SageManifolds 1.0 in Sage 7.5 and development workflow

101 views
Skip to first unread message

Eric Gourgoulhon

unread,
Jan 15, 2017, 9:51:21 AM1/15/17
to sage-devel
Dear Sage devs,

The just released Sage 7.5 contains the integrality of SageManifolds code,
the last SageManifolds tickets having been merged in 7.5.beta4
(cf. https://trac.sagemath.org/ticket/18528 )
Consequently, there is no need for a separate installation of
SageManifolds (*), which is released in version 1.0 as part of Sage 7.5:
http://sagemanifolds.obspm.fr/
thanks to these developers and reviewers:
http://sagemanifolds.obspm.fr/authors.html

The full integration in Sage will of course facilitate the use of
SageManifolds, on a personal computer or in the SageMathCloud, or via the Sage
Debian Live key. But this will also considerably ease the development of
manifold code, especially by newcomers. For them, I've prepared the page
http://sagemanifolds.obspm.fr/contrib.html
It presents some basic workflow, from the installation of the latest Sage
develop version to submission of code via a Trac ticket, with links to some
sections of Sage's Installation Guide and Sage's Developer Guide. I would
appreciate any feedback on this page from Sage developers, in order to
correct/improve it.
NB: of course, duplication is the root of all evil and the above page is not
intended to be a substitute for Sage's Installation and Developer guides, but
rather a kind of synthetic summary pointing to these guides.

Best regards,

    Eric.

(*) Accordingly, a change of terminology is performed from "SageManifolds
package" (SageManifolds was actually a Sage spkg until version 0.5) to
"SageManifolds project". See also the "Direct integration into Sage" section
of the development models discussed at
https://wiki.sagemath.org/CodeSharingWorkflow


William Stein

unread,
Jan 15, 2017, 1:59:26 PM1/15/17
to sage-devel
On Sun, Jan 15, 2017 at 9:51 AM, Eric Gourgoulhon
<egourg...@gmail.com> wrote:
> Dear Sage devs,
>
> The just released Sage 7.5 contains the integrality of SageManifolds code,
> the last SageManifolds tickets having been merged in 7.5.beta4
> (cf. https://trac.sagemath.org/ticket/18528 )
> Consequently, there is no need for a separate installation of
> SageManifolds (*), which is released in version 1.0 as part of Sage 7.5:
> http://sagemanifolds.obspm.fr/
> thanks to these developers and reviewers:
> http://sagemanifolds.obspm.fr/authors.html
>
> The full integration in Sage will of course facilitate the use of
> SageManifolds, on a personal computer or in the SageMathCloud, or via the
> Sage
> Debian Live key. But this will also considerably ease the development of
> manifold code, especially by newcomers. For them, I've prepared the page
> http://sagemanifolds.obspm.fr/contrib.html

404 Not Found

> It presents some basic workflow, from the installation of the latest Sage
> develop version to submission of code via a Trac ticket, with links to some
> sections of Sage's Installation Guide and Sage's Developer Guide. I would
> appreciate any feedback on this page from Sage developers, in order to
> correct/improve it.
> NB: of course, duplication is the root of all evil and the above page is not
> intended to be a substitute for Sage's Installation and Developer guides,
> but
> rather a kind of synthetic summary pointing to these guides.
>
> Best regards,
>
> Eric.
>
> (*) Accordingly, a change of terminology is performed from "SageManifolds
> package" (SageManifolds was actually a Sage spkg until version 0.5) to
> "SageManifolds project". See also the "Direct integration into Sage" section
> of the development models discussed at
> https://wiki.sagemath.org/CodeSharingWorkflow
>
>
> --
> You received this message because you are subscribed to the Google Groups
> "sage-devel" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to sage-devel+...@googlegroups.com.
> To post to this group, send email to sage-...@googlegroups.com.
> Visit this group at https://groups.google.com/group/sage-devel.
> For more options, visit https://groups.google.com/d/optout.



--
William (http://wstein.org)

Eric Gourgoulhon

unread,
Jan 15, 2017, 2:53:08 PM1/15/17
to sage-devel
Le dimanche 15 janvier 2017 19:59:26 UTC+1, William a écrit :
 
404 Not Found


Indeed there is a problem on the server. I am contacting the support team...

Kwankyu Lee

unread,
Jan 15, 2017, 3:24:37 PM1/15/17
to sage-devel
You managed to develop SageManifolds separately from Sage, and then to integrate the code fully with Sage. I am curious how this is done. It looks daunting to me to chop out big code to small patches for gradual merge, without breaking Sage. 

Current Sage development workflow forces an author of a project developed externally to work in this way. Was your experience a pleasant one to recommend this way or do you think we need a better and alternative way for future authors?

