Is software engineering, well, engineering?

[sheldrake]

Writing software is closer to screenwriting than camera design though.

Computer design is engineering, and I'm sure some software design requires deep knowledge of the physics of computers, but mostly software writing is a combination of arithmetic and the sort of creative writing that results in, say, a Haynes manual or a news bulletin.

There's a whole world of fairly abstract technical concerns which aren't directly solved with physics that software engineers have to understand. Concurrency problems & distributed systems design, the applied mathematics of machine learning or graphics engines etc..

[Martin Y]

So apart from the lighting, cameras, microphones, mixing, recording, editing and the entire broadcast chain and streaming service, what have Gavin and Stacy ever done for us, eh?

This feels like the Dunning-Krugerest thread here so far.

[plodder]

Writing software is closer to screenwriting than camera design though.

Computer design is engineering, and I'm sure some software design requires deep knowledge of the physics of computers, but mostly software writing is a combination of arithmetic and the sort of creative writing that results in, say, a Haynes manual or a news bulletin.

There's a whole world of fairly abstract technical concerns which aren't directly solved with physics that software engineers have to understand. Concurrency problems & distributed systems design, the applied mathematics of machine learning or graphics engines etc..

Not dissimilar to a really complicated plot then, where all the characters need to remain consistent and where they make amazing things happen by talking to each other. JK Rowling would make a good coder. Re machine learning etc I'm happy to concede there are crossovers between proper engineering and programming.

[sheldrake]

Distributed system design isnt anything like craftimg the plot for a novel. In particular, its not like ripping off Ursula K Le Guin and repackaging it in twee British edwardianism.

[plodder]

Lol, there's no copy pasta in coding? Pull the other one.

[sheldrake]

You can just f.ck off now

[Little waster]

If you want to call writing a game engineering, then I hope you'll also embrace the engineering that went into writing & producing the Gavin and Stacey Christmas Special. One line out of place and the whole thing would have fallen down.

Writing games is engineering & art.

The Gavin and Stacey Christmas special performed by robots would be engineering. have at least been entertaining

Gavin & Stacey is Two Pints of Lager and a Packet of Crisps for c.nts.

[bjn]

Writing software is closer to screenwriting than camera design though.

Computer design is engineering, and I'm sure some software design requires deep knowledge of the physics of computers, but mostly software writing is a combination of arithmetic and the sort of creative writing that results in, say, a Haynes manual or a news bulletin.

What does a field need to be to qualify as engineering?

I’ve written software that is used to make features films. It involved a f.ck tonne of mathematics beyond the maths native to computer science. It also involved a huge amount of process to design, create, verify and deliver that software.

[Woodchopper]

Yes indeed.

Claiming that making software doesn’t involve science is a bit like claiming that building structures doesn’t because the bloke who built the garden wall just put it up without much thought.

[greyspoke]

^would also be true if you replaced "science" with "artistic reativity".

[bjn]

^would also be true if you replaced "science" with "artistic reativity".

Software in itself doesn’t involve artistic creativity (in the normal sense of those word). Someone writing a bank transaction system doesn’t get to be creative. A games dev may get to be a bit creative. In my old field of visual effects, we had to understand the needs of the artists using our systems, but were much more focused on making the software work for them rather than doing the pretty pictures.

That said, my friends who did the work on the Interstellar black hole sequence had to marry hardcore physics with computer programming with artistic flair.

[plodder]

They didn't marry hardcore physics in the same way a structural engineer does. What they did was write code that uses physics equations to make it look like things were moving realistically. Arthur C Clarke did the same thing with his stories, just with much less sophisticated technology (pen and paper).

[sheldrake]

No he didn't.

[bjn]

They didn't marry hardcore physics in the same way a structural engineer does. What they did was write code that uses physics equations to make it look like things were moving realistically. Arthur C Clarke did the same thing with his stories, just with much less sophisticated technology (pen and paper).

Seeing as they managed to get several real life scientific papers published on the visualisation of black holes, it was hardcore physics. To do that they even had to solve equations, come up with new ones and everyfink.

Structural engineers just hammer sh.t together don’t they? No maths innvolved with that is there?

[plodder]

look, I’m genuinely impressed, and using commercial visualisation techniques to help scientists is awesome, but engineering is taking science and applying it to solve problems. coding is a tool that can help with this, an important tool, but it doesn’t merit some sort of special status. Writing complex functional code is important and interesting but it’s no different from writing anything else

[sheldrake]

It is different, because you're creating a dynamic system, just as if you'd assembled a circuit or an engine. Computers are very generalised machines that become specific useful machines due to the code they execute. It's a type of engineering where the implementation tool is mathematically formal language rather than welding torch or digger.

