Posts by Vic

Re: Its the wrong way to get off this planet

The obvious way to get off the planet is to go to the edge and jump

What - climb over the Circumfence?

That way lies madness[1] ...

Vic.

[1] And turtles.

OK, so I'm going back on my promise to not write any more

Well then, you're going to have to start doing your own research. This is the last time I'm going to spoon-feed you.

For your argument to work you'd have to explain why subspacing preferentially removes more colliding inputs, proportionately speaking.

We don't care about "proportionally", we care about "absolutely".

For a given hash & block size, there are a finite number of blocks that will cause collisions in a given hash. By removing some of that finite set, we have fewer potentials for collision. It is that simple.

It is clear that collisions are a problem in the general case - we know the number of collisions for any particular block by way of the Counting Argument. The only way this cannot be improved by llimiting the input set to textual data is if the hash were designed to maintain the same bit spectrum for similar input data. This would make it a very poor hash. And so, the number of possible colliding blocks is necessarily reduced by limiting the inpuit dataset. LIke I said.

And from this point on, you're on your own.

Vic.

I think that a well-designed hash won't have significantly different collision rates over the two kinds of data for a given capacity range (ie, the number of messages to be stored).

But we already know that the hashes used in git were not specifically designed for text data; thus we can expect that the data blocks that will cause a hash collision with our desired block will be spread randomly throughout the possible input space. If we now dramatically subset that input space (by requiring our input to be text data), we have ipso facto discarded all those possible collisions form data blocks that were not text data - thus we have a much reduced number of possible collisions, and thus a much reduced probability of collision.

Within the filesystem view, we cannot do that input selection; any possible block variant is permitted. Thus we cannot achieve that probability reduction, and the process is much less safe.

Mathematically speaking, you're absolutely right.

I know.

All I'm saying is that in practice assuming that the same hash implies the same contents can be a reasonable engineering assumption

And I'm saying that is not a reasonable assumption in the general case. I have outlined my reasoning above.

you can plug the numbers into some formulas to find the expected collision rate and choose your digest size to make the risk "low enough" that it's not worth worrying about

For the general case without subsequent block-checking? The hash is required to be of at least the same size as the block. And that is not exceptionally useful...

Vic.

By your logic, git having less entropy in the source should imply less entropy in the output hashes, which would in turn imply more collisions.

No, not at all.

The set of possible inputs which could generate a given hash is much reduced by the requirement that the inpout in question is source code (with a limited character set). The same cannot be said for the general case of a file.

this risk is "vanishingly small enough"

And I'm, saying that, in the general case, this is simply not true. Only if you constrain your input set - such as is the case with git - can you make it so.

Vic.

The git tool uses large digests (sha-1) to identify all commits

git is working on source code; there is vastly reduced entropy per character in the input data.

This is not the case for random files on a filesystem.

Vic.

it is reasonable to assume that when two blocks have the same checksum, they are in fact the same block.

That is not reasonable.

An m-bit hash may take 2^m different forms. An n-bit block may take 2^n different forms. So if n > m, there are 2^(n-m) different blocks that would generate the same hash. Nothe that n<=m is an entirely useless operation, so we have this risk.

This is the Counting Argument. It's used to disprove "magic" lossless compression algorithms that will store anything in nothing...

Vic.

security experts and mathematicians in the various intelligence services and academia, many of them amongst the best in the world

That's very unlikely.

I just went to the GCHQ careers site. THe first position I clicked on was for an Intelligence Analyst - quite an important role[1]. they're offering a bit ove £25K.

Applied Research offers less than £28K.

At these rates, they might get a few gung-ho patriots, but the real talent will be earning ten times as much. It is likely that the "best in the world" is nowhere near Cheltenham,

Vic.

[1] The job description states "Analysts are at the heart of GCHQ’s mission - turning data into the critical intelligence that helps to protect the UK. Our analysts work with complex data, understanding and interpreting it to find the crucial intelligence within it."

