Julia YouTube channel is invaluable place to learn about Julia language, ecosystem, etc. Since I quite pedantic I from time to time take notice of small typos in mistakes in titles of videos and descriptions. Wrong spelled name of author of talk, missing hyperlink in the description, some typos things badly formatted, etc.
I recently try to learn how to makes timestamps for Julia videos and I find some obvious mistakes in them, I guess they my be a reason why some videos don’t have good progress bar on YouTube (more here). I wonder if is a sense in correcting some other mistakes on Julia YT channel, when I should send these ones that I spotted? Maybe this mistakes aren’t worth correcting, but I don’t know it.
On JuliaCommunity GitHub I didn’t see repository dedicated to that, so I guess it is better to ask about here.
Thank you, I already open an issue about timestamps. Logan Kilpatrick said he will do something with them. My questions is about other errors, like misspelled name of the speaker or something like that.
I understand it very well. Big mistakes are avoidable, small always find their way out.
My English is bad and not very good, so I can’t help with improving text. I can catch some pretty obvious mistakes.
Native Elementary Functions in Julia, https://www.youtube.com/watch?v=B03F6EFUm78. It is classical example of small mistake. In the title we have name of the speaker “patrick Kofod Mogensen” and his name of course should be written as “Patrick” (I check it on the slides).
I have one more thing to ask. Videos from first (?) JuliaCon 2014 are all collected into playlist, but most of them don’t have in the title name of the speaker nor “JuliaCon 2014”. They are from 2014, so from the times of middle Roman Empire (at Julia time scale), but maybe changing their title by adding name of the speaker and “JuliaCon 2014” is worth an effort? From what I seen most speakers names are stated in the descriptions of videos.
Here is example of video description, that can be made better looking. I already make timestamps for it, but since they are not incorporated into description when I write this, I will not use them. Also, I don’t use Markdown (?) quotation, because it often smooth over rough parts that I want to show.
BEGINING
Many numerical tensor manipulation packages exist (e.g. Einsum.jl), but treating tensors at a purely numeric level throws away a lot of potential optimizations.
Often, it’s possible to exploit the symmetries of a problem to dramatically reduce the calculation steps necessary, or perform some tensor contractions symbolically rather than numerically.
SymbolicTensors.jl is designed to exploit these simplifications to generate more efficient input into numeric tensor packages than you would write by hand. It based on SymPy.jl, sympy.tensor.tensor, and ITensors.jl. Time Stamps:
00:00 Welcome! 00:10 Help us add time stamps or captions to this video! See the description for details.
END
Changes are obvious.
BEGINING
Many numerical tensor manipulation packages exist (e.g. Einsum.jl), but treating tensors at a purely numeric level throws away a lot of potential optimizations. Often, it’s possible to exploit the symmetries of a problem to dramatically reduce the calculation steps necessary, or perform some tensor contractions symbolically rather than numerically.
SymbolicTensors.jl is designed to exploit these simplifications to generate more efficient input into numeric tensor packages than you would write by hand. It based on SymPy.jl, sympy.tensor.tensor, and ITensors.jl.
Time Stamps: 00:00 Welcome! 00:10 Help us add time stamps or captions to this video! See the description for details.
END
I suggest to change YT title of this talk to “State of Julia | JuliaCon 2021”, since we have already one talk of this name and hopefully we will have many more. You can find my propositions of timestamps for this presentation here.
I have different issue with this video. You can see opening slide “JuliCon Baltimore 2019” for first 3m 33s, while whole video has 14m 59s. It will be to nice to cut out first 3m 20s, after that I will try to make timestamps for it. But, I can’t promise anything.
Abstract now have a form (no Markdown, it can smooth things over)
BEGINNING
Charge transport and reactions in electrochemical and semiconductor devices are described by systems of nonlinear PDEs for the electric field and the diffusive and convective movement of charged particles.
A finite volume discretization approach turns these into a system of nonlinear algebraic equations. The Julia package VoronoiFVM.jl provides an infrastructure or this discretization approach and uses forward mode AD to set up the sparse Jacobi matrices for the solution with Newton’s method.
