Posts by logicalextreme 606 posts • joined 8 May 2020
Cheers — missed the editing window, but closing and reopening the application sorted my missing mailbox. Looking good so far! I see what you mean about Vivaldi…just as Vivaldi Mail's been released :) I think in my current job I prefer a dedicated mail client though, so good to see Thunderbird coming along.
If it had a smidgeon of type safety, I'd be inclined to agree. But if I need object and methods and I'm on Linux, I'll use Python — I can't think of anything PS does better than Python except for (occasionally, but even then not often) doing Windows admin. And at least you can rely on Python to throw an exception and stop when it hits one, rather than barrel on regardless waving a sledgehammer around…
Or, not related to the OP's topic, but also a stupid thing that companies say that I like to rephrase within the analogy of burglary and then repeat back to them, is something akin to "we don't need alarms/locks/doors/to hide our valuables/to reduce the number of unnecessary valuables we have; burglary's against the law so it can't happen to us, and if it does the perpetrator will go to jail".
It's not a model that's unheard of. I'm somewhat of the opinion that everyone's healthcare/genetic data should be fully accessible to anyone who wants to see it, as anybody from doctors to script kiddies would be able to mine it for patterns and we'd hopefully be able to propose formal studies and trials that had a lower chance of failure, meaning lower drug prices/decreased mortality. However it would also have to be the case that there was no stigma surrounding medical/mental health conditions (opening it all up would hopefully accelerate the reduction of such stigmas, but not eliminate them) and that it would only be used for good (e.g. not selling people stuff/making outrageous profits from healthcare, no genocide, no eugenics etc). People would only be less prissy about their data if all those conditions were met, which is certainly not a possibility given current evidence; and furthermore even if you could trust everybody right now the potential risks would still be there. Once you'd opened the door there'd be no going back except for new data.
Completely open healthcare data is a common enough pipe dream that it's mentioned in some of the NHS documentation about data protection/consent regs — essentially saying that there are plenty of people who don't care who has access to their healthcare data, but you still have to get their consent to use it or transfer it for the most part (this predates whatever the hell shady crap happened last year where they tried to change the rules on the QT).
PowerShell fans will be disappointed to learn that "Programmatic access to Windows Autopatch is not currently available."
Much as I loathe PowersHell when misused as anything approximating a programming language, it's sometimes quite good for sysadmin tasks (as long as you don't expect anything like robust error handling).
MICROS~1 have heavily implied on a number of occasions that all new sysadmin functionality will be, if not "Powershell-first", then at least available to PS via cmdlets or modules at the point of release.
It's simply not true. Windows config remains the same tangled mess of registry keys, config files, MMC modules of varying antiquity, crappy new Settings screens and various other gubbins, and the likelihood of you being able to administer everything via any CLI is basically zero. No wonder they had to ditch the idea of truly headless Windows shortly after the first releases of Server Core.
I know there can be a fair bit of inconsistency in administering the various parts of your average *nix instance via CLI too, but at least it's possible, not to mention expected.
Yeah, it's something I've tended to not do in the past; I've usually built up enough cruft in the home directories that I figure a clean go at it is nicer and I'm unlikely to truly need anything I haven't archived away already (plus I've got nightly backups just in case). You can get a fair bit of config clutter in home directories too, some of which may be outdated when all the new versions of packages turn up in the new distro.
Glad to hear it; after a disastrous time with 14.04 and a catastrophic loss of my 16.04 system, I've been on 18.04 since shortly after it arrived without too many problems but I'd like to move to a clean install soon for various reasons and also have an extracurricular project that's arrived just in time to try out 22.04 separately. It'll be nice to finally have python3 as the default without having to do anything ungodly, too.
Wiki article for great justice.
The convention's even got its own classes in various popular high-level languages like Python, C#, Ruby, Rust, Dart and Java.
Not that it matters because some tech's apparently ambling towards a dynamically-typed and/or stringly-typed quantum state of who-gives-a-fook.
One place I worked at had the grand idea of dispensing with numbers entirely and naming releases after dead languages, e.g. vPictish, vLatin, et cetera. Went about as well as you'd expect, especially when a drugs bust happened and the local police force turned out to have been using the same naming convention for their sting operations.
