C. E.

general recent conversations

Wolfram Mathematica

Welcome! This is the main Mathematica chat room for mathematic...

C. E.

Feb 18 09:00

Here are the restrictions: stackoverflow.com/help/deleted-questions Assuming that we have the same restrictions as Stackoverflow. With these restrictions, it shouldn't happen too often.

C. E.

Feb 15 20:15

@CraigCarter I would try to leave a comment on one of his answers. That way he'll be alerted to the comment.

C. E.

Feb 13 15:40

@user21 @MichaelE2 This is a fair point. I don't think it's too much to ask for a profile picture that doesn't draw scrutiny in this way. I will change it for him and ask him to pick a new.

C. E.

Feb 13 08:38

@xzczd @bmf @JimB @user21 It turns out that the profile picture is the number 11 ball from billiard. He wanted to keep it and I told him that I don't oppose it as long as it doesn't become a regular thing that it gets reactions.

C. E.

Feb 12 12:30

I made a comment on one of his questions now, hopefully he will see it.

C. E.

Feb 12 08:55

@xzczd @bmf @JimB Generally, it's in breach with the code of conduct if it causes harm, even if the harm is inadvertent. Even if 11 is a reference to the sacred number in Hinduism and signifies spiritual awakening, since the user is from India, it could still be a breach of the code of conduct since it means something else to other users. But this is kind of an edge case and I'm not sure what to do. I'm leaning towards removing it since of all the avatars on the site, this stands out.

C. E.

Jan 24 08:52

The post mentions that Dataset is more general than Tabular in that it supports hierachical data. However, the most efficient ways of working with tabular data in Python such as pandas and polars support nested data. The file format Parquet itself also supports nested data. So the reason for tabular not supporting nested data should not be efficiency, I don't think.

C. E.

Jan 24 08:41

Yes, it will be very useful for AI. It implements what already exists in other languages and needs to exist in Wolfram Language as well. There are very optimized ways of reading tabular file formats such as Parquet, and representing the data efficiently in memory. I thought that Dataset would be the way to represent tabular data in Wolfram Language, but they probably missed out on the optimizations and had to start on something new to include them.

C. E.

Dec 13, 2024 15:54

@xzczd No, unfortunately not but it looks nice in demos. I use Github Copilot and ChatGPT for coding on a daily basis so I have no doubts that it can work well.

C. E.

Dec 13, 2024 12:35

@xzczd No, but I watched the announcement video by S.W. on Youtube where he demoed it, and the Computational X video on Youtube where he uses it a bit more of different problems and I think it looks nice.

C. E.

Oct 23, 2024 20:40

Thank you to everyone who nominated themselves to make the election possible and also thank you to everyone who voted!

C. E.

Oct 23, 2024 20:39

Thank you! Finally I've gotten the role that I've wanted for many years. Now let the work begin.

C. E.

Aug 14, 2024 10:36

C. E.

Jul 9, 2024 10:32

@ТymaGaidash ctrl+z would probably have worked immediately after.

C. E.

Jul 9, 2024 09:47

You can read the comments to this meta thread for some clues: https://politics.meta.stackexchange.com/q/6946/17614

He seems to be a bit of a character. There is a link to another meta thread where he himself wants to ban the term "historical Palestine" which he considers anti-semitic.

C. E.

Jun 30, 2024 20:25

So if I understand this correctly: PacletInstall (on Windows) installs paclet files into <user>\AppData\Roaming\Mathematica\Paclets\Repository, but this path is not in $Path, so once the paclet is installed it cannot be imported using Needs, without figuring this out and manually adding it to $Path. This isn't very user-friendly.

C. E.

Jun 10, 2024 18:26

@kirma Nice

C. E.

Jun 10, 2024 14:52

Considering the ease with which ChatGPT translates the code into Wolfram Language, I wonder if there won't be a day when we will all exchange ideas in pseudocode.

C. E.

Jun 10, 2024 08:23

@bmf My understanding is that we do not. On the other hand, it looks like he's genuinely tries to be helpful and is not posting it e.g. to get reputation points without having to actually write the code.

C. E.

Mar 21, 2024 11:04

@kirma Yes, and of course, in the browser war between Google, Apple and Microsoft, there is hardly any amount of hassle that they will spare, so performing as their browsers will always be a very difficult benchmark to beat.

C. E.

Mar 21, 2024 08:01

@Nasser When you view it locally, the pages are Mathematica notebooks, while in the browser, they are HTML pages. I'm not surprised that it makes a difference.

C. E.

Jul 26, 2023 07:58

m_goldberg has been on my mind recently. He is over 80 years old and stopped using the site suddenly two years ago.

C. E.

Jun 29, 2023 07:25

It's nice that they added the AR feature, but it seems that this time they went directly and only to high-level functionality. For example, with the machine learning stuff, they have some high-level functions but they also have lower-level functions that allow you to implement things yourself. In the case of AR, it would be nice if they could have a camera model, pose estimation algorithms and so on, as needed for 3D computer vision and AR.

