Re: classdef development

[Richard Crozier]

On 17/12/2012 13:07, Danilo Zanatta wrote:
> On Wed, Dec 12, 2012 at 11:39 AM, Серёжа Плотников
> <address@hidden <mailto:address@hidden>> wrote:
>
>     Dear maintainers,
>
>     Could you please say me how far you are from having usable (or at
>     least testable) matlab-like class system (classdef)? Is it possible
>     to be involved into testing process?
>
>     The thing is that I'm about to rewrite some free toolbox which was
>     written for matlab so that it can work under Octave.
>
>     Best regards,
>     Sergey
>
>
> I'm also very interested in having this functionality.
>
> I could also help in the testing process.
>
> Regards,
> Danilo

I am not one of the core developers, and I am not working on classdef, 
but I have gathered some info on this.

classdef is (AFAIK) far from completion and is a large task that 
requires excellent C++, a deep knowledge of the Octave interpreter and 
interpreters in general, and considerable time to develop (see 
http://wiki.octave.org/Summer_of_Code_Project_Ideas#Implement_classdef_classes). 
The classdef syntax has been created, i.e. the interpreter can recognise 
it, but it is not actually attached to anything.

It doesn't help that Matlab does not have a language spec as such.

If anyone has means to provide money for the development time, it will 
happen a lot faster (this is how the old-style classes got implemented I 
think), but otherwise will take some time.

Richard




--
The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.