Re: Misses the point

what fucking 'tard came up with that?

Someone who got what he wanted...

Vic.

Re: Bottom line is ...

This is the 24th November of 2015 and I gave jake and upvote.

Scary, isn't it?

Vic.

Re: Well if Dell says so . . .

Did you see how I used 'critical thinking' to slam you?

Where?

Vic.

Re: Odd

Oh, that's right! We build our own builds and install them, rather than trusting a vendor to not muck things up.

That's still vulnerable to WPBT...

Vic.

Re: Feeling SO fine

Thankfully it is, nor never will be, a Linux feature.

Ten years ago, no-one would even think that possible.

These days? I wouldn't bet my house on it. Miguel will find a way to incorporate it somehow...

Vic.

Re: Feeling SO fine

I think it is time we stopped trusting OEM installs.

Even that might not be enough.

If this is being shipped in the WPBT - as the Lenovo crapware was - even a fresh install is going to get pwned on boot.

Ooops. That was a good idea, wan't it?

Vic.

Re: Nothing to see here

and anyway, you have nothing to hide.

Well, not any more you don't.

Vic.

Re: SOP when buying new laptop (with Windows, obviously)

Uninstall all vendor software

Wouldn't work in this case - as seen in the previous article, once you remove this root CA, a simple reboot and it comes back.

I hope this hurts Dell significantly. This sort of thing must be stamped out.

Vic.

Re: Probably a necessity

If you wrote all your scripts in 5 mins you haven't written many or very simple ones.

That's sorta the point - init scripts are simple. If you're trying to do complexity in init scripts, I would suggest you're putting the work into the wrong thing...

I'm scratching my head as to why people insist it is the worst thing ever

Because it keeps putting its tendrils into things it shouldn't. If all it did was to replace the SysV init system, no-one would really care too much - although I think systemd is clunky, I would get used to it. But when Poettering keeps insisting everyone else change their ways of working to allow him to do what he wants, you'll get push-back. And that is what is happening.

Vic.

Re: Probably a necessity

To use VIM in compatible mode one has to start it and then in command mode type:

"set compatible"

Or you can give it the -C switch on the command line. Not that I would.

Vic.

Re: Probably a necessity

Systemd while not perfect allows one to write a service init script or custom action in minutes rather than hours.

I have never spent more than a couple of minutes writing init scripts. From what you're telling me, systemd solves a problem that doesn't exist...

Vic.

Re: Fuck systemd

I will keep your post in my reference list for all those who incessantly spout off about how "easy" Linux is.

I wouldn't...

I use vi on a daily basis[1] on many distributions[2]. And the only "gotcha" of which I'm aware is the fact that some terminal emulators do weird things with the arrow keys, meaning you have to use the h,j,k,l keys as per the manual.

vim is pretty much interchangeable for vi.

Vic.

[1] vi is my editor of choice; I learnt it many years ago, and I like it. I'm not trying to convert anyone else to my line of thinking.

[2] I even used to use it on Windows. My boss - a big UltraEdit fan - despaired.

Re: "News"

You said it's ethically improper for a lawyer to support a not guilty plea of a guilty client if the lawyer knows the client is guilty. Am I correct that that is what you are saying? If so, you are wrong

He's not. It's ethically incorrect deliberately to seek the wrong outcome. Lawyers are Officers of the Court, with a duty to see fair play.

Lawyers have guilty clients all the time that plead not guilty.

Indeed they do - but if the lawyer knows that the plea is a sham, he has a duty to inform the court.

From the American Bar Association's Rules of Profession Conduct Rule 3.3(b):

A lawyer who represents a client in an adjudicative proceeding and who knows that a person intends to engage, is engaging or has engaged in criminal or fraudulent conduct related to the proceeding shall take reasonable remedial measures, including, if necessary, disclosure to the tribunal.

