Grading papers

[Meat Engine]

I just have to grade engineering stuff, and it makes me want to drink. I can't imagine having to grade composition like Dusty. I would probably want to walk into traffic.

[Verbal]

Lets hope you taught them well and they all got everything right.

[Shredlord]

[quote name="Meat Engine"]



I just have to grade engineering stuff, and it makes me want to drink. I can't imagine having to grade composition like Dusty. I would probably want to walk into traffic.[/quote]



I have to grade compositions--mostly literary analysis from high school kids. Luckily, 4 of my 5 classes are high level classes, but it's still draining.

[Meat Engine]

[quote name="Verbal"]Lets hope you taught them well and they all got everything right. [/quote]



This semester I didn't really teach. I did teach quite a bit during the summer, though, and it was just as bad.

[Lavabug]

Yeah that looks like hell.



Try getting to teach a proof-based mathematics course. Grading will take seconds.



Fluid mechanics. I have a fluid dynamics course coming up.

[Verbal]

reynolds number?

[Lavabug]

[quote name="Verbal"]reynolds number?[/quote]

and Navier-Stokes equations for viscous flow. We should be seeing some magnetohydrodynamics too (plasmas), seeing how the course is oriented towards astrophys (and the prof is a prominent solar physicist).







The scribbles above are basically "F = ma", no joke.

[Verbal]

that isnt showing up at all, but i wont understand it anyway



when i did applied maths in school almmost every question was just f=ma

[Lavabug]

[quote name="Verbal"]that isnt showing up at all, but i wont understand it anyway



when i did applied maths in school almmost every question was just f=ma [/quote]

Its missing a white background, opening it in another tab probably works. It is mass*accel = all the forces on a fluid (gravity, friction etc), per unit volume, hence the rho(density) instead of a m on the left.



F=ma works for everything.

[Meat Engine]

Reynold transport theorem basically gets you F=d(mv)/dt if the change in linear momentum of a control volume is investigated. That is good for looking at the reaction force of a rigid body under the influence of a fluid stream.



The great thing about Navier-Stokes is that a lot of terms drop out for "simple" systems. The most important one is the time rate of change of velocity term. If that doesn't drop out, you get a PDE instead of an ODE, and must proceed with Laplace transforms, similarity methods, or a handful of other equally distasteful methods.



Inviscid flow is another nice assumption. You end up at the Stoke equation.



We finished up with boundary layer theory. I hated it.

[Verbal]

nobdoy understands that except lavabug

[Hallucinating dead]

[quote name="Meat Engine"]Reynold transport theorem basically gets you F=d(mv)/dt if the change in linear momentum of a control volume is investigated. That is good for looking at the reaction force of a rigid body under the influence of a fluid stream.



The great thing about Navier-Stokes is that a lot of terms drop out for "simple" systems. The most important one is the time rate of change of velocity term. If that doesn't drop out, you get a PDE instead of an ODE, and must proceed with Laplace transforms, similarity methods, or a handful of other equally distasteful methods.



Inviscid flow is another nice assumption. You end up at the Stoke equation.



We finished up with boundary layer theory. I hated it.[/quote]



They teach "Navier–Stokes equations" to the students of 11th grade here in india. I remember i studied it last year, because of the over pressure of bullshit education system. Fuck the teachers and the way they teach

[Lavabug]

I vaguely recall the Reynold's formula from my "mechanics & waves" course, I think it was related/some form of the continuity equation. But we arrived to F = ma just considering all the forces acting on a fluid in compact vector notation(gradient of grav potential, friction etc), and rewriting acceleration as its two components out explicitly. Indeed, considering inviscid and irrotational/conservative flow, you could arrive to other famous equations like Bernoulli's (the one that describes Venturi's tube). With the continuity equation in conservative flows, you could get Poisson's and so forth. I thought that was cool.



In my course we will be combining these equations with Maxwell's field equations, should be good.



There's like a million dollar prize for whoever finds an exact general solution to Navier-Stokes equations, in their most general form.

[Lavabug]

I got through my 1st year thanks to tipler's, one huge ass problem book, and Walter Lewin's lectures. (I attended a distance uni, so no profs). I'm actually glad I did it that way, as I really learned how to look stuff up on my own.

[Lavabug]

Tell me about this parallel computing business, mg. What kind of stuff do you do, do you get any hands-on experience on setting up comps for parallel processing?

[Meat Engine]

[quote name="Lavabug"]Tell me about this parallel computing business, mg. What kind of stuff do you do, do you get any hands-on experience on setting up comps for parallel processing?[/quote]



If your computer has a multi-core processor, then it can handle OpenMP stuff already. That is parallelism in shared memory. To use MPI, one needs a Unix cluster, and mpirun/mpiexec installed. MPI is parallelism across multiple nodes with no shared memory between them. Then there is also grid computing, for which software is available. It is a basic matter of installing a software package on the machines to be linked in parallel. When you hear about a large scale computing project asking people to turn over spare clock time, it means they are using grid computing.



Writing the scripts for parallel computing is the pain in the ass. You have to look at your algorim, find independence, and work around dependencies. You also need to take care to minimize message passing between nodes so as not to lose the time you gained by distributing the work in the first place.

[Shredlord]

OK, this thread is now giving me a headache.

[Meat Engine]

[quote name="Lavabug"]I vaguely recall the Reynold's formula from my "mechanics & waves" course, I think it was related/some form of the continuity equation. But we arrived to F = ma just considering all the forces acting on a fluid in compact vector notation(gradient of grav potential, friction etc), and rewriting acceleration as its two components out explicitly. Indeed, considering inviscid and irrotational/conservative flow, you could arrive to other famous equations like Bernoulli's (the one that describes Venturi's tube). With the continuity equation in conservative flows, you could get Poisson's and so forth. I thought that was cool.



In my course we will be combining these equations with Maxwell's field equations, should be good.



There's like a million dollar prize for whoever finds an exact general solution to Navier-Stokes equations, in their most general form.[/quote]



I like irrotational and conservative flows a lot, because then you get into the regime of writing the flow as a complex quantity, the real part being potential flow and the imaginary part being the streamfunction. From there it is easy to get the velocity field.

[Verbal]

A major problem in Physiology texts is they are describing very complicated things with simplified algebraic formulae (that they just introduce without any notion of an explanation). The explanations are not explanations as a result. And the whole thing is incredibly frustrating. They make massive shortcuts. AT LEAST ADMIT YOURE JUST TRYING TO AVOID A CALCULUS DIGRESSION

[Meat Engine]

There are a lot of texts, even the engineering sort, that seek to reduce as much as possible to algebra. It's kind of stupid.

[Verbal]

Maybe some people would read such a text and think they now understand what's going on, but surely 99% of people realise that they are being presented with a hollow sham explanation that has more holes than a fenestrated capillary.

[zim]

i didn't know meat engine was a teacher

we were friends on steam for years but i think he deleted me recently

[morbid board eli]

I haven't graded a paper in three years

[Lavabug]

I've never graded a paper, thanks to my fellowship.

All these years later, I still think fluid mechanics is the coolest thing ever. I've moved on to MHD now.