💾 Archived View for dioskouroi.xyz › thread › 29349398 captured on 2021-11-30 at 20:18:30. Gemini links have been rewritten to link to archived content
-=-=-=-=-=-=-
________________________________________________________________________________
Hello World!
It appears the EpidemicsRunner software toolkit has been made open source. It was developed by a crossfunctional team of the German Goethe Center for Scientific Computation Universität and has apparently been used in real life settings as guidance tool.
Images, description and source code can be found on
https://github.com/devanshr/epidemics
The EpidemicsRunner allows the user to run various algorithms used in epidemiologial modeling. The main feature is its in build editor and support an optimization framework, which allows the user to adjust model parameters to existing data.
It has been designed for ease of use (but still full of features) because its original purpose was as a quick response tool for e.g. local authorities. The vision: A non expert user can quickly generate various scenarios based on local data and decide local measures until expert knowledge arrives.
The tool has been used in various capacities. In order to inspire learning and maybe even help others, it has been made open source.
You can use and modify it however you want. You can also import only the headers to your own C++ program.
If you want to use the GUI, a build script is provided in the EpidemicsRunner folder. All features of the GUI are active if you put it in the plugin folder of the UG4 framework.
jonpoljargo.de
There's no license on the codebase, could someone add an OSI-approved license to this codebase? UG4 itself is LGPLv3 (
https://www.gnu.org/licenses/lgpl-3.0.en.html
), so licensing this under the same terms would make sense.
A license will be added soon and will be probably very permissive
This is great. First visible commit is on Mar 12th. The front-end is HTML? Hope we start to see more gov apps being built as web apps hosted in the cloud. With WebAssembly, the C++ code should be usable directly.
Thank you for making public money sponsored code public.
If you don't mind why was the software written in C++?
I don't mean any disrespect by asking this, but asking why this was done in C++ implies you believe some other option would have been better; or at least it gives me that impression. What language should this have been done in, in your opinion?
The code was written in C++ because of performance considerations of the ODE solvers.
Does it mean you tried something like python or r and discovered that it was to slow?
Maybe they did not have a useless idiot sitting around asking to rewrite shit in Rust.
Would you please stop breaking the site guidelines? We ban accounts that keep doing this, and you've been doing it repeatedly.
If you wouldn't mind reviewing
https://news.ycombinator.com/newsguidelines.html
and taking the intended spirit of the site more to heart, we'd be grateful.
Curious, where in the German government is this used? RKI or BMG perhaps?
How can one model (or simulate) what one do not fully understood?
Remember that simulation of covid from UK with all the pomp and bravado? Not even close to reality.
By the way, all simulations have the same relationship to reality as cartoons. This is just philosophy 101.
"The best and brightest built a state of the art forecasting model and it says I should do X" sells much better than "I assume this year will be exactly like last year and last year we should have done X".
Basing your decisions on complicated scientific methods inspires confidence in many people, even when there is no benefit.
It also mitigates responsibility for decision makers, because they can claim that they would have made the perfect decision if only the model's predictions would have been perfect as well.
> How can one model (or simulate) what one do not fully understood?
On the contrary. When you fully understand something, there's no longer a need for a model.
IMHO “King of the Hill” is a pretty accurate representation of reality
Why have you used UG4, any specific reason? Would like to know, how this software works and how can we use in our country?
find_package(Threads), people...
Yeah there's a lot of bad stuff happening in the CMake config. Further, this doesn't appear to be used by the government. At least not RKI.
What's the issue with that?
The CMakeLists.txt is very poor. Just use find_package(Threads REQUIRED) and do target_link_libraries(... Threads::Threads), and use target_compile_features or target properties to set the C++ version. Right now the CMakeLists.txt requires the user to set that, at which point you might be better off with a Makefile...
Living in Germany I have to say that no one predicted anything remotely correct regarding Corona. It is mostly anxiety and panic driven and short sighted and working against people not with the people.
Corona brings live and death questions to the doors of normal people and also plays havoc with our political parties. It is embarrasing to observe to watch them for the most part. Seems for some of them the first problem they really have to work on instead of only talking about it.
Also from Germany, I have to disagree. The models pretty much nailed the 4th wave down to a T. Missed it by one or two weeks. Based on the assumption nothing was dine against it. Nothing was done against it. The models are good, they are just not used to prepare or base decisions upon them.
