Re: Well deserved
"Citation Needed", or in plainer English:
C'mon @batfink, this is El Reg and we're all stuck at working at home^H^H^H with streaming. What scenes should we be looking for, in which movies?
290 posts • joined 7 Sep 2016
Reading your last comment reminded me of working in Birkbeck College library some 30 years ago. Unlike the rest of the universe, Birkbeck used a cataloging system called bliss that was...anything but. Instead of Math being at 510, say, it would be at some code like QEL/EF. They had the loveliest librarians to help you but my word, you were sunk without them.
Put it this way: the second edition of the catalog has been in development for a decade longer than the HURD kernel...They started in 1977.
Oh lord, OS/2 from floppy was painful in the extreme. For some reason OS/2 was really, really picky about the quality of the disks. It wasn't wrong to be picky, but back in 1994 at the peak of the cheap-is-fine era in Windows it was painful. I still have nightmares about the summer I had to install OS/2 on thirty Viglens in a library.
Am I imagining it or wasn't there also a little flurry of (dunno what to call them) character-mode GUIs? They would split up your screen into "windows" and you could do different things in each, eg calculator in one and a wordstar-y sort of WP in another.
Is my memory playing me false?
aha! Another word nerd. Love it.
"Dumpster fire" is another phrase with a mysterious etymology. Obviously US English, but seems to come from nowhere to prominence in the last, what, ten years? And, these days its often coupled with the adjective "raging". Where did that come from?
That era used freakishly heavy cabling too. You'd have cables about 1cm across connected via 9-pin D shells to the NIC, and woe betide you if you didnt screw the little buggers in tight. And at the other end you had these things called MAUs where the connector was the size of a baby's fist. I had a job at the time that involved a lot of moving PCs and moving the cables was an utterly miserable experience.
At the time the successor to TR was supposed to be ATM. 155Mbits compared to the measly 10 of Ethernet (and IIRC 100Mbit was still stuck between Fast Ethernet and HP's 100Mbit VG technology). But ethernet was 10x cheaper and worked well enough. The rest is history.
Inspired by the discussion here about " UPS batteries dont last forever". UPSes never seem to kick in when they really need to: we've all heard stories of the UPS silently failing, or ruthlessly shutting down the server and causing its own damage, or the battery pack expanding until one day it looks like Jabba and cant do anyone any good, etc etc etc.
I posit that UPSes are actively harmful these days and what you should do is design applications and server subsystems with redundancy. e.g. independent power feeds into your data center, app pools/clusters than can survive nodes going AWOL, etc.
I could be convinced, but whenever I bring this up to colleagues they look at me with horror. Am I insane? (Don't answer that last bit.)
I am firmly of the opinion that absolutely nothing will change until businesses feel a direct, immovable financial impact from failing to secure their systems. I'm not talking about having to pay for a year of credit monitoring here, that is a fig leaf. You need something of similar monetary force to an audit statement that basically says, with force of perjury behind it, "my network is clean" like an audit statement says "my books are clean". Failing to have that, or having to retract it, is then to be ruthlessly punished by the market.
When businesses feel fear -- real, brown-trousered fear in the C-suite -- then they act. Until then, nothing much.
I think The Great Torvaldo is making one good point and missing the larger one. It is a scary thought to not be writing on the same platform that you are deploying on. Absolutely right. But it does not follow that you should stay on one, nor that choosing not to is futile.
For one, we can all tell stories of how Production turned out not to be like Dev because the latter had some dll or package that the former didn't. So you already have to have very strict processes for avoiding these types of snafus, and once you have those, you've broken the major objection to cross platform dev.
Secondly, the implication that building on x86 is the only way to establish quality is demonstrably false, when you think about it, there are many, many more non-x86 (eg arm) devices out in the field than x86, and they work well. Phones, routers, etc etc. So clearly the cross-building folks know what they are doing.
I do agree with ARM that having some rocket fast reference hardware would be a Good Thing though. One of the sad changes in Linux over the years has been the gradual dying off of non x86 Linux distros. It would be good to have some decent hw at a lower price point than x86 to resurrect that.
In the Olden Days when bobajob12 was a wee lad, a small number of stores would have a reputation for their shop window display. The toy store would have, say, a model train running up and down the Alps with all the props (even if, inside, the best you could buy was a circular track and a map of Bradford). A couple of stores still do this for Christmas, e.g. Bergdorf Goodman in New York.
I wonder what incredible store displays an RPi store could come up with? You could even make it a competition.
