OT: Review of Ferguson/Imperial model

[Vir Campestris]

<snip>

We do have a reliable antibody test.

What we don't have is a reliable _cheap_ one we could use on large numbers.

Downthread there's a pointer to this having been done elsewhere.

Andy

[Pancho]

That's not really fair. Modern software engineering practices are mainly developed for teams working on big complicated projects.

With such a little project I can't see any particular objection to a single file. A programmer can know 15,000 lines intimately.

As long as they regression test, and tag release versions it should be fine.

I kind of suspect he wasn't ready for the attention and tried to clean up his code before releasing it and introduced problems doing so.

[Paul]

Do you know who John Carmack is ?

You've probably heard of Doom and Quake.

He's a genius at getting amazing performance from scrawny video cards.

He blew a fair bit of his resources on this.

*******

I could write a long reply about simulators, their uses and abuses, but what would be the point ? Everyone here has already made up their minds, and wants to scapegoat the guy.

I won't spoil your fun.

There have been successful tech transfers, of dusty decks from universities, to the "real world", but this isn't one of them. It can take a long time (twenty years), to take something worth having, and beat it into shape so you can have a copy for a reasonable sum from a commercial source.

And you don't have to write a simulator from scratch. There are simulation environments you can buy, and you can create a series of configuration files on top of that. That can solve some of the provisioning of base logistics in your simulator. Now you can be "half-blamed" for the mess.

I was in a meeting room at work one day, the presenter was late, and we were shooting the shit. For a giggle (because the God of Simulators happened to be in the room), I asked for a show of hands, how many people had written simulators of their own. Because I wanted to get the God of Simulators engaged. I was shocked, when almost every person in the room, put up their hand. The dirty dirty habits of professionals :-) Too funny. Now, if I hadn't asked that question, I would never have known precisely how many dusty decks were out there. And, just at work.

The trick then, is to not give your deck away, nor use it for critical purpose, without a lot (and I do mean a lot) of extra work. The person who has done this, deserves all the abuse he gets, because he's broken one of the cardinal rules. There's a way to be shed of stuff like this, without getting blamed for every fart that comes out of it. But it also means, it could take

20 years, for the ideas in your deck, to be exposed to the light of day.

SPICE for circuit simulation (electrical engineering) is an example of a deck that successfully made it to the real world. ( Anyone who has used it, is entitled to make their ugly face now :-) )

Paul

[michael adams]

As was explained in those sections of my post which you chose to snip, the suspicion of a hoax and/or a politically motivated fraud arises immediately you come to compare the damning criticisms of the original code in the lockdownskeptics link as made by an anonymous critic as posted by Tim Streater and the comments made concerning that very same code by John Carmack who was a member of the GitHub/Microsoft team who actually did the conversion

To wit:

"Vir Campestris" [e-mail]

An anonymous poster wrote -

Whereas John Carmack who was a member of the GitHub/Microsoft team who actually did the conversion, and who writing under his own name wrote of that very same code

the original code

<quote>

" fared a lot better going through the gauntlet of code analysis tools than a lot of more modern code"

and

" Similarly, the performance scaling using OpenMP was already pretty good, and this was not the place for one of my dramatic system refactorings."

</quote>

So again, in the absence of any more substantial evidence* it's simply a question of who to believe ?

An anonymous poster on a UseNet DIY group read by around 100 people at most, of whom about 5 would have any sort of clue about what he was talking about and with no reputation to defend ?

Or a GitHub/Microsoft employee posting on a publicly social media platform read by thousands maybe, and under his actual name ?

cue:Harry Hill

There's only one way to decide !

Fight !!!!

michael adams

...

*An unvarnished "argument from authority" is a logical fallacy of the first rank when there is already sufficient evidence or argument to decide a question either way. Howver in the absence of any such evidence, or where deciding the question would be far too complex a matter, then an argument from authority can be regarded as a legitimate compromise

[Paul]

Are you sure the word wasn't "refactoring" ?

Chopping it up into modules. 50 files of 100 lines each, instead of 5000 lines.

Making subroutines out of code chunks that are used more than once (parametrization).

So far, I'm not seeing where Microsoft are involved here. I saw the name John Carmack, who was a software designer at IDSoftware and worked on Doom and Quake and the like. You can think of him as a "STEM guy" with super coding skills (like recognizing that some fossil code came from a Fortran deck).

The file I looked at, had a copyright of 2004-2020, or 16 years. And somewhere I saw a claim this code was 30 years old ? Then perhaps it was a Fortran deck at one time, and got ported (translated) to another language at some point.

The code is not exactly "generic" looking. It's quite specific to a particular task. The notion of it being "The SIMS (without graphics)", is pretty close. With graphics it would look like John Conways "Game of Life" (which you wouldn't appreciate unless you coded your own copy when the idea came out in Scientific American long ago). Today, when people see that as a screensaver, they haven't a clue what it is.

's_Game_of_Life Paul

[tim...]

there's no problem with it being a single file

if you'd said 15,000 line single module, I'd agree

but there no reason why you can't store a multi module solution in a single file (assuming only one person is required to work on it)

tim

[newshound]

Isn't the other point that the airwaves auction was designed by a *very* smart economist. I think economists get a bad press. I know that we are all ultimately paying for that spectrum, but it did generate a pile of cash for the government.

[Pancho]

What do you mean module? This is C code

About four people on it this month. The git history is a little too busy. Normally you'd only expect to see working code commits in the logs.

[tim...]

is it, I had assumed from previous comments that it was something akin to an excel macro or VB solution

but function is the word you want

tim

[Andy Burns]

Did it though? I thought the winning bidders used the large cost to make a loss on the books ...

[Vir Campestris]

I rather think Pancho isn't a software engineer.

I know a darnn sight more about software than I do about DIY - which is why I came here in the first place.

Andy

[Andrew]

Overpaying for a Kew Gardens 50p piece (before they became collectors items) could have been proftable though.