Posts by Norman Nescio

It turns out that even hard-wiring and writing a notice don't work:

BBC: Major research lost after cleaner turns off fridge, lawsuit says

As I said, cleaners have different incentives to day-workers, and (lack of) training is often given as a cause. That's a cop out.

As accident investigators will tell you, relying on human operated procedures means that things will go wrong, Railway signalling systems were automated to remove fallible humans from the loop as much as possible; so if you want a process to work more reliably than people, you need to remove people from the process. Some do this by banning access to cleaners and letting staff qualified to use the technical area (and understand the working environment) also do the cleaning of their working environment, others by using hard-wiring, enclosures, and safety- and security- interlocks to prevent mishaps from occurring.

I would ask why the cleaner had access to the circuit breaker, and did Prof. Lakshmi formally and directly inform the cleaning company that cleaning was not required? I don't count an informal notice on a fridge - there is no guarantee that a cleaner is literate.

I have not, but I read an interesting case (which I can't find on the Risks Digest) of a machine room on the umpteenth floor of a tower block being flooded.

Water mains pressure is insufficient to feed the higher floors of tall tower blocks. As a result, they have holding tanks, often at several different heights with pumps to move water up to the highest level needed.

The company affected had a computer/machine room fairly high up, and flooding was not regarded as being a potential problem, until the plumbing leaked catastrophically and flooded the computer/machine room on the umpteenth floor from above. Of course, the holding tank will have a float switch, so the pumps will merrily keep pumping water up, which then cascades down. Raising the floor multiple stories above street level didn't help.

Speaking of pumps, Peer 1 Hosting of Manhattan had to resort to employees moving fuel up 17 flights of stairs to keep their data-centre running after Hurricane Sandy. The pumps to move fuel from the basement tank were submerged in flood water.

Top cunning plan, only foiled by the fact that the cleaner only realises their equipment won't plug into the server socket once they've removed the server plug to check. It's absolutely standard that people do an extremely cursory search for an unoccupied socket, shrug, and pull out a plug from an occupied socket. If you are lucky, they put it back in afterwards. If you are lucky. I've also seen cleaners removing plugs from sockets by pulling hard on the lead/cord - which shouldn't work for UK plugs, but does*. It saves them traipsing back over the bit they've just cleaned to remove the plug from the socket, so very efficient time-and-motion-wise.

The only system that is pretty immune to that is central vacuum systems, where the hose connects to a faceplate in the wall. Doesn't work for floor-polishers 'though. The other possibility is to hard-wire everything that shouldn't be unplugged.

*I suspect this lead to the interesting experience of finding a cleaner using some equipment with no plug, simply bare stripped wires. Said cleaner would unplug a piece of equipment, place the bare wires into the live and neutral of the socket, and replace the plug. It meant the cleaner could get their job done, which was what was important to the cleaner, who had an entirely different set of incentives to the day-workers. Ingenious, but not entirely within regulations.

Well, I'm just an old buffer, and I'm late to the party, so I'd better wax lyrical to bring a shine to the proceedings.

Re: All that glitters ...

Pull the floor tiles a few at a time

Ah yes. Sage advice.

At one of the places that employed me, they had a (quite serious) workplace accident.

A new cable needed to be laid in a machine room with a raised floor. So the person doing the work removed a line of floortiles two or three metres long along the front of a line of racks.

The racks tilted over and collapsed on the person, causing life-changing injuries. I saw the pictures of the aftermath, without the blood.

People were most assiduous in removing only two adjacent tiles at a time after that.

Re: Alternatives for sanity check

My thoughts on 'sanity check' are similar. In my experience they are 'not-obviously-wrong' or 'not-obviously-inconsistent' checks. They check for a minimally necessary condition, but not a sufficient condition. They are often an efficient way of determining if some presented data are in a fit state to be submitted for further (possibly expensive) processing e.g. that all the required fields in a form have some content, before going on to process the content. Or, running the Luhn algorithm to check that a submitted number is plausibly a credit cad number that can be sent to a payment processor for further validation and processing.

Check 1 - has the submitter put something in the field marked credit card number?

Check 2 - are the contents purely numeric?

Check 3 - do the submitted numbers pass the Luhn check?

All of which can be performed before sending the sanitised data onwards for processing.

Re: Abort, Retry......... Fail

Well I formed a committee of learned folks (at least 3 PhDs in there) who have pronounced that folks named Graham are racist, sexist, transphobic wankers.

They are also crackers.

Re: abort

Marx Brothers, A Night at the Opera: "Everybody knows there ain't no Sanity Clause".

Re: And by "solving" a non-problem ...

The Novel "1984" by George Orwell had as a central theme the emasculation of the English language in order to prevent people being able to express concepts, and as a final aim, to prevent them being able to think independently at all. This committee would appear to have similar aims, if less in magnitude.

That's the strong Sapir-Whorf hypothesis of linguistic relativity (aka linguistic determinism) which states "language determines thought and that linguistic categories limit and determine cognitive categories". As Wikipedia goes on to point out "This version is generally agreed to be false by modern linguists.".

You can think about things without having a particular vocabulary. In fact, it is additionally quite possible to generate a vocabulary de novo (like quark, but not quiz), so banning particular words certainly does not de-reify the concept - as can be seen in the mutation of terms of abuse adopting a previously sanitised or medical term.

I'm happy to use different terms if inoffensive technically accurate ones can be found or formulated. It would be churlish not to.