I detect a little snark against Apple in this article, but FB and GOOG were being quite naughty here, and the rules of the Apple walled garden are very clear: you're a guest in their domain, and if you misbehave, you will be thrown out. FB wanted to slurp data, couldn't get that past the Apple censor, and tried to sneak around it.
So yeah, getting official Apple approval is like having a root canal, and they can be hellishly annoying, but on this one, I'm glad they slapped the miscreants down.
Sent my from Android device :)
Not just MSFT. Filesystems are *hard*, though. It's not just a question of sed s/32/64/g and recompile. I mean look at the pain that Microsoft, who had a compelling need to fix theirs, had with the whole WinFS project for NT. They announced it as part of Cairo in --what, 1992? and it never completely shipped, even ten years later. I consider SGI XFS and Sun ZFS to be two of the best filesystems ever created and even they had issues.
As for the PST file format, though, you got me. That should have been set on fire at the earliest opportunity.
Yes! It seems that a commonly missed factor in migration planning is "what does the system look like now" - right down to inspecting log files for what "normal" really looks like. The gold standard is that a good system "runs clean" - you can explain EVERY entry in the logs, and there's nothing there you can't explain. Got a tiny little assert logged? Run it down. If you can't do that, then you had better know well in advance that these things are "normal" and that you shouldn't freak out after a migration when you see them again. It just cuts down on stress for everyone.
We should take a lesson from people who deal with drug users: harm reduction not behavior elimination.
Most people will be exposed to malware on a stick, but unlikely to be exposed to chip-level attacks. And they are never, ever going to stop plugging things they find into their computer.
So, what actions does a user need to take -- a real user, not a Reg reader, mind-- to protect themselves from nasties on the stick when they plug it in? This kind of stuff is too much.
+1. In the classical period (1995-2005) being an enterprise DBA was pretty hard. You had to know your OS inside out, you had to understand how to do fast I/O and then you had to design the database, as in tune how tables should be laid out for maximal performance. Wasnae easy, cap'n.
These days, I'm not so sure (I'm out of the biz). Do Oracle admins still worry about tuning the SGA for example? Does anyone use raw block devices? Or is it all self-tuning and autonomous for all but the most esoteric setups?
Let's say - they had their data processing operations in Osterly and were known for a shop building in Knightsbridge that was bedecked in lights?
Frankly, it's a miracle that they were running anything at all. I remember a summer working in their credit department where the workers had the thrilling task of microfilming credit card applications. One page at a time. It was like the sheet feeder had never existed.
I support a name change. Frankly the only people who would find "NIPS" amusing are those with the minds of 12-year-olds, but --breaking news, not!-- there are a remarkably high number of such folks in attendance at conferences. Including this one. So why give the puerile crowd anything extra to exercise their little minds over?
There's nothing inherently wrong with a name like NIPS. But the record of this and other similar conferences makes it susceptible to a rule I may as well call This Is Why You Can't Have Nice Things.
Ignoring the very wealthy markets of North America, Chinese cities and Western Europe, this level of repairability could be huge. One of the things holding back smartphone adoption globally is that not only are they costly but they are a royal pain to fix.If you have a smartphone that your guy/gal in the market in Kigali or Bangladesh or the wildest corners of Romania can fix, you're on your way to global adoption.
It's what you might call the Toyota truck model (or, for older Reg readers, the 2CV effect, or perhaps the Land Rover Defender effect) - things that are fixable in market tend to get very widely adopted.
True, although I'd rather be pranked with a rotated screen by the tech support people in the basement than stabbed in the back by an accountant with a spreadsheet on the top floor.
Time and place for everything, but a good manager who makes it very clear where the boundaries are also leaves space for the humor to grow, which contributes to better productivity for all.
Is that a good sigh or a bad one?
Took me a year of cursing at WP5.1 before I knew enough keystrokes to be useful in it (the little plastic Fn key template being as useful as a chocolate teapot). But once I got it, ye gods, what a productive tool. Much as I like Word these days I still miss the fullscreen blue mode. Word even had a blue mode itself for a few years but sadly they ripped it out in recent versions.
I have a bit of nostalgia for the sheer variety of computing back then - 68K, 8080, even the weird-and-frankly-quite-broken stuff like the 80186 (IIRC Research Machines used to have a demo mode on their 80186 PCs that could play a little Bach fugue). And the tools, being forced to run in such compressed environments, were pretty clever (and cheap). TSRs like Sidekick. Zortech C for 29.99. The BBC Tube (I mean! Co-processors across an interconnect! For kids in schools!)
