[ase-users] [EXTERNAL] Pure python ASE viewer

[Noam Bernstein]

> On Aug 15, 2019, at 10:59 AM, Michael Walter <Michael.Walter at fmf.uni-freiburg.de> wrote
> 
> As I inderstand you are working on an alternative implementation for the gui. Would it be possible to reuse some of the work done for the present gui (which I regard as quite powerful) or even set up on this?

More or less, at least in the context of static configurations (or lists, i.e. trajectories), not as a front-end to arbitrary ASE-using codes (like kmos, which I’ve seen use the GUI).   

My guess (not having looked at the source) is that it’d be quite hard to add vtk to the GUI, because vtk has its own event loop, interactors (mouse rotation, object picking), etc, but I haven’t really looked into that at all carefully.  

> 
> In my experience alternative implementations nearly always miss some of the features.

That’s certainly true, but on the other hand, it probably has some features that the existing GUI does not.

If I’m right about the fundamental difficulty of merging the ase.gui with vtk, the remaining alternative is to re-implement ase.gui in vtk, i.e. do what I’ve done, but use an interface (API and/or UI) closer to the existing GUI.  I’m torn about whether I’m interested in doing this myself, because I have a long history of using tools with similar interfaces to the one I’ve been developing, and I’m not sure I want to change.  I realize I’m swimming against the tide (or just showing my age?), but I don’t really like menu-driven things for visualization, especially when they don’t (or don’t document) how to achieve the same effects by scripting, config files, etc.

But I do take your point that there’s substantial overlap with the existing GUI (at least from what I can tell from the docs, since my pythons, 3.6 from CentOS SCL and 3.6 from macports, don’t seem to have Tkinter).  I’ll think about whether there’s anything I can about that, either in the direction of using vtk within the existing ase.gui, or redoing the interface of my vtk-based tool to make it more like ase.gui (perhaps ultimately replacing ase.gui in the future).  

								Noam
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.fysik.dtu.dk/pipermail/ase-users/attachments/20190815/8512bcee/attachment-0001.html>