Proposals for a xAct virtual event?

[Jose]

Hi all,

This thread has number 1000 in the xAct forum. On behalf of all xAct developers, I want to use the opportunity to thank you all again for your interest and support of xAct.

In another thread, Soham just suggested that we have a hackathon on xAct. It can be certainly be useful to gather and talk about xAct, but I don't know what would be the best or more useful format of such a gathering. What are your preferences?

1. An introduction like the ICERM 2020 course? https://www.youtube.com/watch?v=MQqbyR8hCrA . Different developers should present their own packages.

2. A general discussion on xAct: current problems (bugs or design issues), proposals for extensions, development methods and tools, etc

3. General Q/A about xAct.

4. Should we organize the discussion by individual packages instead of having general xAct meetings?

What else would you suggest?

I think this type of meeting would be more successful if we plan its structure in advance.

Cheers, and many thanks,

Jose.

[Subhodeep Sarkar]

Hello,

This sounds like a great idea. Along with a thorough introduction to xACT like the one in ICERM, I think new users like me will benefit a lot more from hands-on sessions showing some specialised applications of xAct. They can be based on some of the publicly available examples.

I must mention that recently, in the Einstein Toolkit workshop, Prof Witek gave a very through introduction to performing the 3+1 decomposition of the Einstein equations using xAct and it was extremely helpful. So more tutorials like that will be greatly appreciated.

Thanks and regards,

Subhodeep

--
Subhodeep Sarkar 
Research Scholar (RSS2019005) 
Department of Applied Sciences (Physics) 
Indian Institute of Information Technology, Allahabad
_________________
"Perfection of character is this: to live each day as if it were your last, without frenzy, without apathy, without pretence." 
~Marcus Aurelius, Meditations (7:69)

   

--
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+uns...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/xact/91fb398f-ac13-4920-a0e3-9943df29afcen%40googlegroups.com.

[Leo Stein]

Dear Jose et al.,

I think this could be a successful idea, assuming that the schedules of a good number of people line up in the right way to make it happen. 

On organization:

I have heard about 'hackathons' employing the following mechanism to organize the work:

- People who have ideas for projects should pitch their project(s) to everybody else, in a presentation or document, before the work begins. This could be e.g. a prototype or sketch of the functionality that somebody wants to implement.

- After all these presentations, people can decide which project they want to contribute to, and each group can discuss the organization for their group, and the programming tasks they need to accomplish.

- If it seems like there are going to be a very large number of proposals, then some set of people probably need to be an organizing committee, to pre-screen the proposals and down-select for the most mature ones.

- I think this hackathon idea would be easiest using github (which is currently used for many xAct extensions), or somebody may want to host their own version control server using gitlab or other software. In general, I think having an official public git repo for xAct/xTensor could be quite beneficial -- github already has a mechanism for issue tracking and bug reporting, creating project boards (kanban style development), etc. I have seen Mma packages hosted on github with a mechanism to check whether or not there are updates from with Mma, by querying github (see e.g. https://github.com/bshoshany/OGRe ).

On projects:

If I was to pitch a project, it would probably be to make an interface for users to create a Poisson algebra and do calculations with Poisson brackets. You can see an example with a nontrivial Poisson algebra in the ancillary notebook from this paper: https://arxiv.org/abs/2012.06586 . Although in that example the Poisson algebra is on a phase space for finite number of degrees of freedom, I imagine that some folks might want to be able to do field space Poisson bracket calculations, like for the Hamiltonian formalism of GR or other field theories. Being able to automate constraint algebra calculations seems like a nice goal.

I could alternatively propose a project to xTerior for creating surfaces which can be paired with forms to represent integration; so that the variational derivative can actually act on an integral, and keep boundary terms. Doing this for arbitrary tensors instead of forms would be harder, so we should start with forms.

Best

Leo

--

[Soham Bhattacharyya]

Dear Leo, Jose, et al.,

I think Leo explained the process best. In terms of how hackathons go, say one person in a group of people has a particular feature they want to get implemented into the system. They have done the 'feature' or the calculation by hand so far and it seems repetitive to them when it comes to utilizing one calculation in calculations that come afterward. Following Leo's example say that I want to have an algorithm that helps me get variational derivatives on non-scalar objects (which I noticed is present as a footnote in the xTensor.nb notebook). If one has done such 'atomic' calculations by hand, they have a rough idea of the algorithm they follow in their head while doing it on pen and paper. So that person can suggest the feature and a rough algorithm (which others might agree to as well) to the entire group. I think there might be a lot of such expected features in different sub-fields of field theory. The moderator can make a list of all the expected features and the corresponding algorithm and then break up into groups focussing on a specific problem.

This particular format makes the audience play a part in the development process compared to a lecture series where the audience is just a listener.

I hope this made sense.

With regards,

Soham