Then I remember just how hard it was to get anything done. CONFIG.SYS not exactly right? Sorry. Don;t remember the exact PEEK and POKE? Sorry. Segment:offset addressing? Yuck. It's a miracle anything happened.
If the point behind your anonymous comment is that the ACLU only ever take on cases that reflect some left-wing bias of theirs, I can assure you that they don't.
For example, right now they just filed suit in support of the NRA against the state of New York.(link). They came out in support of the white nationalists' right to protest in Charlottesville (link). Those are just two recent examples. If you want to go further back in time, they even defended Col. Oliver North in the Iran-Contra scandal. (link)
So, less of the cheap shots, eh?
I happen to think the defendant is a loon, but this sort of case was never going to fly. The Internet, notoriously, interprets censorship as damage and routes around it. Doesn't matter whether the 'censorship' is 'good' or 'bad'.
The really interesting cases will start when people start making these guns and hurting themselves or others. I predict an upswing in hand and facial injuries as the first experimenters discover that machining parts to close tolerances is, uh, quite important if you want that explosive projectile to go in the direction you want.
True, but @Mk4 didn't claim it was a hack job. For all we know, they could have documented it up the wazoo, pointed out it's criticality, and even gotten sign off from a Higher Being...but that's no guarantee that a busybody in another part of the org couldn't insist it be taken down.
Being real life, of course, this can go both ways:
A: "Ach, I'll just roll my own crypto". Busybody: "Hell no" - BB probably saved the day there.
A: "Ach, I'll just write some glue code" Busybody: "Hell no" - BB probably cost the company $$$ as now the swarms of IBM/SAP/$expensive consultants arrive to tear apart the business.
Check out Derek Lowe's chem blog for more on this. Like: "Imagine 6 skunks wrapped in rubber innertubes and the whole thing is set ablaze. That might approach the metaphysical stench of this material"
In many parts of the world, Bangladesh included, the concept of make-it-work trumps adherence to "code", or saftey, or anything we take for granted in the developed world. See https://huqelberry.wordpress.com/2011/03/05/crazy-cables/ for an example (from Pakistan). You see it and you are simultaneously impressed and horrified!
An AC asked for a source on Singapore. I use that as one example. The key tenet is that the EU needs to have made an "adequacy decision" for a country, and Singapore isn't on that list, ergo, is not considered adequate by the EU.
Here's the official word from the EU: https://ec.europa.eu/info/law/law-topic/data-protection/data-transfers-outside-eu/adequacy-protection-personal-data-non-eu-countries_en
And here's a link from a local Singapore law site: https://lawgazette.com.sg/feature/the-new-european-general-data-protection-regulation/
IIRC, India (and Singapore, and few others) is one of the countries that the EU considers to *not* have adequate data safeguards in place for any company subject to GDPR to use as a data processor. You can imagine, that this puts a dampener on all that yummy outsourcing.
See: https://www.pwc.in/consulting/cyber-security/blogs/how-can-indian-organisations-prepare-for-the-gdpr-regime.html for a brief explanation.
El Reg readers should be aware that if they are working for an EU company (data controller) that uses an external data processor (such as an Indian outsourcer), they need to tread extremely carefully. The GDPR has real teeth. You do not want to mess it up.
Ah, PA-RISC. A lovely system. But this battle is lost. It's not so much what "makes commercial UNIX better" as what made, past tense, it better - and those advantages (deep talent pool, corporate backing, intensity of focus, knowledgeable customers, dedicated hardware) are no more.
To whit: All the Solaris talent left Sun/Oracle years ago. Judging from HPE's April 2018 roadmap I imagine HP-UX is supported by three people in a garage who weren't even born when the Superdome and Integrity lines came out. HPE is a shell of its former self and barely seems to know what it is any more. Both Oracle Solaris and HP-UX are running on fumes and all the innovation in their OSes is coming from other communities (eg OpenStack on HP-UX). Customers have their own business imperatives and Linux is good enough for 99% of what they do, plus the talent to run it gets cheaper every year, so "good enough" is fine. Hardware margins are brutal.
When I retire I fully intend to recreate the computing environment of the late 1990s and have a blast just playing around with all the heyday UNIX systems I can lay my hands on. The UNIX Wars were ridiculous at the time, but with nothing to look forward to but hearing aids and senility, I shall have a blast firing up IRIX and OSF/1 and HP-UX just for the hell of it. But not AIX. Please God no.
Biting the hand that feeds IT © 1998–2020