* Posts by ibmalone

1414 publicly visible posts • joined 6 Jul 2017

New lawsuit: Why do Android phones mysteriously exchange 260MB a month with Google via cellular data when they're not even in use?

ibmalone

You can reorder them and change what's available in the quick drop-down.

ibmalone

Well, yes, that's an issue regardless, there being cheaper data available doesn't change that fact. But also, how are people paying that much for data in 2020? That's definitely an issue too.

(Also, I'd always taken it as a given that things would be tryin to use data if not switched off. If it isn't google play itself it'll be some app that's installed and poorly written, or something trying to access network when you just wanted to open it to access saved data.)

Edit: also 99% of users should know how to turn data off on their phone, it's one of the most obvious controls in the default UI and the fastest way to usage monitoring, which, if you are paying $8/GB, you probably want to keep an eye on regardless. Again, if google is exfiltrating data and people are paying for it, that's definitely wrong, but if this amounted to service level things like connectivity pinging then it'd be a bit like complaining to philips that their lightbulbs are stealing electricity from you.

ibmalone

1. "$8 per GB of data in the US" Wow. Just. Wow.

2. And people think it's odd that I turn all data services off when I'm not actually using them?

[Checks meeting agenda...] Where does it say 'Talk cr*p and waste everyone's time'?

ibmalone

I've got a bizarre ongoing issue with both Teams and Zoom where I (apparently) sometimes sound like a chipmunk. This was funny the first couple of times in April, not so much now. Almost certainly sample rate related, but very hard to diagnose as one-to-one or echo calls are usually fine, suspect it's partly caused by other people's setups (reliably happens in meetings where certain people are involved and not in others). Switching microphones (from webcam to soundcard) usually fixes it. Don't know if it's pulseaudio or buggy software (the soundcard is envy24 based, wouldn't be too surprised if something is getting set up for 96kHz and then being surprised to get half that).

Microsoft warns against SMS, voice calls for multi-factor authentication: Try something that can't be SIM swapped

ibmalone

Re: OMG !

What baffles me is my bank used to have a card-reader based 2FA and then changed it for SMS. Was pretty astonished to learn how common SIM swapping is, there's apparently a pretty thriving industry.

Somebody's Russian to meddle with UK coronavirus vaccine efforts, but GCHQ won't take it lying down

ibmalone

Re: Still confused about the 90%

20,000 patients in hospital at peak, over 3000 on mechanical ventilation. I'll not spend the evening hunting for workload stats, these are probably hard to process, but BMA covid tracker survey through April makes interesting reading, 25% of doctors redeployed (30th April survey), to the question "During this pandemic, do you consider that you are currently suffering from any of depression, anxiety, stress, burnout*, emotional distress or other mental health condition relating to or made worse by your work?" nearly 30% answered "Yes - and worse during this pandemic than before". https://www.bma.org.uk/advice-and-support/covid-19/what-the-bma-is-doing/covid-19-bma-actions-and-policy/covid-19-analysing-the-impact-of-coronavirus-on-doctors

Possibly being in the relatively quiet environment of a non-covid ward skews the perspective.

My local GP surgery was closed - and that was not because it was overwhelmed. The doctors and staff were at home no working.

Not sure if mine were at home or not, what I do know is they were doing remote consultations.

ibmalone

Re: Still confused about the 90%

Can't find any direct report of rAd5-S or rAd26-S challenge trials. Ultimately it's down to ethical approval, not outright illegal to do (lack of effective treatments is an issue). Not sure if the Russians would be keeping quiet about it if they'd already proved it to work (their current roll-out is actually just a phase 3).

Additionally all medical evidence is statistical in nature, science generally is anyway, but in medicine things tend to be noisy and small numbers. Even if the Russian vaccines achieve a 90% efficacy (which is quite high for vaccines I'm told), one in ten of the treatment group will get an infection. Not all the control group might. Got to separate the two somehow. (You've still got a control group, otherwise how do you know the protocol worked? Is it as big as the treatment group? Maybe it's smaller because you don't want to give too many people a severe coronavirus challenge, but that makes noise more of a problem.)

ibmalone

Re: Still confused about the 90%

I likely wouldn't conclude without reading the study. But that's just me. I am hard to convince.

