xAct communication and working examples.

[Cyril Pitrou]

Dear xAct users, and dear JMM,

First apologies for not having taken time to check if xPand is OK with the new xAct distribution. I will do it soon.

I write because I noticed that for all beginners or at least late users, in xAct, it is always the same feeling.

They are expected to learn by reading carefully all the documentation, exactly like one is expected to learn maths, starting from the basic axioms, through set theory group theory field theory etc...
However learning a programming language for people expert otherwise in physics and computing science has more to do with learning a foreign language in the country. Infusion. Modifying examples, and only after some times reading the documentation about a specific point.

So I think what is missing is actually much simpler than what has already been written in the detailed documentation. What we need is just a page full of working examples. I think all experts in xAct who are animating it are somehow ashamed of providing super simple examples as they would think it is super trivial. But for new users this is exactly what is need to bridge the gap between knowing just mathematica, and solving complex problems in xAct.

This is all the more true for xCoBa, where working  examples in free access are missing. I think just a list of 20 to 50 working examples would be enough to convince new users. I am sure many are abandoning even though xAct is quite simple after all. I deeply think this is what is now missing that the system is more or less complete. I am bad in argumenting in emails, sorry. I do not try to argument too much because I have the feeling that everybody would agree on that.

So far on xact.es website I can only see the working examples of Teake Nutma, whose link is broken, and the file by Jolyon Bloomfield.

So the questions I am sking are :

1) How do we do to have such list of working examples ? I know Leo Stein has a few ones on one of his pages. I think since these examples are meant to be very basic, they will not evolve and one could put them on xact.es pages in the documentation page (addding a section 'basic working examples'). If you have better, but simple  ideas, we can debate. I know nothing about the web. Or one would want to have these examples be improved by knowledgeable people (not me), but then one would need some co-working solution.

I think as a starter we should just prepare some examples, and we will think later in how to improve them with the best programming practices.

I think a list of links to the notebooks with one line of description for each notebook would be enough. As long as this in the page 'documentation' it would be higly visible.

2) What list of working examples (if you agree) ? I guess if we all come up with what is in our mind with 3 to 5 examples, that would make a huge list of simple but interesting problems. I remember there was a KerrNewman Example a long time ago on the xact.es site. It was extreely useful and simple at the same time.

 

For instance, from my usage of xAct I can see as obvious working examples that I could provide in just a few hours of work. If you agree to that project I can prepare them. (I guess other people who use xAct 100 times more than me could come with some simple and basic examples very easily. I think we tend to be ashamed of simple notebooks and they are not worthy, when in fact they are for beginners)

1) Variation of action wrt to metric BUT ALSO with respect to fields, so as to get field equations. Everybody needs that. If people had clearly seen that xAct can do that, the server would be victim of overload.

2) xCoBa in the TensorValues framework for Friedmann-Lemaitre spacetimes (playing with various types of coordinates, for instance).

2bis) The same thing for static and axisymmetric spacetimes. With the proof that the only such spaces are Schwarschild. That would even be nice to give some examples of usual change of coordinates (I am not an expert but I know there are a few standard coordinates systems).

2ter) The same thing for Bianchi I spaces (This I know how to dit it).

3) Maybe some working examples for the 1+3 splitting in GR. (This would require some more work but I did some things about that). This is a xTensor example.

4) Some basic example for Clifford Algebra. (OK for this I meed Guillaume Faye). So that people from particle physics can compute |M|^2 and cross sections for nasty theories. I did it for weak interactions (in the Fermi theory) in  a recent publication, and xAct was so powerful it was amazing. We can only fear that once we have a good working example, 10^3 particle physics researchers will use it for their computations. I would need to work a bit on that example since I do not master what is needed really there but I think this is a very important example for the communication about xAct.

I am willing to start and I will send soon the examples to show you what I have in mind. I know you will say it is poorly written but I have no reputation to defend here.

This is just to convince you it is easy and should be rapid for each of us and should help a lot xAct.

Bon dimanche,

Cyril Pitrou

[Leo Stein]

Dear Cyril et al.,

I strongly agree, more example notebooks should make xAct easier to learn! That was the hope behind the public repository you mentioned. But I guess we did a bad job of advertising it, because people don't seem to know of its existence. And nobody else has contributed example notebooks.

For anybody who doesn't already know: there is a repository of example notebooks here: https://github.com/xAct-contrib/examples

If you have an example notebook that you want to contribute, you can email it to me (or, if you're familiar with git and the pull request model, you can send a PR).

More comments inline below...

On Sat, Mar 24, 2018 at 10:28 AM, Cyril Pitrou <cyril....@gmail.com> wrote:

Dear xAct users, and dear JMM,

First apologies for not having taken time to check if xPand is OK with the new xAct distribution. I will do it soon.

I write because I noticed that for all beginners or at least late users, in xAct, it is always the same feeling.

