Re: Inconceivable
Worse yet when you only want black ink, but it refuses until you replace the colour cartridge...
136 publicly visible posts • joined 6 Nov 2019
I agree that sounds wrong - A few guesses for what maybe it was meant to mean...
Firstly, while the net force of gravity is the same, it's more spread out, so maybe some shear forces come into play?
No crush, but maybe there's a shred.
Alternatively, if a planet's orbit is within the new bounds of the stellar atmosphere, collisions will start to slow the orbit of the planet and it will spiral in.
It's already well roasted before that, and I'd guess torn apart as it gets absorbed by the star - but maybe whatever isn't torn to plasma could descend and be crushed (I'd guess more by density than gravity)
Correction - they're not doing anything a huge number of humans with large data farms and no morals (and no chance of being caught) couldn't have done
Because of the scale, this is more equivalent to a rogue nation state than what a single human could feasibly do
"Buy our bread, it doesn't have sawdust in it!"
While presumably true, in some countries that can get you in trouble since you're implying your competitors do put sawdust in their bread - why else would you say it?
And PR folks (including politicians) have been saying things that aren't strictly false without giving you an honest view of matters for years
That would present guarantees about it being tested... But that's also a little unfair.
The "move fast and break things" crowd is relatively new. Plus going back in time is the opposite of fast.
OSes have had clocks that assume time is monotonic (i.e. can only go forward) for a long time.
A negative leap second would break that assumption, allowing us to finish a calculation before it's started (with appropriately perverse definitions and consequences).
You could put caps on it to limit it to reasonable levels - if you as an individual are spending $10k on your case, and losing means you pay another $10k? That might work.
Won't cover the legal fees of a large company you sue, but if you're suing another normal citizen (not a large corp or a millionaire) that's some reassurance to them that they can afford to tell you to stick it
> but spaces now make more sense "because it always works and it's always consistent."
If you think spaces for indentation is consistent, that means one of two things:
1) you're not comparing against tabs for indentation, you're mixing tabs and spaces (which is *always* stupid - you get the downsides of both), or
2) you think your indent size is appropriate for everyone else's screen, font size, eyes, etc...
Tabs are more consistent - one tab is one indent level.
Not the 1 vs 2 vs 3 vs 4 vs 8 spaces per indent level some people use (I've seen all of those).
I use 4 spaces because that's most common, so it's what my editor is set to.
Consistency and standardisation is one of the most important things.
Except if I have to edit a Makefile, of course
Maybe, yes, and no. In that order.
First the yes:
Part of the justification was that they had proper role-based access, people could only access the information they were meant to, so they have obviously heard of minimum privilege.
And applied it well enough that inspectors accepted they had taken reasonable steps.
Second, the maybe:
I'm not sure of the CRM in question, but a quick googling suggests Qantas' main CRM is from Salesforce. I'll assume that it's the one in question here, but that may be wrong.
Salesforce is a big enough player that they have some trust.
So if the ability to exfiltrate was due to a bug on Salesforce's side - liability should lie with them.
If it's a configuration issue, then maybe Qantas, maybe Salesforce. Depending on defaults and documentation. Unless they had Salesforce consultants set it up for them, in which case blame is still on Salesforce' side.
It's reasonable to assume that Salesforce should have a secure system, just like you need to trust Microsoft, AWS, Cloudstrike, Visa, whoever else.
Yes, you still need to do risk analysis on it. But there comes a point where you have to trust someone - you can't have every company in the world auditing the source for every bit of software they use, it's just not viable.
Finally, the no:
The person giving the verdict may well be a paper pusher, relying on others to delve into technical details, but they've applied the law as it stands and done a thorough analysis. Read the actual report if you haven't yet, it's linked in the article.
Now, do you disagree with the findings? That's reasonable if you can make a solid argument.
Should we punish anyone with a data breach even if they did everything right on their side - or at least did everything that the laws require and government recommendations on cybersecurty say (essential eight for now, although that's being reviewed)? That won't work well in the long term.
Unfortunately, security while maintaining functionality is hard.
There aren't really a large number of options for CRMs for big players - who has the time and technical capability to audit it all?
Qantas is big enough that they should be able to audit stuff, but also big enough that the amount of functionality they want is large (and hence there's more to audit)
It sounds like Qantas have complied with the legal obligations, and probably have a reasonably well locked down system and training program (people will always fail eventually)
It sounds like the attacker knew how to walk someone through a few steps in Qantas' CRM that allowed external access... The CRM probably shouldn't have allowed that, but the phishing using their own CRM means the employee probably didn't do anything "wrong"
As for moving it from a SaaS tool to internal? That's no quick task for a major system - 12 months isn't enough for planning or testing (let alone development - assuming you can hire enough developers and testers and security experts) if you want it done right; your suggestion would almost certainly reduce security
There are issues with Qantas for sure, but in this case they seem to have followed best practices as far as they reasonably can, and better than most companies (how many can say "we schedule deletion of data we no longer have permission to hold, and none of that data remained in our system at the time"?)
If they don't like that, there are things many of them will dislike even more!
In one of Deviant Ollam's stories of pentesting, one of the times he got rumbled after hours was because one of the workers saw his car on the weekend (driving past, not at work) and went to check it out.
The workers all knew each other (so claiming to be from another site didn't work) and cared about the company - because they'd been there years/decades, because there was a good culture and management looked after staff
> Where UI designers, in order to justify their employment, change the UI in every edition...
Oh, if only it were that simple, we might fix it!
No, you missed a bit - where not only was there research on UIs, but there was also training for UI designers. Plus UI was its own field (as opposed to getting the front-end developer to do it)
As for UI changes as justification - that's at a company level, not just individuals - how will the customer know we spent heaps of time building this new version that they should spend more money on, unless it looks different?