Re: One thing you've missed

Assuring data consistency in a snapshot is harder than it looks. It depends on your application programmers 'doing the right/write thing'.

StackExchange: Is it overkill to shutdown a VM before taking an LVM snapshot for backup purposes?

But what if an application or series of services is in the middle of an operation that consists of multiple independent transactions? For example, a user is registering. The database is updated but the registration e-mail has not been sent out yet or something like that. Taking a snapshot at this point would not reflect a correct/complete system state.

Yes, programmers should assure consistency for a whole transaction (which can consist of many updates). In real life, that doesn't always happen. Snapshots are great, but they can't protect you from non-atomic transactions.

Re: Capital idea comrades!

Service grants citizenship!

Oddly enough, that is pretty much the philosophy described by Robert Heinlein in his novel Starship Troopers, where Federal (read military) service granted the right to vote. Perhaps Putin is using the novel as a textbook?

Possible issues with MS curl to be aware of

I don't know about tar, but the Windows 10 curl is a bit complicated.

curl.se: curl shipped by Microsoft

1) "The curl tool shipped with Windows is built by and handled by Microsoft. It is a separate build that will have different features and capabilities enabled and disabled compared to the Windows builds offered by the curl project." ... "If you have problems with their curl version, report that to them. "

2) " The curl tool comes installed in addition to the dreaded curl alias that plagues Powershell users since it is an alias that runs the invoke-webrequest command and therefore isn't acting much like curl at all. A work-around is to invoke curl as "curl.exe" to prevent powershell from treating it as an alias. "

(1) Is just like any distribution repackaging curl - you should report bugs to the distribution first, so it's not necessarily a Microsoft/Windows specific thing, but worth taking note of. (2) has enraged people for quite some time.

Re: but why?

Look, everyone just uses variations on the .norm normal file format. Just go with the flow...

Compression

Dial-up modems developed the possibility of compression as well. I fiddled about with MNP 5 and V.42bis in my time. The problem was that they obviously did not work if the data you were transferring had already been compressed (like zip files) - a point my less technical colleagues failed to appreciate: "but it says 4:1 compression...".

Re: Advantage of 7-Zip

I don't see how this helps. It just kicks the can down the road. Hiding the filenames is security by obscurity. What happens when the recipient doesn't know what filenames to expect?

If you encrypt a zip archive of a collection of files, the filenames are in cleartext, even though the contents are in encrypted text. Someone can replace one of those encrypted files with a different file of the same name, without needing to know the encryption password. The example given in the StackOverflow answer shows this. This individual file in the updated zip archive does not have encrypted contents, but when you decrypt the archive, the zip decryption utility does not tell you this. So what might have been, for example, an encrypted .dll is replaced with a non-encrypted .dll which might do something completely different.

If you take an archive of encrypted files, then zip+encrypt it again, you end up with a new archive with a single file in it - the encrypted archive. So an attacker cannot replace a single file - they would have to replace the entire archive with a new (non encrypted) one, but without knowing the underlying file names within the encrypted archive.

So, in one sense, you are right: if the recipient has no idea of the expected filenames, we are no further forward.

If, however, you 'double encrypt', and say to the recipient on a different communications channel to the one in which you are sending the archive: "I'm sending you an encrypted archive. If you unpack it with the password 'Horse-Battery-Staple' you will find an archive that can be decrypted with the same password with a file within it named 'Squeamish-Ossifrage'. If Squeamish-Ossifrage is not there, do NOT use the archive."; then you have a chance of preventing an attack. Assuming encryption is working adequately, an attacker will not know the filename 'Squeamish-Ossifrage' and cannot replace it indetectably.

Obviously this is not ideal, and open to all sorts of social engineering to get people to accept unexpected files, but if zip is the only tool you have, it is a possible workaround. I've had to use it when communicating with people with locked-down corporate PCs that had no other (easy) possibilities than using WinZip.

NN

Re: "it has earmarked “significant investment” for FY24 towards customer experience and branding"

I agree. It's always a bad sign when a company in trouble decides to spend money on marketing rather than improving the product.

Classic signs a company is in trouble:

1) The HQ is refurbished

2) The C-suite get new, expensive limousines

3) Re-branding to 'freshen up'.

4) Company flags go up outside office locations

Essentially, when the C-suite start valuing image over substance, the rot has set in.

Re: Paper doesn't crash.

GDPR applies to paper records. Covered by Chapter 1, Article 2 'Material Scope' section 1.

Article 2

Material scope

1. This Regulation applies to the processing of personal data wholly or partly by automated means and to the processing other than by automated means of personal data which form part of a filing system or are intended to form part of a filing system.

Legal text on EU website

So a set of handwritten records in a filing cabinet are subject to the GDPR. Even if the light is not working, and you have a sign reading "Beware of the Leopard" on the door to the disused cellar toilet you keep it in.

Article 4 'Definitions' section 6 tells us what a filing system is:

‘filing system’ means any structured set of personal data which are accessible according to specific criteria, whether centralised, decentralised or dispersed on a functional or geographical basis;

If you do an Internet search on the question "Does the GDPR apply to paper records of personal data?" , you'll get more background. The simple answer is "Yes".

Re: Apologies in advance for the super-pedantry here.........

Sapphire is not an element. That always triggered me. And neither for that matter was t'other one.

That said Joanna was as lovely as ever. Both actors deserved far better than that show.