END
I propose changing it to
BEGINNING
Charge transport and reactions in electrochemical and semiconductor devices are described by systems of nonlinear PDEs for the electric field and the diffusive and convective movement of charged particles. A finite volume discretization approach turns these into a system of nonlinear algebraic equations. The Julia package VoronoiFVM.jl provides an infrastructure or this discretization approach and uses forward mode AD to set up the sparse Jacobi matrices for the solution with Newton’s method.
END
This is poster video not even 2 minutes long, but as showcase I made timestamps for it.
One small thing. In abstract we have “up and downsample signals”, maybe more correct is “up- and downsample signals”. Someone with better English that I should decide.
I think we should replace line in the description
“Keynote ( Soumith Chintala): PyTorch and my journey in open source”
with
“Keynote by Soumith Chintala: PyTorch and my journey in open source”.
My first proposition is to change the title to State of DataFrames.jl | Bogumił Kamiński | JuliaCon 2021. After that I will to slightly improve it description.
Traditionally, here is my proposition for timestamps.
BEGIN
In recent years, we have seen a large body of research using hierarchical
matrix algebra to construct low complexity linear solvers and preconditioners.
Not only can these fast solvers significantly accelerate the speed of
large scale PDE based simulations, but also they can speed up many AI and
machine learning algorithms which are often matrix-computation-bound.
On the other hand, statistical and machine learning methods can be used
to help select best solvers or solvers’ configurations for specific problems
and computer platforms. In both of these fields, high performance computing
becomes an indispensable cross-cutting tool for achieving real-time solution
for big data problems. In this talk, we will show our recent developments
in the intersection of these areas.
END
In the text above we have unnaturally short lines. I propose to change it to the version below.
BEGIN
In recent years, we have seen a large body of research using hierarchical matrix algebra to construct low complexity linear solvers and preconditioners. Not only can these fast solvers significantly accelerate the speed of large scale PDE based simulations, but also they can speed up many AI and machine learning algorithms which are often matrix-computation-bound. On the other hand, statistical and machine learning methods can be used to help select best solvers or solvers’ configurations for specific problems and computer platforms. In both of these fields, high performance computing becomes an indispensable cross-cutting tool for achieving real-time solution for big data problems. In this talk, we will show our recent developments in the intersection of these areas.
END
Today I will complaining about timestamps that I made myself. It is quite fitting, after all if I criticize what others did, I should also criticize what I did wrong.
Now we have in timestamps
21:34 Problem of "X is more specific that Y. Example 1
23:32 Problem of "X is more specific that Y. Example 2
I probably should write something like text below.
21:34 Problem with “X is more specific that Y”. Example 1
23:32 Problem with “X is more specific that Y”. Example 2
But someone more skilled with English should judge which form is better. In any case, closing quotation mark need to added.
BEGIN
LoopVecorization.jl can be thought of as treating loops like a familiar DSL for specifying dependencies between operations (such as arithmetic and loads or stores) and loops, without regard to any order aside from that inherent in the dependency chains.
The library has infrastructure for modeling the cost of evaluating a loop nest using different orders of the constituent loops, and different unrolling and blocking factors of the loops.
The advantage is demonstrated in allowing writing high performance code that is generic with respect to the data layout of the underlying arrays, with the order of evaluated loops and data access pattern shifting in response to transposed arrays without any change in the user’s code.
END
I think we can format it in better way.
BEGIN
LoopVecorization.jl can be thought of as treating loops like a familiar DSL for specifying dependencies between operations (such as arithmetic and loads or stores) and loops, without regard to any order aside from that inherent in the dependency chains. The library has infrastructure for modeling the cost of evaluating a loop nest using different orders of the constituent loops, and different unrolling and blocking factors of the loops. The advantage is demonstrated in allowing writing high performance code that is generic with respect to the data layout of the underlying arrays, with the order of evaluated loops and data access pattern shifting in response to transposed arrays without any change in the user’s code.
END
Apparently youtube is now using OCR, e.g. on slide titles, to automatically add titles to timestamps. But it probably won’t do that if we have the “help us add timestamps” placeholder timestamps. Not all videos use slides, but for those that do it could be helpful to rely on youtube’s automation.