.. [I used to do this a lot with wifi routers a while back]

A lingo that is 'fit for purpose' for actual KERNEL programming

seeing as I've actually done kernel programming for Linux, FreeBSD, and even Windows (including bit-wiggle peripheral setup things and netfilter modules), I'd like to chime in with my significant concern that Rust inside the actual *kernel* is PROBABLY a bad idea. You need the low-level predictability and lightweight code of a language like C. Even C++ has been wholeheartedly rejected from being in the kernel, for similar reasons I might add.

My understanding of RUST suggests that this language is really _NOT_ well-suited towards the kind of programming that KERNEL programmers do. I'd like to suggest that a language like RUST might have a negative impact on kernel efficiency, memory footprint, and even reliability.

Remember: not everything running Linux has multiple cores, >32Gb of RAM and TB's of disk space available. Some machines might have a 4Mb flash (or even 2Mb) to load the kernel+userland from and still needs to run Linux. Saving that nickel on the BOM to reduce FLASH size down to "tiny" like that is an important feature. A lingo with too big of a footprint would RUIN this.

This is not to say there isn't room for what I'd call "Rust Support" in glibc (for example), which seems to be hinted at in an article linked to from this one. That would be fine with me, of course!

Though glibc and other things may actually be a part of (or closely associated with) the Linux kernel tree, this does not make it part of "the kernel". Still I couldn't see a Linux kernel ever NOT shipping with glibc.

Still, putting Rust code inside the actual kernel: It would (In My Bombastic Opinion) be WORSE for Linux than SystemD, Pulse Audio, or Wayland... combined!

Re: Reply to Linus Torvalds

If certain phrases offends some people it is worth considering changing it.

And it's also worth telling overly-sensitive people that make demands upon others (to do work, change, whatever) where to go and "how far" etc. because *BEING* *BULLIED* *BY* *THEM* and then *CAVING* is actually *WORSE* since they *NEVER* *STOP* and *JUST* *DEMAND* *MORE*.

Now the article doesn't say, but if those "offended" people actually DID ALL OF THE WORK, then it's like "sure, whatever" and no harm just adopting the edits. They're just names for things.

And I suspect it _WAS_ this... and moving forward, it's nauseating, but I *SUPPOSE* 'new terms' could be used instead in the future. Whatever. At least there's no pile of extra WORK to do just to please people who are (in their own way) INTOLERANT.

DOWN side: I suspect that this is ONLY the beginning. *THEY* never *STOP*. *THEY* are never *APPEASED*. It *WILL* happen. Again. and Again. and AGAIN.

We shall see.

(so, WAS it the case that those who CARE about changing terms to be more 'inclusive' or whatever did the actual EDITING? I hope so)

Re: And so it ever was.

And I thought _I_ was Captain Obvious!