Proposal discussion for Cantor: Python 3 as the only backend officially supported

[Alexander Semke]

Hi Filipe,

>> Which backends are causing the most problems at the moment? Is it Sage?
>> Maxima?
>>
> R was broken for several years, but there is a fix now (not released
> yet). Sage breaks for each new release. Julia had some problems with
> previous versions. Lua has bugs after a new release.
>
> Scilab was unstable for some previous releases but now it is working.
> Octave and KAlgebra/Analitza are working fine, but the first one
> sometimes had small bugs to be fixed. Python 2 and 3 are ok.
>
> I don't know about the status of Maxima and Qalculate.
>
> Cheers;
>
Having a frontend to different systems is one of the major arguments in 
favor of Cantor. By giving this up you'll heavily reduce the importance 
of Cantor.

There are some bugs now where some combinations of different versions of 
Cantor and Maxima, Octave, etc. don't work. But it's not like everything 
is broken and always.
I just tested on openSUSE 42.2 with Maxima (two different versions), 
Octave, julia and python2/3 - everything works as expected. There are of 
course problems later with each
of these backends and there are already couple of bug reports for these 
issues. But at least the "connection" to the backends is working and I 
think this is what partially causes
troubles in your mind now. But instead of giving up this code 
completely, simply keep it as it is - it's not perfect, but it's usable.

You can put your priorities on python now and also on the worksheet 
itself in order to improve the quality in general in this area - a lot 
of stuff can be implemented in Cantor completely
independent of the actual backend being used. All systems will benefit 
from a better worksheet in Cantor and from the tooling around it anyway. 
Better quality in Cantor's frontend
will attract maybe new developers who can take care of the backends they 
are familiar with.


Regards,
Alexander