They won't notice half as many crashes, but they will notice a ribbon menu!
> > how do you not also accidentally hit the wrong tab close button?
> I do. Hundreds of times. This is why I prefer keyboard control to mouse, and why I keep banging on about standards-based keyboard UI.
Firefox has keyboard controls - new tab, select address bar, paste, back in history, close tab, cycle to next/prev tab - can all be done without mouse - so if you struggle to use a mouse, why keep doing it?
Leave it to the side and just use the keyboard, if that's better for you
There's nothing you need to use a mouse for around the tab bar area if you don't want to, or are unable to use a mouse with precision
Touch interface has inherent imprecision, and needs more spacing, but that's a different kettle of fish
> > The audio playing/muted/waiting to start icons don't take much space
> Which is a problem, because they leave less space on the already very small click target of a tab.
It's not a small target for me, but maybe that's usage because I don't have a lot of tabs open most of the time (I'll bookmark if I want to save for later, or use another window if I want to run through a list of things)
But even then, minimum tab size if I shrink the window right now is favicon, a couple of letters from title, and a close tab button
If I want to select, mute, or close - those are all about the same size clickable areas
What do you actually want? tabs and buttons so large that you can't misclick them? If you're struggling with that, it's not a tiny audio icon that's the problem
> > (no UI clutter on a desktop screen) - and they're more visible that knowing a magical phrase for the address bar.
> Controls that only appear when they are needed, i.e. when there's something for them to do, are better for skilled users, and that is the core market for Firefox: it is the web browser for power users.
An icon that shows "this is the tab that's playing audio" is sensible - tabs playing audio should be a rare event (rare that you'd want more than 1 tab, very rare you'd want more than 2)
And since it only shows when there's audio playing, and you can click that to mute - that's the *definition* of only appearing when needed and there's something to do!
More broadly, UI should be intuitive and discoverable.
Audio plays? an icon shows up. Click the icon? audio is muted. Easily discoverable, and I'd say intuitive (or at least a common pattern - as I said, KDE's menu bar does exactly the same).
Typing something other than an address or search term into the address bar? Not very discoverable unless you know already. Not intuitive unless you're already hooked on siri/chatgpt/some other nonsense
And I say that as someone who learnt to use the kio protocol addresses in konqueror and other KDE apps - that's a power user feature, muting audio should not be!
> Yes, I know. I regularly hit it by mistake and wonder why the radio went quiet. I could happily lose that button.
If you clicked and the radio turned off, it should be pretty obvious, so you shouldn't be wondering long - and if you're having trouble accidentally clicking a tab mute button, how do you not also accidentally hit the wrong tab close button?
Also, how often do you have multiple tabs playing audio?
For me that's mostly if I opened a few youtube links and visited those tabs (as until you open that tab it pauses waiting) - if I want to stop audio in something, I usually want to go to that tab to pause the video as well.
I'll acknowledge that sometimes ads play sound in random tabs if you don't have good blocking, but other than that, is it common?
The audio playing/muted/waiting to start icons don't take much space (no UI clutter on a desktop screen) - and they're more visible that knowing a magical phrase for the address bar.
And if I've muted a tab, I usually want to know which one it is - either because it's something I want to restart later, or it was unwanted and something is misbehaving on that site.
Not identifying which tab/site it was isn't necessarily a feature
Sure, maybe a mute-all is useful sometimes, but it's something I need less once a year (KDE's task bar has playing/mute icons for each window - so I've used something equivalent but faster than typing in the address bar).
> use Wayland and it works just fine
wayland nowadays mostly works pretty well, but it's a bit short of "just fine"
As far as I'm aware there's no way to dock thunderbird into the systray (in x11, you could use kdocker), and I think we're still waiting for ability to script sending input to a window with any sort of standard tooling (i.e. don't need to download from non-distro sites). But I haven't checked for a little while, so maybe they finally addressed those?
I realise there were some security minded reasons for why they locked some stuff down - but anything that was sending input to other windows could presumably also be executing arbitrary other commands so it's not much help
I'm sure there are other issues as well, those are my pain points that come to mind (until very recently, I was also getting memory leaks which seem to happen a while after suspend/resume)
> the algorithms themselves are inherently non-deterministic.
> or, more realistically in practice, use them to seed a good (deterministic) PRNG
The second time you got it correct - algorithms are deterministic.
If you know which algorithm it is, what seed was used, and how many samples have already been taken, you can predict the next sample because it's deterministic (unless there're additional inputs being used at generation time - in which case you'd need to know them as well)
> Also, he appears to be going a bit soft, with age. Where is the volcanic outrage ?
Linus' rants are mostly limited to people who do the same stupid thing over and over and ignore being told no.
His rants have gotten headlines often, but when you read past the headline, it's usually something like "on date, you were told this was wrong because of reason X. Then on date you tried again, and were again told that it was wrong because of reason X. Now you're trying a third time to do something that's wrong even though it's been explained multiple times? what's wrong with you?"
You generally need to have a track record that Linus himself is aware of - and imagine how many people he interacts with (occasionally it's a rant about how stupid the design from a certain company is - that's a different kettle of fish)
No, Linus is correct
If it's a security bug found by generally available LLMs - that can be assumed to be public knowledge
Hence there's no need for a private list - it's presumed public, so public lists are fine
You could /maybe/ have a generic LLM bug report list, but there's not really much benefit - just ensure anyone submitting the bug report properly documents and checks it's a new report and it's all good
There is one exception - not generally accessible LLMs. New models that do different security analysis, etc., before they're public
But those are uncommon and run by people who can coordinate a unified set of reports that don't result in overlap - they can be handled like any other security specialist company