Skip to main content

Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.

Messages - Frenzie

2
Browsers & Technology / Re: E-readers
I don't think the person who wrote that has improved on it.
3
What file name/location precisely? I probably don't know what you're talking about but I'm still curious. ;)
4
There's also https://github.com/Ernir/Kate-bbcode

But Markdown highlighting seems to do the job reasonably well too?
5
Oy, that's pretty annoying. Glad it worked. :)
6
The stuffy image of the Office branding? Office means pretty decent products that do arguably stuffy things well. Calling it 365 won't make those stuffy tasks any less stuffy... I predict that in a decade we'll see them moving away from the stuffy 365 branding then.
7
Really, huh. Here's a Dutch article about it from last week. The headline says you can't just do these modern calculations on a (beer) coaster anymore.
https://www.nrc.nl/nieuws/2020/03/22/berekeningen-die-je-niet-meer-op-een-bierviltje-doet-a3994580

(Translations from DeepL with potentially some minor corrections by me in case I noticed a glaring inaccuracy.)
Quote
Ze zeggen het stuk voor stuk, de vijf wetenschappers die NRC voor dit artikel sprak, allemaal gespecialiseerd in modellering van uitbraken van infectieziekten. Zonder wiskundige modellen is de verspreiding van een infectieziekte niet te voorspellen, en is het effect van een maatregel niet door te rekenen. Bierviltjesberekeningen op Twitter werken niet in een hypercomplexe werkelijkheid waarin ziektes anders uitpakken per leeftijd, geslacht of bevolking. In een wereld waarin virussen hun eigen verspreidingssnelheid veranderen, omdat ze dwingen tot ander gedrag, of hun eigen dragers uitroeien, lopen lijnen niet recht. Het barst van de onverwachte drempelwaardes, zelfversnellende effecten en elkaar versterkende parameters.

Het was de uitbraak van hiv in de jaren 80 die het inzicht bracht dat we niet zonder complexe modellen konden, zegt Mart de Jong, hoogleraar kwantitatieve veterinaire epidemiologie aan de universiteit van Wageningen aan de telefoon. In het Westen was in de decennia daarvoor het idee gegroeid dat infectieziekten iets van het verleden waren, te danken aan hygiënemaatregelen, vaccins, antibiotica, de uitroeiing van pokken.
Quote
They say it all, the five scientists NRC spoke to for this article, all specialized in modelling outbreaks of infectious diseases. Without mathematical models, the spread of an infectious disease cannot be predicted, and the effect of a measure cannot be calculated. Beer mat calculations on Twitter do not work in a hyper-complex reality in which diseases turn out differently by age, gender or population. In a world where viruses change their own rate of spread because they force different behaviour, or eradicate their own carriers, lines do not run straight. It is bursting at the seams with unexpected threshold values, self accelerating effects and mutually reinforcing parameters.

It was the outbreak of HIV in the 1980s that brought the insight that we could not do without complex models, says Mart de Jong, Professor of Quantitative Veterinary Epidemiology at the University of Wageningen on the phone. In the preceding decades, the idea had grown in the West that infectious diseases were something of the past, thanks to hygiene measures, vaccines, antibiotics, and the eradication of smallpox.

Quote
De modellen die nu op de servers van het RIVM draaien, zijn uitgebreider en hebben veel meer datahonger. Jacco Wallinga van het RIVM voert ze in ieder geval drie soorten gegevens. Hij gaat er even voor zitten. ,,Ik zie dit maar als mijn pauze."

Eerst de bevolking. Alle mensen die in hetzelfde jaar geboren zijn, vormen in de RIVM-modellen één groep. Van elk geboortecohort voert het RIVM het percentage dat extra risico loopt in. ,,Dat komt van het CBS, dat zijn de mensen die een oproep van de huisarts krijgen voor een griepprik."

Per leeftijdsgroep probeert het RIVM vervolgens te schatten hoe besmettelijk die is. Wallinga: ,,Uit een goed geanalyseerde dataset uit Shenzhen blijkt dat kinderen minder besmettelijk zijn. Precies het omgekeerde als bij griep, waar kinderen de motor van de epidemie zijn. Maar dit is een inschatting hoor, het is nog niet zo hard."
Quote
The models now running on the RIVM servers are more elaborate and have much more data hunger. Jacco Wallinga of the RIVM feeds them at least three types of data. He sits down for a moment. "I see this as my break."

First the population. All people born in the same year form one group in the RIVM models. For each birth cohort, the RIVM enters the percentage at extra risk. "That comes from the CBS, that's the people who get a call from the doctor for a flu shot."

The RIVM then tries to estimate how contagious it is for each age group. Wallinga: "A well-analysed data set from Shenzhen shows that children are less contagious. Exactly the opposite as with flu, where children are the engine of the epidemic. But this is an estimate, it's not so hard."