Actually I agree, my own area is plagued with press release before publication which is incredibly frustrating. Interesting to see what we can infer from this limited data though, thankfully more will be coming.

B-but I thought Evil Micro$oft Billy killed all of Africa with botched super malaria vaccines!!!???

It's easy to forget these things in 2020.

ibmalone

Re: Still confused about the 90%

There are hotspots of activity, some people live farther out than others, some have different travel routines, some work from home while others go to an office, etc. I think this is more what I was trying to say, and my paint metaphor may have been misleading.

I understand what you're saying, and can see why it might seem like that, but the randomisation also means those factors are spread evenly.

This is why I keep comparing it to a coin toss, because, sure it could happen that every single person in the treatment group lives in the countryside and works from home, and everyone in the placebo group works in an ICU, but it's exceptionally unlikely, you calculate the probabilities by multiplying and at first it starts out plausible, 50:50, then 25:75, then 13:87, 06:94, even up to the first ten it's about a thousand to one for that kind of imbalance and it's not utterly implausible. But it's exponential, so by the time you've got over eighty in one group and under ten in the other it's really unlikely it can happen by chance.

(For purists, I've skated over the combinational aspect if you allow a few of the placebo group to be living in the country, but as we're at the tails of the distribution it doesn't make much difference.)

ibmalone
Pirate

Re: Last post from me

Alas, poor AC! I knew him, Horatio

(closest icon to a skull...)

ibmalone

Re: Still confused about the 90%

To put your mind at rest then, in a randomised trial each participant is randomly assigned to the treatment or placebo arm by an independent mechanism (there are services set up to do it, generally they provide a code which gets used at the pharmacy). The point of this exercise is that there is nothing about the patients that influences which arm they get into, so the 94 subjects being split 9:86 (the lowest guess compatible with that 90% number) really is analogous to doing 94 coin flips and getting only 9 heads. The only thing that makes the overall trial size relevant is that on that scale randomisation is really good, rather than just the best we can do. The paint analogy would be spray painting an area and finding the paint mysteriously not sticking to half of it.

I realise 94 seems small, but if this was a cancer treatment and I told you in a balanced study 9 people on treatment had died and 85 on placebo had died, what would you conclude?

With all the misinformation and conspiracy theories surrounding the virus, wouldn't you want more concrete results too?

I don't think clinical evidence will influence misinformation or conspiracy theories much either way. In any case they tend to focus on the safety, if not absolute fantasy. The people who believe Bill Gates is planning to inject us all with microchips wouldn't be persuaded even if this was 160 infections to 0.

ibmalone

Re: Still confused about the 90%

Let's be clear to start with that people dying of cancer and other conditions because they couldn't get to hospital would have been in that situation anyway, hospitals (especially staff) were overloaded and had the additional burden of trying to deal with infection control while having staff self isolate too. This is not an effect of lockdown, this is a direct effect of the coronavirus burden on the health system. Even now we are having to clean scanners between patients which reduces the number that can get through. The excess deaths data is pretty persuasive because it corresponds pretty well to coronavirus deaths, not to our response to coronavirus, and suggests if anything we've missed cases https://twitter.com/ganeshran/status/1326115141147037701/photo/1

It's also a bit late to be worrying about the economy, if we cared that much about it then we wouldn't have the mess of Brexit. If we think poverty leads to deaths then maybe we should be addressing poverty, rather than just letting people sink. And if we are going to say we can't properly account for covid deaths because we're worried about exactly what people died of shortly after being diagnosed, then it seems a little speculative to be trying to offset it against some uncertain death rate due to uncertain economic impact.

ibmalone

Re: Infantile

Except they sort of are, what is described is a misinformation campaign to persuade people (well, persuade idiots, but there's an overlap), not to take a vaccine because it will turn you into a monkey/control your brain/cause 5G to erupt from your ears/be a betrayal of brexit.

ibmalone

Re: Still confused about the 90%

Well, it's the end of the day, so sure, I'll bite:

"It has been suggested "

To quote wikipedia, "by who?" Because numbers dying are much higher than those for flu.