They are expected to learn by reading carefully all the documentation, exactly like one is expected to learn maths, starting from the basic axioms, through set theory group theory field theory etc...
However learning a programming language for people expert otherwise in physics and computing science has more to do with learning a foreign language in the country. Infusion. Modifying examples, and only after some times reading the documentation about a specific point.

So I think what is missing is actually much simpler than what has already been written in the detailed documentation. What we need is just a page full of working examples. I think all experts in xAct who are animating it are somehow ashamed of providing super simple examples as they would think it is super trivial. But for new users this is exactly what is need to bridge the gap between knowing just mathematica, and solving complex problems in xAct.

This is all the more true for xCoBa, where working  examples in free access are missing. I think just a list of 20 to 50 working examples would be enough to convince new users. I am sure many are abandoning even though xAct is quite simple after all. I deeply think this is what is now missing that the system is more or less complete. I am bad in argumenting in emails, sorry. I do not try to argument too much because I have the feeling that everybody would agree on that.

So far on xact.es website I can only see the working examples of Teake Nutma, whose link is broken, and the file by Jolyon Bloomfield.

So the questions I am sking are :

1) How do we do to have such list of working examples ? I know Leo Stein has a few ones on one of his pages. I think since these examples are meant to be very basic, they will not evolve and one could put them on xact.es pages in the documentation page (addding a section 'basic working examples'). If you have better, but simple  ideas, we can debate. I know nothing about the web. Or one would want to have these examples be improved by knowledgeable people (not me), but then one would need some co-working solution.

I think as a starter we should just prepare some examples, and we will think later in how to improve them with the best programming practices.

I think a list of links to the notebooks with one line of description for each notebook would be enough. As long as this in the page 'documentation' it would be higly visible.

2) What list of working examples (if you agree) ? I guess if we all come up with what is in our mind with 3 to 5 examples, that would make a huge list of simple but interesting problems. I remember there was a KerrNewman Example a long time ago on the xact.es site. It was extreely useful and simple at the same time.

 

For instance, from my usage of xAct I can see as obvious working examples that I could provide in just a few hours of work. If you agree to that project I can prepare them. (I guess other people who use xAct 100 times more than me could come with some simple and basic examples very easily. I think we tend to be ashamed of simple notebooks and they are not worthy, when in fact they are for beginners)

1) Variation of action wrt to metric BUT ALSO with respect to fields, so as to get field equations. Everybody needs that. If people had clearly seen that xAct can do that, the server would be victim of overload.

There is a simple example in the file Lagrangian-variation-xPert-VarD.nb in the repo I linked above. But of course more examples would be better!

 

2) xCoBa in the TensorValues framework for Friedmann-Lemaitre spacetimes (playing with various types of coordinates, for instance).

2bis) The same thing for static and axisymmetric spacetimes. With the proof that the only such spaces are Schwarschild. That would even be nice to give some examples of usual change of coordinates (I am not an expert but I know there are a few standard coordinates systems).

2ter) The same thing for Bianchi I spaces (This I know how to dit it).

In general, it may be nice to have a library of exact solutions in common coordinates/bases coded up. That is a lot of work, but we could start with the common ones—for example Kerr-Newman is already done. Barry Wardell started a project along these lines (here: https://github.com/barrywardell/Metrics) but it looks like it's been dormant for 6 years.

A very ambitious goal would be to include every exact solution from some standard reference works, like Stephani et al., or Griffiths and Podolsky.

 

3) Maybe some working examples for the 1+3 splitting in GR. (This would require some more work but I did some things about that). This is a xTensor example.

I have an example notebook for this that I need to clean up and upload to my repo. I will try to do it in the next few days (but I will also be traveling to a workshop...).

 

4) Some basic example for Clifford Algebra. (OK for this I meed Guillaume Faye). So that people from particle physics can compute |M|^2 and cross sections for nasty theories. I did it for weak interactions (in the Fermi theory) in  a recent publication, and xAct was so powerful it was amazing. We can only fear that once we have a good working example, 10^3 particle physics researchers will use it for their computations. I would need to work a bit on that example since I do not master what is needed really there but I think this is a very important example for the communication about xAct.

I am willing to start and I will send soon the examples to show you what I have in mind. I know you will say it is poorly written but I have no reputation to defend here.

This is just to convince you it is easy and should be rapid for each of us and should help a lot xAct.

This is a great attitude :) Sometimes worse is better, especially when we get the ball rolling to test some possibilities. We can restructure later once we learn from our mistakes.

Thanks!

Leo

 

Bon dimanche,

Cyril Pitrou

--
You received this message because you are subscribed to the Google Groups "xAct Tensor Computer Algebra" group.
To unsubscribe from this group and stop receiving emails from it, send an email to xact+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Thomas Bäckdahl]

Hi!

I also agree that more examples should be available, so I tidied up an old example proving vanishing of the Lovelock tensor using spinor methods and irreducible decomposition. It is now available on  https://github.com/xAct-contrib/examples

Regards
Thomas

To unsubscribe from this group and stop receiving emails from it, send an email to xact+uns...@googlegroups.com.