I agree with hef19898, here's a link to the prognosis of the Robert Koch Institute from July:
https://www.rki.de/DE/Content/InfAZ/N/Neuartiges_Coronavirus...
Yeah, the models in this program are definitely the ones you want to use for figuring out things in this situation.
Ordinary Differential Equation model, Partial Differential Equation model, Finite Difference model, etc.
Let's be clear here. There was the scientific response, and the political response. They do not work well together.
Science worked fine. Politics did not. This is not specific to Germany.
Same here in NL. Scientific data routinely gets ignored during policy making, either because the data did not originate here and we're supposedly somehow special or because it didn't fit the preferred outcome. The results are predictable.
> no one predicted anything remotely correct regarding Corona.
That is simply not true. The RKI (Germany's main institute for disease control and prevention) almost exactly predicted the current Covid numbers back in July [1]. The top left of the picture on page 7 shows the prediction under the assumption that 65% of Germans get vaccinated; currently 68% of Germans are vaccinated.
[1]:
https://www.rki.de/DE/Content/Infekt/EpidBull/Archiv/2021/Au...
I fail to see how their prediction was "almost exactly correct."
The prediction expected a bell curve peaking at the end of December, at around 400 7d incidence, assuming ~65% vaccinated.
We are at then end of november, ~68% vaccinated, and the 7d incidence is already above 400, and still rising.
We can only hope that they got the shape right, the timing slighly off and the peek not too off.
Where are our goalposts?
I personally think that for such a chaotic, human system, a prediction from six-months previous is pretty good if it's within a few weeks and probably less than an order of magnitude.
Probably depends on the input parameters. Lots of calls for freedom days and such as well as a general removal of measures probably changed the model conditions considerably.
You are confusing what the politician said with the models.
Even now parties like the FDP are blocking the possibility of lockdowns just for the sake of it and not based on any scientific model.
As southern European and living here for about half of my lifetime, the whole management reminds me of the usual mess of Portuguese governments.
I actually never expected that somehow my home country would be able to manage better at decisions, instead of the yes/no/but-elections-are-coming that we had here.
Shitty software written by shitty people.
I hope you all hang for your war crimes at the upcoming tribunal.
Get fucked commies.
Commies are really, really rare in Germany. I find it personally doubtful that so many are involved in that research area.
Also try to assume that they are acting in good faith.
"_Don't feed egregious comments by replying; flag them instead._"
a.k.a. please don't feed the trolls
https://news.ycombinator.com/newsguidelines.html
Commies are not rare in Germany, they just don't call themselved commies.
We have communist parties in Germany such as the Marxist–Leninist Party of Germany or the German Communist Party. However they have no almost support. Communism is about the ownership of the means of production as in people are not allowed to own factories. You will find almost no German who supports that. Not even the Linke supports that and instead hides behind a nebulous "overcoming capitalism".
What's wrong with commies?
Taking something from someone else is wrong. Doing it for the good of others doesn't correct this theft. The degree that this happens in a society represents the level to which civilisation is broken and corrupted.
Open source examples aren't communism, there's no threat of force. It relies on decentralized collaboration and natural willingness to create and share. There's a solid legal framework in place that promotes use, improvement and extension with low cost to all parties.
So I guess Nestle is communist? People who took land from natives are communists as well. In a way, every capitalist corp which takes resources from other countries is communist.
If you think communism = taking away things from people, I guess you don't know political ideologies that well :)
So would you say all things funded through taxation are wrong ? Not trying to be snarky, just wondering if that's the logical conclusion to that axiom.
No. I would argue that less taxation would be needed if anti competition and regulatory structures were better enforced and supported.
well that's one way to see it.
however, in germany, we don't want to have homeless and extremely poor people, so we solve it by taking money away from the rich and middle class. for some people it might be called communism, for germans and in general europeans however that's just the baseline they expect from a functioning state.
Are they that rare? 5 percent of the vote goes to Die Linke, 1 in 20 isn't exactly non existent.
Die Linke is not a communist party. German communists are called DKP (15K votes in the last election) and MLPD (23K votes).
There are certainly many communists in Die Linke, even if they are a "big tent" left party.
If you count trotskyists among communists there might be some, but nowhere near the majority
Can't be very good, or the government doesn't use it.