In any case, nothing to do with hospitalisation risk, this is those vaccinated are 90% less likely to catch coronavirus than those unvaccinated. Meaning a widespread vaccination programme could stop the unconstrained spread because it wouldn't get far.

"I ask as those that do end up dying in hospital with covid are still classed the same as those that die because of covid."

No you don't, because if you had looked at excess deaths this year compared with other years it'd have been pretty clear that however you want to count them there are a lot more.

Let's Encrypt warns about a third of Android devices will from next year stumble over sites that use its certs

ibmalone

Re: Counter intuitive?

Wow, cell phone users using cell phones for phone calls only. Why, they're still in the Neolithic phone age!

Well, a few things to draw from this, it doesn't mean there are users of older android devices only using them for phone calls; for all we know from this 100% of those in service are still surfing the web. Second, these are smart phones, so it's not "cell phones for phone calls only", if you want to do that then you would be much better off without a smart phone, so it's likely the people on them are using them because they want those features.

Voyager 2 is back online after eight months of radio silence

ibmalone

Re: You know you need coffee when..

"A correspondent the regonynmiser has called Dave B"

(just realised this works two ways, see if you can spot both!)

Update to NHS COVID-19 app brings improved warnings, end to 'ghost' notifications

ibmalone

Re: Off with it's head.

What you describe sounds like one of the "ghost notifications", apparently (so we've been told) unavoidably triggered at the API layer, but not meeting the threshold for the NHS Covid app to give you a warning.

I say "apparently", because it's now going around that the warning level for the NHS Covid app had incorrectly been set much higher that intended, so it's quite possible that these host notifications should have been triggering a warning. Came to elReg today to see if they've covered it yet...

ibmalone

Re: Thank you all

It doesn't solve the problem, but it does patch up the difficulty of otherwise untraceable contacts. I can tell you who I shared an office with (working from home makes that quite easy...), but not who I shared a queue with.

Here's a nice graphic to help explain why even if something is not a magic bullet it is still worthwhile in our current situation.

https://www.pslhub.org/learn/coronavirus-covid19/tips/the-swiss-cheese-respiratory-virus-pandemic-defence-r3379/

Defence in depth, useful in IT, useful in public health, generally useful in most large numbers games.

I can't find a thorough assessment of Stop Covid NI effectiveness, but they do occasionally mention numbers: https://www.health-ni.gov.uk/news/stopcovid-ni-app-compatible-tracing-apps-scotland-and-jersey

“Since the launch of the ‘StopCOVID NI’ app in July almost 490,000 people have downloaded it and over 20,000 people have been contacted via the app to let them know they have been in close contact with someone who has tested positive for COVID-19. This is allowing people to self-isolate quickly if they have been exposed to the virus, reducing the risk of them infecting others.”

ibmalone

Re: Good

Sometimes it's interesting to know the scope of a problem, even if it's not an unbiased sample. It's at least information in the form of affects one person versus a few people versus many people.

ibmalone

Re: What?

Back in April (and into May I think) the queue for our local post offices was about an hour long. The big supermarkets I took to timing, and generally about ten minutes, but was longer if you went at the wrong time. Now we're back to supermarkets being crowded and no distancing, which means you're not likely to be next to one person for 15 minutes, but probably doesn't make it any less likely for a carrier to pass it on, just that you'll have no idea who to.

ibmalone

Re: What?