Would you share your experience, in some place like https://wiki.sagemath.org/CodeSharingWorkflow ? 

Eric Gourgoulhon

unread,
Jan 15, 2017, 4:07:30 PM1/15/17
to sage-devel
The support team has just fixed the issue; the server is up again:
http://sagemanifolds.obspm.fr/contrib.html

Eric.

Eric Gourgoulhon

unread,
Jan 15, 2017, 4:49:38 PM1/15/17
to sage-devel
Le dimanche 15 janvier 2017 21:24:37 UTC+1, Kwankyu Lee a écrit :
You managed to develop SageManifolds separately from Sage, and then to integrate the code fully with Sage. I am curious how this is done. It looks daunting to me to chop out big code to small patches for gradual merge, without breaking Sage. 

Current Sage development workflow forces an author of a project developed externally to work in this way. Was your experience a pleasant one to recommend this way or do you think we need a better and alternative way for future authors?

Indeed it took time to split the original SageManifolds package into many tickets, but it was also the occasion to think again about the whole code structure. Above all, the interaction with the reviewers and the whole Sage community (through questions I posted on this list) was very fruitful and instructive to me. One of the reviewers, Travis Scrimshaw helped a lot, improving the quality of the code and its documentation. He also modified some parts quite substantially, introducing new constructions. So yes it was a pleasant experience and I do not regret the choice of such a development workflow, despite one of its main drawback: until all the tickets were merged, its was difficult for people to add new features. But now, I expect that it will be much easier, hence the proposed contribute page.

Would you share your experience, in some place like https://wiki.sagemath.org/CodeSharingWorkflow ? 

Yes, I may add something there.

Best wishes,

Eric.

William Stein

unread,
Jan 15, 2017, 5:10:33 PM1/15/17
to sage-...@googlegroups.com
Kind of like publishing and important and foundational research paper...?
--


You received this message because you are subscribed to the Google Groups "sage-devel" group.


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


To post to this group, send email to sage-...@googlegroups.com.


Visit this group at https://groups.google.com/group/sage-devel.


For more options, visit https://groups.google.com/d/optout.


--
Sent from my massive iPhone 6 plus.

Eric Gourgoulhon

unread,
Jan 15, 2017, 6:27:26 PM1/15/17
to sage-devel
Le dimanche 15 janvier 2017 23:10:33 UTC+1, William a écrit :
Kind of like publishing and important and foundational research paper...?
 
More kind of like writing a book ;-)

Paul Masson

unread,
Jan 15, 2017, 7:33:17 PM1/15/17
to sage-devel
Eric, it seems to me that this is now an area in which SageMath is ahead of Mathematica, and you are to be congratulated for that. Given that Wolfram is a physicist, I've never understood why there is no native support in Mathematica for general relativity calculations.

Have you considered adding a page to the SageManifolds site comparing features available now in SageMath to what's available in the all the M's? Might be a good form of promotion and attract a physicist or two to the project.

Eric Gourgoulhon

unread,
Jan 16, 2017, 4:33:29 AM1/16/17
to sage-devel
Hi Paul,

Le lundi 16 janvier 2017 01:33:17 UTC+1, Paul Masson a écrit :
Eric, it seems to me that this is now an area in which SageMath is ahead of Mathematica,

For me, being open-source and based on Python, Sage was always ahead of Mathematica :-)
More seriously, this may be true for *concrete* tensor calculus (by *concrete*, it is meant that you have to specify the manifold dimension and some atlas, and the computations are, in fine, performed on the tensor components), but regarding *abstract* tensor calculus (i.e. the manifold dimension is some symbolic variable n, no atlas specified), there is no functionality in Sage yet, while there is the very powerful xAct suite for Mathematica:
http://www.xact.es/
But this might be the topic for another extension of Sage.

and you are to be congratulated for that.

Thanks, but this is really a team work:
http://sagemanifolds.obspm.fr/authors.html

Given that Wolfram is a physicist, I've never understood why there is no native support in Mathematica for general relativity calculations.

Have you considered adding a page to the SageManifolds site comparing features available now in SageMath to what's available in the all the M's?

If time permits, yes.
 
Might be a good form of promotion and attract a physicist or two to the project.

Now that version 1.0 is released and integrated in Sage, I'll make more promotion in the general relativity community.

Best regards,

Eric.

PS: and thank you very much for the threejs viewer, very useful for manifold visualizations:
http://nbviewer.jupyter.org/github/sagemanifolds/SageManifolds/blob/master/Worksheets/v1.0/SM_sphere_S2.ipynb

Reply all
Reply to author
Forward
0 new messages