I seem to recall reading somewhere that the safety profiles of various energy generation methods take quite a holistic view too, so it's not just operational safety and risk to the general public that's accounted for — construction accidents and accidents during maintenance and decommissioning are also taken into account and (presumably) totted up against the average total kWh generated, and nuclear tends to come out on top when doing this.
It's probably quite anthropocentric though, and I'm not sure how spent fuel factors into it. I'd also take a lot of things with a pinch of salt as I imagine there's a lot of lobbying money flying around from parties with financial interests in any given energy generation field. Books sound good though, I'll put them on my list.
We had everything stored as UTC. I did suggest presenting the reports in UTC instead of local time, but that got kiboshed because it would "cause confusion". ¯\_(ツ)_/¯
The company's currently looking for a buyer and the founder just jumped ship. Pretty sure some of my reports are the only things that have kept them going as long as they have, and they weren't exactly rocket surgery but I've done backchannel support for the poor sod I left behind at least twice a year since I left (popcorn emoji)
I ended a conversation with the trading director at my last job with an explanation that, as good as I am, I was as yet unable to control the length of the earth's orbit around the sun.
She'd been ever more insistent over the course of the conversation that there could be no more or less than 52 numbered weeks of precisely 7 days in each trading year, and that anything else was unacceptable. She eventually stormed off and I never talked to her again.
Also spent the Monday following each clock change (well, always the March ones and sometimes the October ones) drawing timelines for various members of the finance department (all chartered) to explain why there weren't any sales for the hour of 0100, or why they seemed to have doubled. More often than not the idea seemed to eventually click with them, and they'd appear to back down. Then thank me for the explanation. And then ask when I'd be fixing it.
I've got a slightly better one than that. I'd just installed Microsoft's own minidump file analyser for Microsoft's own minidump files onto one of the production database servers (there was some sort of pseudo-justifiable reason for doing this on prod).
The installer finished and declared that to before using the software there'd need to be a reboot, and asked if I wished to do this now. Two buttons, OK and Cancel (or possibly Yes and No).
Knowing that this message was almost always untrue and I'd be able to use the software without a reboot, and also knowing that if it didn't work I could just schedule the reboot into a maintenance window, and also liking a tidy desktop, I of course clicked Cancel (or No). Cue immediate reboot of production database server.
It certainly taught me a lesson but if I ever get locked in a room with the person responsible for that particular dialogue box it won't be pretty.
MediaWiki. Can't recommend it enough, but of course it can potentially depend on what you want out of a tool.
MediaWiki's open source, so you can do what you want with it, and my number one reason for always suggesting it (especially for knowledge bases, though it applies to other uses too) is that pretty much everybody uses Wikipedia at least once a day — if you leave the visuals mostly alone apart from your own logo and stick to basic formatting and sectioning, your pages will look like Wikipedia articles. I've found this to be invaluable in having people read and comprehend pages quickly. Caveat of course is that you run it yourself somewhere, though the cost of doing that (including learning) probably pales in comparison to paying for an Atlassian product.
I find Confluence to be an unholy mess that's loosely inspired by the idea of a wiki, but with all the good bits cut out and a ropey sociopathic WYSIWYG editor sellotaped onto the side (wiki markup was once supported and is now removed, and markdown isn't a thing AFAICT). Basic needs like linking to a section of a page are either impossible or can very occasionally be achieved with a lot of clicking and copy-pasting of weirdly-formatted URLs that you sometimes have to construct by getting <div> ids out of the page source. Even linking to another Confluence page is a chore. Flowing doesn't seem to be a thing, so you have to put things in "columns" which the WYSIWYG editor helpfully displays as…rows, and keep manually adjusting the widths as you go. To be fair to Atlassian the ⟨I⟩ in WYSIWYG can arguably stand for "isn't", so I think they're covered. The vast majority of basic features that I've been unable to find I've then Googled, and found a ticket (or several) in Atlassian's Jira instance telling the multiple users requesting it that ( you used to be able to do it, but that feature was removed | you can only do it via a third-party extension | it's not possible and we're committed to making sure it'll never be possible ). That's when their own Jira isn't either failing to render half the elements on the page, or down entirely.
I'll admit that Confluence is marginally better than my brief experiences of SharePoint though.