The details the make it into the story don't include the time limit. 13ft is almost exactly 4 metres. It's unclear from any of the linked articles what exactly the points system is, and strikingly the 120 point limit is below the amount of points awarded for the lower exposure, so why have the threshold?

No doubt this is all calculated somehow, they may weight the points based on the filter output for example, but have a look at the only article referred to that mentions the points change (I'm being generous, it's a link off a referenced article), https://healthtech.blog.gov.uk/2020/10/29/how-the-nhs-covid-19-app-is-making-the-most-of-cutting-edge-global-technology/

Now, I've not spent 15 minutes in a shop queue for a long time, yet I got the low exposure notification warnings. So where is the guarantee that walking down the street isn't enough to trigger an isolation warning without sight of the scoring system?

Did I or did I not ask you to double-check that the socket was on? Now I've driven 15 miles, what have we found?

ibmalone

Re: Do I have to put my hand through the flames

Having been in exactly this situation, I can report that my first response was the 'off' switch followed by unplugging. There were multiple sockets with things plugged in and it was quicker to switch everything off than pulling all plugs. (Didn't even contemplate trying to figure out which plug was the correct one.)

ibmalone

Re: Been there ... done that

We had one like that, the contacts were the spark plug terminal itself and a sturdy metal strip with a good gap and it's relatively low voltage (on the magneto side), so its quite robust to a little sparking.

ibmalone

That's one thing, but being able to turn off a socket that something is plugged into is nevertheless useful.

X.Org is now pretty much an ex-org: Maintainer declares the open-source windowing system largely abandoned

ibmalone

Re: Network transparency

Currently a lot of our working-from-home users are relying on X2Go, which in turn relies on X. I remember thinking we wouldn't have coped so well if covid had hit even a few years ago, looks like we might not cope so well a few years from now either.

Researchers made an OpenAI GPT-3 medical chatbot as an experiment. It told a mock patient to kill themselves

ibmalone

Re: Trained on 570GB of text scraped from the internet...

The obvious question is why anybody thought it could do this in the first place. It can pastiche (with a little coherence) the appropriate response in a given context, expecting real content is serious magical thinking. Let's see if it can fly a plane next!

Facebook tells academics to stop monitoring its political ads for any rule-breaking.... on privacy grounds

ibmalone

Founded in 2004

NHS COVID-19 contact tracing app is leaving some unable to access government self-isolation grants

ibmalone

I assumed it was something along those lines, but I haven't heard of people getting spurious self isolation notices with stop covid NI (maybe the people I know have less exposure of course).

They found a way to be able to fine-tune the criteria so that it only actually triggers self isolation in the app if their more sensitive criteria is triggered

That'd be less sensitive criteria surely?

ibmalone

These warnings were generated by the underlying APIs provided by Google and Apple, and are now followed by a separate push notification that tells users to disregard them.

Oh no it doesn't...

I've had a couple of these so far. The separate message turns up a couple of hours before the 'Possible COVID-19 exposure' notification. I also still don't understand quite why it generates this when the other apps don't, something is missing in the 'underlying APIs' explanation.

ibmalone

Except we do actually want people to follow the instructions.

Let’s check in with that 30,000-job $10bn Trump-Foxconn Wisconsin plant. Wow, way worse than we'd imagined

ibmalone

Human cost

What changes this from farce to tragedy is that people were forced out of their homes for it. Reply All has a good episode on how this played out for the residents in Mount Pleasant https://gimletmedia.com/shows/reply-all/wbhjwd

UK test-and-trace coronavirus data may be handed to police to nab those who aren't self-isolating as required

ibmalone

Re: If we didn’t have such ridiculous laws in the first place...

The police aren't used in the way you suggest, you're thinking of social services. So yes, it is a problem, because they are not being given this information to helpfully check up on the bed ridden elderly. They are being given this information to enforce isolation and that creates a perverse incentive (as if there wasn't one already) against people getting tested.