Quote
De wetenschappers die NRC spreekt zijn niet bang om toe te geven dat het bij modelleren om schatten, uitzoeken en bijstellen gaat. Zelfs het simpelste modelletje zit vol onzekerheden.
Quote
The scientists who spoke to the NRC are not afraid to admit that modelling is about estimating, selecting and adjusting. Even the simplest model is full of uncertainties.

Quote
Wallinga vertelt over zo'n ,,totaal tegenintuïtief" effect dat hij in zijn eerdere modellen zag. Griepvaccinaties bij kinderen blijken juist voor hevigere griepuitbraken te zorgen. ,,Dat komt, denken we nu, omdat er door die vaccinaties minder natuurlijke immuniteit wordt opgebouwd in de bevolking, die beter beschermt. Als je dan een keer een mismatch hebt tussen het vaccin en de griep van dat jaar, dan heb je écht een grote uitbraak."

Je bedenkt het vooraf niet, zegt Wallinga. ,,Maar als je het ziet in je model, en je gaat redeneren, dan is het ineens logisch."

Heesterbeek kan meer van dat soort verrassingen opnoemen. Hij zag dat populaties woestijnratten nul pestuitbraken hebben, tótdat hun holen meer dan 35 procent bewoond zijn. En dat suboptimaal vaccineren tegen rode hond slechter werkt dan helemaal niet. ,,Het zijn waarnemingen die je niet zonder modellen kunt begrijpen."

Daarom zijn die modellen zo zinnig, zeggen de wetenschappers. Die kunnen snel uitrekenen dat het sluiten van landsgrenzen geen zin heeft als je zelf een besmettingshaard bent, of dat groepsimmuniteit óók een dempend effect heeft voordat het hoog genoeg is om het virus in te dammen.
Quote
Wallinga talks about such a "totally counterintuitive" effect that he saw in his earlier models. Influenza vaccinations in children appear to cause more severe flu outbreaks. "We now think that this is because these vaccinations reduce the natural immunity built up in the population, which is better protected. If you have a mismatch between the vaccine and the flu that year, then you really have a major outbreak".

You don't think about it beforehand, Wallinga says. "But if you see it in your model, and you start reasoning, then suddenly it makes sense."

Heesterbeek can list more surprises like that. He saw that populations of desert rats have zero plague outbreaks, until their burrows are more than 35 percent inhabited. And that suboptimal vaccination against rubella is worse than none at all. "They're observations you can't understand without models."

That's why those models make so much sense, the scientists say. They can quickly calculate that closing national borders makes no sense if you are a source of infection yourself, or that group immunity also has a dampening effect before it is high enough to contain the virus.
9
Yeah, a worrisome development. I'd rather pay for a new Windows every 5-10 years, because what exactly are they making money from then? On the other hand, if you buy a new laptop or something, you do pay for that. Building your own PC is niche.

Supposedly the PC market is/was "declining rapidly." I find that questionable, but of course back in 2000 our computer from '96 was hopelessly outdated and right now my computer from 2014/2015 may not be top of the line but it's hopefully years from needing replacement.[1] Plus that computer from '96 was literally our first.

A market that's "mature" or whatever you want to call it isn't "declining." It just isn't the big new thing.
But I did get a new GPU a couple of years ago, and I've bought probably two SSDs to upgrade my storage in the meantime.
10
It's just half-yearly updates, I wouldn't call that rolling release. I'm not sure how annoying it is or isn't though.
11
In part that's because in the old days, what we have now would be called Windows 10 Service Pack 2 or 3. Now it's called Windows 10 build 1909. But it's still way worse than Windows 7 because they #$% up a lot of stuff for no reason that they're still slowly rebuilding.
12
Word 97 (yes, it opens and edits DOCX)
Wow, really? Good on Microsoft. I know they released plugins for older versions of Office but I didn't realize they went back a full decade. And that right there is why Microsoft is a much less annoying evil than Google.

Markdown in which editor? Or is there an editor called Markdown?
I tend to use Geany for my Markdown, which is my preferred format. In practice as soon as someone else is involved it's easier to get started in Zim or LibreOffice and possibly switch to DOCX.

Some things that I want to look nice I do straight in LaTeX without the Markdown step. It's slightly more annoying to edit, but otherwise it's just a useless time wasting step. LyX can also be a nice option.
13
Was track changes already properly developed back then? I think these days 2007 or higher is kind of required due to DOCX. I bought a used Office 2016 Pro Plus key 'cause everyone keeps sending me these Office documents and I lost access to Office 365. You can also get Office 2010 and 2013 for about €20 but I'm not aware of any concrete advantages to that. Actually it's more like €20 for 2016 and €25 for 2010 and 2013.

Practically speaking you can get by just fine with WordPerfect 5 or MS Works 2, arguably a lot better than with GSuite/Office 365 in your browser. The only problem is that Office 2000 and higher can't read old MS Works files.