C. E.

Jun 19, 2023 07:19

He also talks about it in other places, such as the book that he published and this blog post: writings.stephenwolfram.com/2023/03/…

C. E.

Jun 19, 2023 07:13

Actually, the [most recent blog post](https://stackoverflow.blog/2023/06/14/hype-or-not-developers-have-something-to-say-about-ai/?cb=1&_ga=2.158008698.1307856021.1687158257-1441724409.1662123942) on the Stack Exchange blog mentions that the most highly voted question Stack Exchange AI related to ChatGPT is [Why is ChatGPT bad at math?](https://ai.stackexchange.com/questions/38220/why-is-chatgpt-bad-at-math).

ChatGPT is not that smart, is the simple answer. Stephen Wolfram has an interesting take on this in the Lex Friedman podcast mentioned by @kirkus. He says that what ChatGPT (and other…

C. E.

Jun 8, 2023 23:18

@Nasser You can see that here. We have about 5-10 thousand page views per day.

C. E.

Jun 6, 2023 20:54

I see. Good observation.

C. E.

Jun 6, 2023 20:13

It woud be interesting to see S.W.'s response about Binomial[-1, -1]. Did WR pick the wrong generalization for negative numbers or was Knuth simply a bit hyperoblic when he said that anyone who has worked on this in the last 50 years knows that Binomial[-1, -1] should be zero and not one like in Mathematica.

C. E.

Apr 7, 2023 17:54

I'm saying this in jest, of course, but that would be TreeForm, no? On a more serious note, what are you looking for? Because TreeForm shows what will be evaluated and in what order.

C. E.

May 20, 2022 00:43

@b3m2a1 I feel the same. I have at least two packages that I think would be a good fit for the repository but I would need to write documentation and so on…

C. E.

May 19, 2022 16:36

We have talked about a paclet repository for how long? And now it is here. I can’t help but think of when we finally got the undo function.

C. E.

May 2, 2022 19:51

@b3m2a1 Sorry for not answering. Really nice. Still haven’t tried it out but I’m looking forward to.

C. E.

Apr 22, 2022 16:12

@dtn You can learn a lot by simply perusing the documentation of functions. It is pretty great at showing not just what the syntax is but also what some applications are. And there is some tutorial style content as well, such as this: reference.wolfram.com/language/tutorial/…

C. E.

Apr 18, 2022 21:10

@b3m2a1 ok

C. E.

Apr 14, 2022 12:58

@b3m2a1 thank you

C. E.

Apr 13, 2022 21:31

@b3m2a1 this looks really good, I bet a lot of people would probably be interested in being able to quickly put together forms like that. Do you have any plans to make this publicly available?

C. E.

Apr 8, 2022 17:48

@b3m2a1 It looks really nice!

C. E.

Apr 5, 2022 22:01

@b3m2a1 ok, sounds interesting. I should take a closer look at it.

C. E.

Apr 5, 2022 21:42

Are you able to for example print a slider and have the slider change some variable in the Jupyter notebook with this approach? And if so, is it possible to update other output cells that also depend on this variable? One problem I have with the built-in interact is that when I move the slider, I want to update not just things in that output cell...

C. E.

Apr 5, 2022 21:36

But based on what you say it sounds like there is a fundamental limitation to how good it can get anyway without great amounts of effort.

2024 Mathematica Stack Exchange Moder…

Open discussion for mathematica.stackexchange.com/election/4

C. E.

Oct 23, 2024 20:47

Thank you!

C. E.

Oct 23, 2024 20:36

@rhermans Thank you very much!

C. E.

Oct 23, 2024 20:29

The result seems to only be viewable using the OpenSTV tool for now. Thank you again to everyone who nominated themselves and made the election possible!

C. E.

Oct 23, 2024 20:25

C. E.

Oct 10, 2024 07:42

@Rabbit I have now made a comment on one of Domen's answers to encourage him to participate in the election.

C. E.

Oct 9, 2024 16:21

@681234 No, please read the previous answer by Sasha. If after the one week extension there are still fewer than 3 candidates then the election will fail.

C. E.

Oct 7, 2024 14:46

I think it means that if we don't get two candidates then the election may be canceled, but if we get two or more candidates then it will proceed.

C. E.

Oct 7, 2024 14:45

@xzczd Yes, ready to serve! Regarding your question, we at least have this description from the nomination page to on: "After 7 days, the top 30 nominees, ordered by reputation, advance to the primary phase. However, if there are 10 or fewer candidates, we skip directly to the election phase. If there are not enough eligible nominations at the end of the nomination period, it will be extended by one week. If there are not enough nominations after an extension the election will be canceled."

C. E.

Oct 2, 2024 12:06

@681234 You're right, I missed that.

C. E.

Oct 2, 2024 07:44

Is the election for just one moderator or for two? (The community recently lost two moderators.)