Coding4Engineers Book feedback

[Mech-eng-au]

Hi Lin,

The book is really coming together! Every time I read a chapter, my understanding of both OOP (object-oriented programming) and C# improves, and I become more proficient with PicoGK.

I thought it would be helpful to open this discussion so that the computational engineering community can provide feedback and help you proofread the Coding4Engineers book.

Here's an example from your latest chapter (11):

When writing the constructor for the BasePlate class, you use Vector instead of Vector2. This will result in a CS1729 error when Fixture.BasePlate is called (see screenshot below).

As I'm new to programming, I apologize if my terminology isn't accurate. I hope that it makes sense anyway.

[LinKayser]

Thank you — I fixed the mistake in the book example!
Lin

[ahuller1991]

Thank you for creating and making open-source both PicoGK and this book. I sincerely agree with your philosophy on current CAD software vs. physics-driven design. I have made it through the chapters you have posted thus far, and I have been able to create a rocket nozzle using a combination of NASA CEA, Compressible flow, and Rao Contour equations. I still have a lot to learn (I am not proficient in C#) but this has been a lot of fun, and I can see the power in this capability. I am looking forward to future chapters of your book!

[digitaldoctor]

Greetings, ALL and wishes for a most productive New Year.

I would appreciate feedback on a line of though that I'm parsing, as it pertains to Computational Engineering and C#. The line is driven by an earnest interest in democratizing awareness of, and access to, the likes of CE.

1. in the age of AI, is it absolutely necessary to go through the gauntlet of learning C# from the ground up?

2. is there a sound skeletal understanding of CE principles that can be had, and then the hardcore coding be offloaded to a C# coding assistant?

3. if the answer to question 2 is yes, then what does that fundamental knowledge base look like? Can it be presently so clearly and simply that high school and community college students master it?

Looking forward to constructive responses.

Rafiki Cai

[digitaldoctor]

It may have been better, if I had read the Preamble to the CFE book. LOL

 "you should **not** attempt to encode logic into an LLM". (Lin Kayser)

This seems to be exactly what billions of dollars are being deployed to do. There can be no AGI
without logic being wired into LLMs.

If the basic workload of logic is to process information systematically, draw valid conclusions, and avoid contradictions,
why shouldn't this capacity exist in LLMs.

[LinKayser]

Logic should be encoded in computational models, they are computationally light, and always give you the right answer. You can bombard an LLM with many petabytes of question-answer pairs, and it may still give you a wrong answer at times. Moreover, you cannot be sure that it will give you the right answer. LLMs are great for things that are not easily explainable through logic (natural language, "good" driving, why one person in a picture is a man, and another a woman).

The people spending the billions largely know that and are looking at other ways to build general intelligence. ChatGPT, for example, now synthesizes answers as code, and then runs the code (how many days to Christmas?), which is an interesting way to do it, but probably not the final way. If you read some of the public writings of Yann LeCun, you can follow these arguments.

Stephen Wolfram also has some good takes on it. I have summarized some of it here:

https://jlk.ae/2024/01/01/on-tribal-knowledge-llms-and-computational-knowledge-in-engineering/

Best,
Lin

[digitaldoctor]

TYPO In FUNDAMENTALS (Second Paragraph)

" that we will built upon in the following chapters""

[eherrmann8]

The book is definitely helpful. I am looking for a comprehensive list of functions (and examples) within PicoGK & ShapeKernal so that I don't have to surf through the code to figure out what is available outside of what you have already put in the chapter examples. Also I'm looking for functions like volume, volumetric center and moment of inertia for a given closed voxel object.

[eherrmann8]

I found the CalculateProperties function which returns the volume of the voxel field. Can this be extended to also return the volumetric center and moments of inertia about the local frame coordinates?

[dvd101x]

Hi, thanks for the book.

I found a typo in Chapter 4: Classes: Class Definition last sentence.

Now, it’s time to instatiate our first object!

Should be:

Now, it’s time to instantiate our first object!

Missing an "n"

[dvd101x]

Hi, just a few more typos

phrase	fix	to	source
it should’t work	should’t	shouldn’t	Preamble: My Story
I found it extremly hard	extremly	extremely	Fundamentals - The skeleton of a class
one very straighforward one	straighforward	straightforward	Fundamentals - Constructors
Let’s do the straighforward one first	straighforward	straightforward	Fundamentals - Constructors
to expose the mimimal interface	mimimal	minimal	Fundamentals - Visibility
compiled for different archictures	archictures	architectures	Running code
And basically everyhing in our current	everyhing	everything	Inheritance - Summary
to form a class hiearchy	hiearchy	hierarchy	Inheritance - Summary
the name of the function and it’s signature	it’s	its	Interfaces - Has a…, is a…, behaves like a…
Slighly confusingly, there is no	Slighly	Slightly	Interfaces - Has a…, is a…, behaves like a…
being a class of the interited type	interited	inherited	Interfaces - Summary
you can stil agree	stil	still	Design an aircraft in an afternoon - Designing an airplane using interfaces, placeholders, and stubs
maybe we should just used a bool value	should just used	"should just have used" or "should just use"	Design an aircraft in an afternoon - Designing an airplane using interfaces, placeholders, and stubs
attachment point and and mechanical	and and	and	Design an aircraft in an afternoon - Geometry is the last thing on your mind
wings were just flat recangles	recangles	rectangles	Design an aircraft in an afternoon - Building an airplane
something that is often suprisingly difficult to do	suprisingly	surprisingly	Design an aircraft in an afternoon - Summary
the code you are executing is interupted immediately	interupted	interrupted	First steps in PicoGK - Catching exceptions
Building a fixture is not a trival task	trival	trivial	Let’s build a computational fixture maker (Part 1)
make sure it sits at a specifed height above	specifed	specified	Let’s build a computational fixture maker (Part 1) - Loading and positioning a mesh
you should keept it alive	keept	kept	Let’s build a computational fixture maker (Part 1) - Cleaning up our namespace with nested classes
But we we can make	we we	we	Let’s build a computational fixture maker (Part 3) - Drilling holes
and then create the modfied version	modfied	modified	Let’s build a computational fixture maker (Part 3) - Drilling holes
We have alread worked with all	alread	already	Computational geometry (Part 1)
Of coutse, In a physically correct manifold	coutse	course	Computational geometry (Part 2)
let’s have the actual implementaton	implementaton	implementation	Computational geometry (Part 4)
	No link for "Modulating a grid"	Add link for "Modulating a grid"	Computational geometry (Part 4)