Re: classdef functionality

[Michael Goffioul]

On Fri, Feb 24, 2012 at 4:39 PM, Jordi Gutiérrez Hermoso
<address@hidden> wrote:
> 2012/2/24 John W. Eaton <address@hidden>:
>> On 24-Feb-2012, Jordi Gutiérrez Hermoso wrote:
>>
>> | On 24 February 2012 08:44, deraltefritz <address@hidden> wrote:
>> | > So I'll ask this once more: will there be classdef support implemented
>> | > sometime (soon) in the future?
>> |
>> | This is the closest we've got:
>> |
>> |     
>> http://octave.org/wiki/index.php?title=GSoC_Project_Ideas#Implement_classdef_classes
>> |
>> | If you want to help, you can mentor or find a student who is willing
>> | and able to do this.
>> |
>> | Assuming Octave gets accepted as a mentoring organisation, of course...
>>
>> I would also be interested in doing this work if someone wanted to
>> hire me to do it.  That's how the "legacy" @CLASS classes were added
>> to Octave.  My experience doing that work tells me that this job will
>> require significant time and expertise.  Probably more than can be
>> handled by a GSoC student unless they are already quite familiar with
>> Octave internals and Matlab's classdef classes.
>
> Could you reword that task into a more modest goal, then? Perhaps only
> implement some classdef functionality? I wrote it initially, but I
> didn't think it was such a difficult task.

I kinda agree with John, this is not an easy task, if you want to do
it right: it's a full meta-foo system to implement (metaclasses,
methods, properties, access-levels, events...). A few years ago, I
started to work on a skeleton and posted the result on the list. You
should be able to find it back in the archive, if you want to have
look.

Michael.