Yup. I've never been able to get into my account. Coming up on eight years now. Cable's fast and the connection's solid as a rock, so I've rarely needed to, but goddamn. Website's broken, contact forms don't work, email addresses bounce back. Occasionally I phone them up to tell them to put the price down and just leave it at that. I assume there's no way for me to cancel the account and I'm stuck with it till the end.
I started watching the lot of them after hearing enough references to it to know it was my kind of thing, but gave up after about four episodes due to the format and, mostly, the sheer amount of filler. If somebody could distill the ~10 minutes of content per episode from all the rest of the stuff that was there purely to help them limp up to the next ad break, and also defragment the different myths in each episode so that it's not like watching fucking Memento, I'd give it another shot.
I'm not sure how it is for other drivers, but I can tell you that that type of indicator slows me down as a pedestrian. I can no longer ascertain indicator status from a brief glance; I have to pretty much stop moving completely and stare at the car for a few moments before I'm confident that the light isn't doing something fancy and useless. Last I checked use of, or even possession of functioning, indicators wasn't a legal requirement in the UK and you're blessed if anybody decides to use them so I'm already slowed down more than I'd like to be; but I do wish things would improve on the whole rather than backsliding.
And I agree with you — I used to happily use the mail and torrent clients in Opera and might get around to using the mail and calendar in Vivaldi if Thunderbird stops improving, but if I want my browser to chew up my CPU I'll just try to perform basic browsing activities on any modern website.
I wasn't aware that it was happening in existing applications in general, but Vivaldi's inherited the (OG) Opera approach of implementing non-web browsing features in the browser (mail client and calendar; torrent client hasn't appeared as yet but was in Opera) so I think he's probably responding to or anticipating a feature request.
This is how the NHS tend to do it and it drives me up the wall. yyyy-mm-dd and dd/mm/yyyy are our official date formats. Nobody uses YDM anywhere, dd/mm/yyyy should be retired, and yyyy-mm-dd is ISO-standard (even though ISO are nobs and 8601's messy, it's the closest thing we have to an international standard). No need for Anglocentric nonsense, especially when it's presented as a "solution" to an i18n problem that was solved decades ago.
they were able to say what was actually taken, as opposed to what you usually see from companies that have been breached which is the prolix equivalent of "customer security is paramount to us, ¯\_(ツ)_/¯".
It indicates that they've got some sort of auditing in place at the bare minimum, and that presumably someone working there at least knew what infosec was, once.
Ah, nice to know it sucks whatever ecosystem you're putting up with.
Tying to contacts is fine for me because I mostly know individuals or couples with mobile phones, so that all fits in nicely. That said, I'm at the age where some of them have started having kids, so I can see how that'll cause problems.
What continues to enrage me about Gcal reminders is the fact that you still can't get reminders about contacts' birthdays. They sync nicely into their own calendar that you can display in the interface, but I'll only find out about one in advance (or on the day, or possibly a few days later) if I happen to notice it in said interface.
People have been asking for this basic feature since Calendar was promoted to a "real" product, and the "solutions" range from the manual to the scripted — but all pretty much along the same lines, "copy the events one by one into your main calendar or a new calendar, then add reminders to those events". Meanwhile I believe Apple users laugh merrily. Hell, I think even Facebook can do it.
Never had issues creating alerts for other events but I don't really use it for meetings, so sounds like your issue might be related — perhaps something to do with them being "second-hand" events. It's unutterably daft.
I wouldn't say that alerting is part of a calendar's one job though — the one job is to record events. The basic extra things that can be achieved via electronic calendars however are invitations, integrations, viewing other people's calendars and alerting. Trouble is that if you these extra features don't work you may as well just have a paper calendar and remember to look at it every day.
Yeah, I think with a lot of these "young people don't use X" there's an implicit assumption that said cohort of young people will continue to not use X when they're at the age of the people who do currently use X, which often simply isn't true. This may even apply to e.g. Facebook, not sure.
As for email specifically — it doesn't matter how young you are, the overwhelming majority of retail websites aren't going to let you check out without an email address. I can't see an entire generation refusing to register accounts and using BugMeNot. No, email will survive like the cockroach that it is and people like us will continue having to write horrible code to make it do things it wasn't designed for like HTML, or authenticating the sender.
The Register - Independent news and views for the tech community. Part of Situation Publishing
Biting the hand that feeds IT © 1998–2022