Note that rule 3.3(c) backs this up:

The duties stated in paragraphs (a) and (b) continue to the conclusion of the proceeding, and apply even if compliance requires disclosure of information otherwise protected by Rule 1.6.

So, in the event that a suspect pleads not guilty and the lawyer knows that plea to be false, he has a professional duty to disclose that knowledge to the court - even if he only finds out about it mid-way through the trial.

Of course, we're talking about lawyers. A significant proportion of them - not all, but way too many - don't actually care about ethics; all they want is results, at any cost.But the preamble to above Rules has the following to say:

[19] Failure to comply with an obligation or prohibition imposed by a Rule is a basis for invoking the disciplinary process

It is quite clear that a lawyer knowingly supporting a false "not guilty" plea is subject to the Disciplinary Process. That this does not happen regularly speaks volumes...

And what does it matter what they know from the client? It's privileged communication, and the lawyer is safe.

Not so; see the above quote. Rule 1.6 is the rule relating to confidentiality - rules 3.3(b) and 3.3(c) clearly state that such confidentiality should be breached if the lawyer knows that the suspect is guilty.

Vic.

Re: Dear Hillary Clinton, and Clipper chip

Doing it in software, you can't keep the algorithm secret

If there's sufficient value in discovering the secrets, hardware isn't going to keep it secret either.

This is the trouble with all these "eggs in one basket" scenarios; they turn the crypto mechanism into a very tempting target for criminals. Once that target is worth enough, someone will put the cash into breaking into it...

Vic.

Re: Restrict guns, not encryption

Ban the manufacture and supply of ammunition

Chris Rock said something similar - NSFW, of course, because it's Chris Rock...

Vic.

Re: Virtual currency? How about "follow the money" instead?

they never do anything "quickly" (for some value of quickly) so this may never come to pass.

Many of our Dear Leaders have been trying to get spy data flowing throughout Europe; I suspect this attack will be the stroke of luck they've been waiting for...

I note that the article talks about "unintended consequences". I believe this to be an empty set.

Vic.

Perhaps, but wouldn't this also work if the traffic lights were set to take into account a faster speed?

No.

Traffic density falls off rapidly with speed - so if you tune for a higher speed, you get fewer cars getting through the lights in a given time period. That means more people sat stationary at red lights, so more traffic problems...

Again, it is all green/safety idiots causing frustration to the majority that don't buy into it.

No, it's engineers who have studied the dynamics of traffic flow and are trying to minimise commuters' aggravation.

Its supposed to be a democracy, not whichever group shouts loudest gets its way system.

And that is exactly why no-one is changing anything just because you've decided to shout loudly.

Vic.

If it has failed an MOT you are still permitted a single trip to a point where it can be repaired.

It's not quite that simple; from the gov.uk page,

You must not drive the vehicle on the road if it fails the test, even if the MOT hasn’t run out, except to:

• have the failed defects fixed
• a pre-arranged MOT test appointment

You can be fined up to £2,500, be banned from driving and get 3 penalty points for driving a vehicle in a dangerous condition.

[We'll ingore the horrific grammar in that]

The upshot is that, although you might not be committing an offence by driving a car without an MOT, that doesn't mean you can drive anything; some cars are not in an appropriate condition to put on the road, even if you are just going somewhere to repair them...

Vic,

Re: @ Matt Bryant

I don't want to learn that some things are caught. If you're invading everyone's life then I want all things to be caught.

I think I'd settle for "substantially all".

What we have so far is "a handful of cases that we're not going to tell you about in any detail". And that's not enough.

Vic.

Re: Hands up if...

Because I accidentally triggered a remote shutdown on a live server.

It was 3am and I was in the wrong terminal window...

This is, IMO, the one practical use of command blacklisting in the sudoers file.

I had a group of users with a group of servers. Against my advice, most of them had sudoer privilege. And we had regular shutdown accidents...

I blacklisted the shutdown and reboot commands - that won't stop the users being able to perform the fundtions, of course, but it does stop the doing it accidentally...

Vic.