As for all the rest, which is pretty much unrelated, it turns out large chunks of people disagree with you about what rules society should abide by, so whatever my own opinions on any of it you're building fairy castles unless you can convince them otherwise.

Lastly of course, as they wield force with state backing the police tend to be the first tool of authoritarian governments, this is why hooking them directly into omniscient surveillance is regarded as a bad idea. Not doing so is a safeguard against government overreach. It's also relatively well accepted that having confidentiality in place increases people's willingness to engage with all kinds of services like addiction and sexual health clinics. Not everyone wants to live under scrutiny all the time.

In any case, almost nothing to do with your "BBC parroting official line" and failure to understand what was going on or, indeed, read.

ibmalone

Part of the problem is that T&T for Covid-19 is a centralised operation whilst in the past such operations have been done locally.

But this wouldn't generate consultancy contracts, or headlines about how many people are being recruited for the purpose.

ibmalone

Re: BBC take the opposing position . . .

If I may repeat what an AC has said, the "official view" is test and trace will be sharing data with the police, indeed the "parrot the official view" BBC do exactly that in this article as it's linked at the bottom: "Coronavirus: Police get access to NHS Test and Trace self-isolation data." The contact tracing app on the other hand, as many people saw this coming and kicked up a fuss (waves one "told you so" voucher), got relaunched using the privacy-preserving framework, in addition to which the instructions from the app aren't enforceable under the regulations (you should follow them though of course, the point being to try to keep a dangerous virus under control).

Of course the fact that people will confuse the two and the possibility of having the police turn up at your door, is going to put people off getting tested and using the app (even if the app has nothing to do with this). Brilliant!

Anyway, work on reading comprehension.

Elizabeth Holmes' plan to avoid her Theranos fraud trial worked out about as well as her useless blood-testing machines

ibmalone

Re: Hold on lads!

breaking enigma was thought impossible, but Turing did it.

No, he built on the work of other experts to automate it. Breaking enigma, wikipedia.

We tend to build in hindsight these maverick stories and ignore the work that led up to them. Breaking the speed of sound? The tip of a bullwhip does it, bullets were doing it before the invention of the jet, as did the V2 rocket.

"Extraordinary claims require extraordinary proof." If you are going to lay down millions or billions on something the experts say is impossible, do some diligence, get your own experts to look at it. Yes, maybe they'll be influenced by group-think, but scientists at least are meant to be able to look at and evaluate evidence. The board of Theranos, stuffed with business people and politicians, was itself very guilty of group-think.

Excel is for amateurs. To properly screw things up, those same amateurs need a copy of Access

ibmalone

Re: It’s not the database, it’s the amateur at fault

I was with you right up until "excellent database", let's not over-egg this. Most databases don't helpfully refactor stored procedures into broken ones with faulty syntax.

ibmalone
Joke

Re: At least they used a database

One manager said he wanted a list of defects of people working for him

Don't we all?

Scre-EE-m if you wanna go faster: BT's mobile network reigns supreme in UK-wide speed and latency tests

ibmalone

Re: What about?

Revisiting to say I've tried this on my 2017 android pie phone bought through CPW (UK base version, not EE firmware) and it does seem to work there. Maybe it was the case in the past that it needed the network's version of the phone, but certainly worth trying.

From the Department of WCGW: An app-controlled polycarbonate lock with no manual override/physical key

ibmalone

Re: Welcome back

Not sure if you know it, or care about it, but you really have a *personal* impact in your readership, Dabbsy.

Much like the Cellmate.

Regardless, very happy to see the return of the Dabbs.

ibmalone

Re: USB devices

I think 8 pin mini B is the smallest :(

But the only one I own has a choke on the cable!

Excel Hell: It's not just blame for pandemic pandemonium being spread between the sheets

ibmalone

Re: Alternative?