MS Office activation is really obnoxious. Only one copy. Well, maybe it works great for most people but for me I decided to put it on real Windows 10 on my desktop, which is sufficient but limiting. It takes a special reboot to use Office. But for myself I just use LibreOffice.
14
I own this album on CD, and I can't quite tell if I think this YouTube vinyl recording is better or worse.
https://www.youtube.com/watch?v=PzVTuKb6vzc&list=PL0YcRc9TaumX6y7I6nYqGf3qSIuETThgd&index=3
15
Der Spiegel writes some odd stuff sometimes.

https://www.spiegel.de/international/europe/dutch-exceptionalism-will-holland-s-looser-corona-policies-pay-off-a-fe2dd266-910c-4b62-9ddd-77b9189fc348#ref=rss
Quote
Rutte's most important adviser in the corona crisis -- Jaap van Dissel, Director of the Center for Infection Control at the Dutch National Institute for Public Health and Environment -- holds views on the disease similar to his prime minister. Van Dissel explains that the virus must not be allowed to spread freely, otherwise the health system would be overwhelmed. And that herd immunity is not the goal. But he says it could become a weapon in the fight against the epidemic.

The RIVM holds views similar to Rutte? Oh yes, because as we all know Rutte has studied epidemiology and his favorite hobby is modeling SARS. What an odd way to phrase it.
16
Like I said, Office 2019. :) You can't get it any cheaper than ~60-100 as a used key from the looks of it, so I don't really see any reason to pay the premium over less than 20 for Office 2016. Which is less than two months of Office 365. :lol:
17
Google Docs Spreadsheets doesn't have the Excel defect I mentioned, so it may be the opposite. Practically speaking they're all junk in my experience, but good enough for basic simple stuff. I doubt Google Docs is better than LibreOffice at opening MS Office documents, possibly worse.

That's something Office 365 online Word probably does just fine. And it's like Office light on every OS with a modern browser.
18
Now with 365, the thing struggles opening a new document when there is already a document open.
I may have been slightly confusing -- well not me really, but Microsoft. There's 365 the online interface and the 365 the subscription Office. The latter is pretty much the same as it always was afaict. It's just Office 2019[1] but you have to pay up every year.
Which I can't meaningfully distinguish from Office 2007. But that's probably a good thing.
19
Same in Office 365. Filters are an integral part of using Excel, and the whole point of 365 is collaboration, right? But applying a filter applies it for everyone. It's useless.

https://excel.uservoice.com/forums/304921-excel-for-windows-desktop-application/suggestions/13151544-enable-user-filter-views

I appreciate that it must be astonishingly difficult to implement. People first publicly noticed at least 5 years ago after all, but the simple fact is that a traditional network drive with user edit lockout works a million times better and has worked that way just fine since presumably at least the mid-'90s. I would expect it hasn't fundamentally changed from a user perspective since MS Office 3 or 4, albeit possibly 95, 97 or potentially even as late as 2000.

Tbh I mostly prefer LibreOffice over MS Office, or at least Writer over Word and possibly Calc over Excel. I've never really played with Publisher for example.

Unlike Writer I don't think Calc is really any better, probably not really any worse either, but Excel doesn't show you what you've selected in inactive windows. It's incredibly annoying. I have no idea what they were thinking. Also Excel has or had this weird program-wide undo instead of document-based undo.

Incidentally, if you use the ribbon in LibreOffice then Alt+letter doesn't work to activate a tab like in MS Office. So they both have their ups & downs. The MS Office ribbon has grown on me somewhat after using it for a decade, although I still think traditional menus are easier to find your way around. In any case Microsoft made very sure that other than visually, ribbons behave exactly the same as menus from a keyboard perspective. This is a lesson that no one else seems to have learned. Rather than contorting your fingers weirdly, you do something like Alt+o, d for "clear direct formatting". This automatically works for every menu entry, so if you use a function a lot you don't have to make up custom keybinds or whatever.

LibreOffice's experimental ribbon ("tabbed") doesn't work that way, unfortunately. It does underline letters though, so I guess it's planned.
20
The weird thing is that Outlook is better at a lot of that stuff and has been since time immemorial (mid-'90s).

