Interesting Links for 26-07-2023

[andrewducker]

1. A Growing Share Of Americans Think States Shouldn't Be Able To Put Any Limits On Abortion

(tags:usa abortion polls )

2. Labour drops pledge to introduce self-ID for trans people - but will still streamline the process, removing the need for a panel

(tags:lgbt transgender labour politics )

3. "What I've learned about having sex with ADHD (and why you should never go nude trampolining)"

(tags:adhd sex )

4. There is no good data on whether attention spans have been declining

(tags:attention psychology research fail )

5. Scottish government approval for £500m hydro scheme

(tags:Scotland electricity hydroelectric GoodNews )

6. To understand Musk's renewed obsession with X you need to understand his history with it

(tags:internet ElonMusk history business finance )

7. On Monotropism, Autism, and ADHD

(tags:autism adhd research )

8. Programming Advice: Before you try to do something, make sure you can do nothing

(tags:advice programming )

9. Wagner gold smuggling critical to keeping Russia's economy afloat

(tags:Russia gold Sudan )

Comments

[cmcmck]

1. Being I was long post op when the system came in I only needed one 'expert' and no panel so it would be an improvement if it all went that way if not ideal.

I still think a lot of the politicos hold to the 'two kinds' theory which goes back an awful long way- those who 'mean it' (ie are heading for GC surgery) and those who are 'playing at it' (ie everyone else).

[andrewducker]

I think a lot of othem are just panicking because The Daily Mail might say something mean about them.

[fub]

8. One could argue that software framework have gotten too complex -- so complex, that making something that does nothing requires actual software engineering work. That should not really be a thing.

[altamira16]

There is too much going on in that article about attention. I question their definition of attention.

Some of the definitions of inattention are suspect as well. Doodling and fidgeting were considered inattentive behaviors in one study in that article, and these things may be helping us retain information long-term. Alternating or divided attention seems like it is the opposite of attention.

Some of the tasks being measured as "attention" are simple ones that are measured quickly. Are these short-term memory tasks really measuring attention?

[andrewducker]

If you want, say, a service that doesn't do anything, then you still need all the bits that make it a service (as opposed to a console app or a GUI, or some other kind of application). And so it makes sense to make sure you have gotten that right before you start making your service do something. Because it might look like "nothing" to you, but there's a bunch of stuff that's necessary in there.
And, as far as I can tell, all of that stuff is actually necessary - it's not something which could be left out.

[dewline]

I need to revisit this later. I worry about attention span issues for various reasons these days...

[simont]

My corollary to #8 is: before you try to change something, make sure you've tried changing nothing.

An obvious case (and yet I do see people forget it) is: clone someone's git repo, check you can build it successfully from a pristine state, and only then, start making changes toward the feature / bug fix you want. If you omit the middle step, you can confuse yourself into thinking that a build failure was the fault of your changes, when it wasn't.

Another case happens in distributed debugging. Bug reporter says: I put this input into the compiler and got that output. Investigating engineer reads the report and says: ok, let's try this modified input and see if that sheds any light. Fine – but first make sure you still get the reported output from the original input, because if a change in the meantime has made that no longer true, then you'll end up comparing (input,output) pairs from two importantly different compilers, and further confuse yourself!

I learned this at my dad's knee, more or less literally. He started off as a physicist, and when he moved into computing, he brought with him his existing attitude to experimental control, including the general principle that a set of differential observations (inputs differing this way produce outputs differing that way) are always suspect if they weren't all observed in the course of the same experiment, because variation between experiments (or between experimenters) can so easily introduce changes bigger than the phenomenon you were trying to observe. And my own life experience provides plenty of support for that policy in computing as well as in physics. But I think people who didn't have that drilled into them as a child are more prone to see replicating the previous observation as a waste of time, and skip it wherever possible.

[calimac]

1. It would be nice if people would stop thinking that allowing late-term urgent medical abortions were a stalking horse for "casual" discretionary abortions.

6. I read a newspaper article (probably behind a paywall, so I'm not going to look for it) telling the same story, so that helps to give it credibility. I find it hard to imagine a circumstance in which Peter Thiel is the good guy, but I suppose that when you're faced with Hitler you have to ally with Stalin.

Remember also that Musk has another company called SpaceX. He just loves him some X, don't he? Maybe he should scratch his itch by changing his middle name to Xavier - he can say it's for Professor X if he wants. "Elon X. Musk" - sounds good, no?

[andrewducker]

Don't forget the Tesla model X.
To go with the novel S and the model 3.

Because he is, apparently, 12.

[calimac]

At least he's older than Donald Trump, who is usually cited as being 4.

[altamira16]

I have read at least two books on this topic this year.

[fub]

Yes, I get that. But ideally, you just had a button in your IDE (or a single statement in your language) that says: "this is a console app" or "this is a service", and the system would take care of it. The code to make it an console app is, arguably, boilerplate because it does not do anything (other than define this piece of software as a console app), and yet it still requires software engineering.

(This is also why I love RPA: it allows you to do "software things" without requiring you to be a software engineer. It can fill a lot of gaps that we simply don't have the development capacity for to fill with 'proper' software.)

[andrewducker]

Oh yes. One button to create it, I agree. Visual Studio does that.

But then, once you've done that, test it, to make sure that it works at all. And then do something simple to make sure that *that* works.

[jducoeur]

To be fair, there are plenty of fancy modern platforms where a "Hello World" program is still only a few lines of code. They sometimes compile to a megabyte of output because there are so many libraries involved, but the actual source code can be pretty concise.

[fub]

That is how it should be! And yes, running a service is complex, so it's only natural that there are a lot of libraries involved -- but all that complexity should be taken care of by the libraries, not by the software engineer! (In most cases, obviously.)