Though I'll modulate this to say that SQL concepts like tables and joins are transferable to other tools for manipulating datasets, it's just not necessarily where I'd start from if you've already got the data to analyse.

ibmalone

You left out source control (unless it's hiding under traceability). How is this version of the spreadsheet different from that version of the spreadsheet, and why was that change made?

ibmalone

Re: What should I use instead

Ok we should not use Excel. What should a non programmer use instead?

Are they writing macros? Are they creating calculations that depend on other calculations? Then congratulations, they're a programmer already.

If you run out of memory slots in your calculator, need to knock up some graphs or do some quick data entry that isn't big enough to warrant setting up a database (and doesn't have horrible GDPR implications), then excel will do. If the task has gotten bigger and you've got sheets that depend on other sheets and calculations that need to be filled down and data that needs to be pasted in and a spreadsheet named quarterly_report_v2.2_2020-q3-tmp-newmacros.xls, then stop and think what you're actually trying to do. Maybe that's what should be used instead, some introspection.

ibmalone

Re: Alternative?

SQL would not be my choice for data analysis (see reply further down). For storing, sorting, extracting data yes, you can even do some of your data cleaning and preparation in it, and I'll confess to having done the occasional avg() and group by, but the statistical tools it offers are very limited and how are you going to draw a graph?

ibmalone

Re: Alternative?

Stata is the most welcoming of the stats packages I've seen, simple tasks can be done through the gui, you can do an entire basic analysis without typing (and generate the commands used at the same time, so the analysis can be stored for reuse independent of data). It's not cheap though. SPSS can also be used from the gui and isn't cheap either, but they've been teaching psychology students to use it for years, so you don't need to be particularly into programming to make progress.

R with R studio has a bit more of a learning curve, but the open source version is of course free. You can view data, graphs are plotted in a window. Yes, you'll need to learn a little R, but it cuts out much of the boilerplate you need to do analysis in a generic language. It has its own idiosyncrasies that make it a little awkward for people from a more traditional procedural background. Strongly recommended to start with the tidyverse framework (includes dplyr, ggplot2 and others) which make it much more powerful. Here's a quick start https://www.datacamp.com/community/tutorials/tidyverse-tutorial-r

There is a learning curve with all of these, as you have to think in a more structured way (a bit like going from a word processor to latex/lyx), and excel will get you a lot of the way for simple noodling around with simple sets of numbers provided you're careful, but with a clear defined problem to work on people can often manage to get things done with these kind of tools more easily than with a general task like "learn python".

ibmalone

Apparently excel can't do that either, it just pretends to.

ibmalone

Re: Relax...

It's a good point, but it shouldn't have to be like that. Horizon in particular started out as something else that was Too Big to Fail (before that was cool) and got shoe-horned into its final role.

What a Hancock-up: Excel spreadsheet blunder blamed after England under-reports 16,000 COVID-19 cases

ibmalone

Yes, that's bizarre. I have a couple of thoughts, which may well be wrong:

1. long data format that records different value types to different records. Maybe nasal and throat swabs reported separately for example and each patient therefore has a separate row for each with different test type.

2. Fixed column width format where multiple rows are needed for any field that requires a line wrap. No, you shouldn't need to do that in excel.

ibmalone

Re: CSV?

CSV handling libraries of course can do that too, why wouldn't they? But this is all a bit of a sideshow, because the argument isn't really that they should have used csv, it's that they should not have used xls.

ibmalone

Re: CSV?

CSV in general is schema-less. CSV following a schema you've agreed for data interchange is by definition not schema-less and import libraries will let you specify your expected data types and fields (often trivially). Anyway, the concept only applies here in watered-down form, since we're talking about multiple records of the same form, it's only the field types and order to be specified, not the relation between different types of observation.

There are definitely other choices which may be better suited to a task, but what csv has over excel is that it's easily generated and read, doesn't have the complexity of multiple sheets and non-data objects, isn't primarily intended to be loaded into software where a field can be altered with an accidental keypress, and doesn't have hard-coded limits on record or field counts (but yes, there are better formats for really big data). And to top it off, xls / xlsx are also schema-less.