Microsoft Lync was fine for chat/call/video, although no more than that. But was it worse than Teams... disturbingly, probably not.
21
The CPB[1] director says it's "practically unavoidable" that there'll be a recession.[2] But besides how worrisome that sounds, the definition of a recession is simply that the economy shrinks for two quarters straight. So it's mostly up to what happens after those two quarters whether the practical effects will actually line up with what we mean by the word.
22
DnD Central / Re: Grammatical Mutterings
I suspect there aren't much Dutch pronunciation in Indonesia
They won't be pronounced quite like in Dutch if that's what you mean, but they're definitely not pronounced like in English.[1] Some rarely used words that went Greek/Latin/French → Dutch → Indonesian might be at some minor risk of Anglicization, but I imagine thousands of Dutch words should be quite safe.
A lot of words are French-style in English, think of Anglo-French police vs. Dutch politie, pronounced and in Indonesian also spelled polisi, albeit politsi in more careful Dutch speech.
23
Russia apologists seemingly do little more than invent bizarrely irrelevant tu quoques and whataboutisms. Russia's annexing Crimea? What about our annexation of Indonesia? It was also in the '10s and '20s after all... you know, the 1610s and 1620s. And remember when we kicked the native Americans out of New Amsterdam? How dare we say anything about Chechnya and Georgia?!

They might at least try to come up with something we're actually doing right now (e.g., absurd immigration policies), even if "we do some bad things too" is hardly a defense.
24
His comments about the past come down to him being a doofus. They were wrong about AIDS, therefore they're wrong about COVID-19.

In a banal sense this is true of course. We're in the middle of this thing and there are still many unknowns. But AIDS was 40 years ago! We quite literally have our modern epidemiological models because of AIDS (tested/improved on SARS etc.), as it turned out to be absurdly wrong to assume everyone's effectively the same with regard to how likely they are to get it, how contagious they are, etc. From my understanding, modern models take every age and gender independently, and various other variables too. Still an abstraction of course, but one that's a heck of a lot less inaccurate and has proved fairly dependable on all kinds of infectious diseases -- albeit on more information than we have now. At the RIVM they apparently have a Japanese intern who can translate Chinese, which has been instrumental for their ability to model COVID-19.

Unless you want to call The Lancet propaganda, see for example https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(20)30567-5/fulltext to get it straight from the horse's mouth instead of filtered through mass media.

Going straight to the science, you'll find a distinction between what they found and government policy (because presumably they feel they can't explain not closing schools to the population):
Quote
School  closure,  a  major  pillar  of  the  response  to  pandemic  influenza  A,14  is  unlikely  to  be  effective  given  the  apparent  low  rate  of  infection  among  children,  although  data  are  scarce.

That's also where you'll find the age thing:
Quote
Data  from  China,  South  Korea,  Italy,  and  Iran  suggest  that  the  CFR  increases  sharply  with  age  and  is  higher  in  people  with  COVID­19  and  underlying  comorbidities.18Targeted  social  distancing  for  these  groups  could  be  the  most effective way to reduce morbidity and concomitant mortality.  During  the  outbreak  of  Ebola  virus  disease  in  west  Africa  in  2014-16,  deaths  from  other  causes  increased because of a saturated health­care system and deaths  of  health­care  workers.19  These  events  underline  the   importance   of   enhanced   support   for   health­care   infrastructure   and   effective   procedures for   protecting   staff from infection.

Of course I'm no expert on medical science, let alone virology. Perhaps all the studies are super flawed, or trying to confuse readers with the difference between statistically significant and actually meaningful. Significant only means that something isn't plain noise, but if you find that 5 in 100 elderly people die and 4 in 100 young people, that may not meaningful even if it is statistically significant. But this doctor isn't engaging with the science, basically just saying the UK government or media is stupid or evil. Perhaps he's mostly or even entirely right about all of that. But regardless if he is, he's using fallacious scientifically irrelevant reasoning, practically scientifically literate although surely he can't be (?), to prove his argument.

I decided to ignore all that because his conclusion doesn't exactly follow from his own premises. We've pretty much already lost more money than we could save, unless he assumes the UK government is both incredibly incapable and simultaneously evil. There could be something to that[1] but it sounds more like he considers them callous and clever.
At least for some governments; not quite sure the UK is one of them.
25
Heh, he sure starts funny. "Don't worry, this isn't going to be one of those conspiracy theories. I'm going to explain what the hidden agenda is."

Can offing some 2-4% of "over-70s" save even a tenth as much money as has already gone up in smoke over the past few weeks? The basic premises seem a bit math-challenged.

In Belgium, the budget for healthcare is some 30 billion, pensions some 40 billion.[1] Let's say all of that is for over-70s, so 70 billion a year spent on performing our social duty.

Imagine 10% is killed off by corona, which gives us 7 billion in savings per annum.[2]
Four days ago, the economic damage from corona was estimated at 16 billion euro, so more than two years of worth of grossly overestimated savings. By the end of the crisis, even in terms of these grossly inflated numbers it'll probably have cost us two decades. Or in other words, basically breaking even.

Now if we put back in the reality that the costs for "over-70s" are maybe a fifth of that napkin calculation, the argument doesn't seem to have a leg to stand on.
This is comparatively more than most places.
The actual number is obviously less than 1.4 billion, because it's "only" about 2% and healthcare isn't just for the elderly anyway. Just grossly overestimating to give the argument its best fighting chance.