[bjn]

look, I’m genuinely impressed, and using commercial visualisation techniques to help scientists is awesome, but engineering is taking science and applying it to solve problems. coding is a tool that can help with this, an important tool, but it doesn’t merit some sort of special status. Writing complex functional code is important and interesting but it’s no different from writing anything else

No it’s not. How many novels are created that involve teams of dozens to hundreds of people working in collaboration with strict processes that involve applied mathematics? Entirely different skill sets, techniques and principles to deliver an completely different product.

It’s like saying the design phase of structural engineering is no different to painting pictures.

PS I have Shelly on ignore so have no idea if you are arguing with him or me.

[bjn]

look, I’m genuinely impressed, and using commercial visualisation techniques to help scientists is awesome, but engineering is taking science and applying it to solve problems. coding is a tool that can help with this, an important tool, but it doesn’t merit some sort of special status. Writing complex functional code is important and interesting but it’s no different from writing anything else

No it’s not. How many novels are created that involve teams of dozens to hundreds of people working in collaboration with strict processes that involve applied mathematics? Entirely different skill sets, techniques and principles to deliver an completely different product.

It’s like saying the design phase of structural engineering is no different to painting pictures.

PS I have Shelly on ignore so have no idea if you are arguing with him or me.

Also, coding is only part of software development, there is a f.ck tonne of stuff around it to make things happen.

Also, computer science is thing, it’s an entire discipline, with journals and papers and departments and everyfink. That science is used to solve a vast range of problems.

[plodder]

It is different, because you're creating a dynamic system, just as if you'd assembled a circuit or an engine. Computers are very generalised machines that become specific useful machines due to the code they execute.

Like a choose your own adventure book. I understand.

Please don’t muddle designing computers to designing and writing code.

[sheldrake]

It is different, because you're creating a dynamic system, just as if you'd assembled a circuit or an engine. Computers are very generalised machines that become specific useful machines due to the code they execute.

Like a choose your own adventure book. I understand.

Please don’t muddle designing computers to designing and writing code.

Designing computers is just like writing code.

[bjn]

It is different, because you're creating a dynamic system, just as if you'd assembled a circuit or an engine. Computers are very generalised machines that become specific useful machines due to the code they execute.

Like a choose your own adventure book. I understand.

Please don’t muddle designing computers to designing and writing code.

Don’t confuse code with algorithms, computer hardware and software are both ways of executing algorithms. You can prove out algorithms in code and then turn that into optimised hardware to execute those algorithms. For example the video decoders in most CPUs. Or the ASICS used for crytpowank mining. Talking of CPU hardware, they are mostly microcode, the hardware needs extremely low level instructions to work. Microcode is below the level of X86 machine code. It can be updated to modify how the CPU works. Then we have the joys of field programmable gate arrays. Prove an algorithm in code, push a button, turn that into a set of transistors that will execute the algorithm. There is no hard line between software and hardware.

[plodder]

Ok, what proportion of coders work on that stuff? All fields get multi-disciplinary at their boundaries, but that doesn't really help with definitions. My outsider impression is that the vast majority of coders do things like remove glitches in websites, and that's not something that needs loads of scientific theory to accomplish.

[plodder]

Here's the syllabus for software engineering at De Montfort : it's loads of modules on coding, and sod all core science.

https://www.dmu.ac.uk/study/courses/und ... gJnyfD_BwE

Here's the equivalent first google hit for civils, this time at Portsmouth. Spot the difference: https://www.port.ac.uk/study/courses/be ... gID7_D_BwE

[sheldrake]

Here's the syllabus for software engineering at De Montfort : it's loads of modules on coding, and sod all core science.

https://www.dmu.ac.uk/study/courses/und ... gJnyfD_BwE

Here's the equivalent first google hit for civils, this time at Portsmouth. Spot the difference: https://www.port.ac.uk/study/courses/be ... gID7_D_BwE

You can't tell from that description how much of 'Computer Programming I' or 'Computer Programming II' in the first year involve algorithm design and complexity analysis, for example.

'Construction Management and Practice' at Portsmouth might mostly consist of how to hold your clipboard to project gravitas, or how to moan about cyclists getting in the way of the van. Hard to tell.

[plodder]

That's pretty much exactly what construction management is, but there are also lots of modules on actual applied physics, like hydraulics etc. That's the difference. Algorithm design sounds suspiciously like "how to write a really tight plot" in a creative writing course. Perhaps some hardcore music theory might also be a good analogy. It's not hydraulics though, or water chemistry, or geotechnics, or structures. Or electrical engineering etc for that matter.

I'm pretty sure a lot of coders are self taught, and that's simply not true for engineers. The closest thing to an informally educated engineer is 10+ years of working experience followed by a very rigorous career appraisal process in order to achieve chartership (and the subsequent authority to self-certify).