Wolfram Mathematica

12:08 AM

@b3m2a1 You were the one with a Notebook -> Markdown converter, right?

2 hours later…

2:28 AM

@halirutan yeah I have one. Works well enough for what I do (I also use it in the paclet server build process). I think @Kuba and @C.E. also have ones that may have features mine does not.

I have a converter that converts from notebooks to WordPress HTML (e.g. the markup that WordPress' editor would have created) but not to Markdown.

I have R.M.'s notebook to Markdown converter if anyone's interested, I'm not sure where to get it online anymore.

@b3m2a1 @C.E. Thank. I guess it would work anyway because the Rubi files contain many math-formulas and I just discovered, that GitHub displays PDF files directly.

2:52 AM

I've been working on the WordPress converter in the last week in fact. I'm going to put up a website where anyone can publish articles on Mathematica. I will be able to accept notebooks, and then my script will automatically convert it to a beautifully typeset article on the website.

I've missed having some place to put articles. I want to have a website that looks great and where the articles look great, so it's something that you can link to and feel good about how it is presented.

rm -rf

3:39 AM

@C.E. bitbucket.org/rsm/senotebook Not updated since 2013, but feel free to copy to Github/modify/whatever :)

@rm-rf ah, there it is. I don't have any use for it right now, but it is where I first learned about how to write converters :)

4 hours later…

7:16 AM

@halirutan if you want to play with mine it's here: github.com/b3m2a1/mathematica-BTools/blob/master/Packages/Web/… and I can write something akin to this for it. (That was made with the converter, btw)

Kuba

@halirutan mine is not perfect either but who knows: github.com/kubaPod/M2MD

7:33 AM

@rm-rf @b3m2a1 @Kuba Hehe.. 3 different implementations :)

@halirutan one nice thing about notebooks is how amenable they are to conversions. Nicer than working with some ill-structured thing like XLSX I'm sure.

@b3m2a1 Yes, and the real good thing is, that notebooks are expressions and we can do whatever we like with it.

3 hours later…

10:07 AM

@b3m2a1 The – or at least one of them – language for manipulating XML is XSLT, and I while I do not know it, there is a chapter in the Mathematica Cookbook called "Transforming XML Using Recursive Functions (à la XSLT)" which describes how parse symbolic expressions using definitions like parse[XMLObject[a_, b_, c]] := .... which to the author is roughly similar to how XSLT works.

And it turns out that the author of Mathematica Cookbook – and this is probably not a coincidence – is also the author of the XSLT Cookbook

It's a great compliment to Wolfram Language that it is "roughly similar" in the way it is used to parse XML as a language that is purposefully designed for this task, I'd say....

2 hours later…

Szabolcs

12:29 PM

@b3m2a1 It might be a good idea to create a separate room for the pacletserver project. Recently I find it difficult to follow all the discussions in chat, which is why I don't always respond. Some separation by topic would be useful.

Szabolcs

1:11 PM

I wonder why this was ignored. I think it is a good question.

Q: Draw circle on ellipsoid

Define a circle $\cal{C}(p,r)$ on the surface of an ellipsoid $E$ in $\mathbb{R}^3$ to be the set of points on $E$ whose shortest geodesic distance from centerpoint $p$ is $r$. Let me assume that $r$ is small enough so that $\cal{C}(p,r)$ does not self-intersect. Given an ellipsoid centered on ...

I mean, it seems clear that the OP could work out the math, but what he's looking for is a more convenient, simpler numerical solution made possible by the myriads of build-in features of Mathematica.

Kuba

1:53 PM

@b3m2a1 @Szabolcs are there lists of contents associated with live ceo-ing videos?

1 hour later…

kirma

2:54 PM

Probably not the wisest choice of naming: techzim.co.zw/2018/06/…

Google suggested this piece of news to me, and I had to scratch my head for a second...

@kirma The smells like onions and lawsuit..

kirma

3:16 PM

@halirutan Sort of, but I also guess WRI could easily consider business interests in Zimbabwe next to impossible to defend, at least without great loss of capital.

@kirma Yes, quite likely.

@C.E. yeah this is how the Transmogrify package works. The problem with something like XLSX is that it’s got multiple XML files attached to it and each of these only describes part of the document so one really needs to go through them simultaneously....but since it’s just XML a sense of ordering isn’t nicely preserved.

@kirma in the US there is also a policy research firm called Mathematica. I’ve met people who didn’t realize it was separate from WRI (and who were confused as to why it wasn’t).

JimB

@halirutan The PacletInstall worked flawlessly on MMA 11.3 Windows 10.

@JimB Thanks for testing!

kirma

3:58 PM

@b3m2a1 I'd say trademark confusion (after initial puzzlement) between Mathematica Policy Research and WRI is not too realistic, but with a company creating mathemathics education software/services... well, there's a real chance.

4:19 PM

@kirma valid point

@Szabolcs chat.stackexchange.com/rooms/79041/paclet-servers

Kuba

@b3m2a1 If I recall you were playing with this: mathematica.stackexchange.com/q/175543/5478

1 hour later…

bobthechemist

5:28 PM

Any stats folks around who can help clarify something? I'm trying to calculate [Cronbach's alpha](https://en.wikipedia.org/wiki/Cronbach%27s_alpha) and there's a statement "[c-bar is] the average of all covariances between the components across the current sample of persons". I don't understand what covariances I'm supposed to calculate and wonder if this would be correct: `Mean@Flatten@
Covariance[N@ExampleData[{"Statistics", "EmployeeAttitude"}]]`

5:51 PM

@bobthechemist I think you want @JimB

JimB

6:48 PM

@bobthechemist You'll need to take out the diagonal of the covariance matrix. So maybe something like the following:

cov = Covariance[N@ExampleData[{"Statistics", "EmployeeAttitude"}]];
n = Dimensions[cov][[1]];
cbar = Total[Flatten[cov - IdentityMatrix[n] Diagonal[cov]]]/(n (n - 1))

@bobthechemist I also note that the Wiki page does not even discuss a measure of precision for alpha which makes me suspect that most folks don't bother with a measure of precision or account for how the observations were obtained.

bobthechemist

@JimB Thanks. I will keep digging about precision. At the moment, I'm mostly interested in answering the question "do student teaching evaluations agree on instructor performance?" It is for my own personal benefit and not something that is to be published, so I don't know how much rigor is warranted.

The wikipedia page provides the two different definitions. It's not clear to me whether or not these definitions should provide the same result or that I would have to indicate which method I'm using (again, assuming that I were using this analysis in research).

1 hour later…

JimB

8:06 PM

@bobthechemist The definitions are really the same. The first one can be written as FullSimplify[(K/(K - 1)) (1 - Sum[v[i]^2, {i, K}]/(Sum[v[i]^2, {i, K}] +
2 Sum[cov[i, j], {i, 2, K}, {j, 1, i - 1}])) /. {Sum[v[i]^2, {i, K}] -> K vbar,
2 Sum[cov[i, j], {i, 2, K}, {j, 1, i - 1}] -> K (K - 1) cbar}]

which simplifies to the second one:

K cbar/(vbar + (K - 1) cbar)

And while the article skips over estimating precision (standard errors, confidence intervals, etc.) one or two of the referenced articles talks about estimating precision.