Re: [O] ob-async

[Nicolas Goaziou]

Hello,

Alex Bennée <address@hidden> writes:

> Ken Mankoff <address@hidden> writes:
>
>> An RSS feed I follow mentioned ob-async here:
>> https://github.com/astahlman/ob-async
>>
>> I haven't seen it mentioned on the list yet. Perhaps others would be
>> interested in asynchronous Babel processing. I've seen the feature
>> requested often on this list.
>
> This is not the first attempt to my knowledge. I know of:
>
>   - my hacky attempt https://github.com/stsquad/async-org-babel
>   - John Kitchen's python specific version 
> http://kitchingroup.cheme.cmu.edu/blog/2015/11/20/Asynchronously-running-python-blocks-in-org-mode/
>   - this matlab version 
> http://emacs.stackexchange.com/questions/21301/async-execution-in-org-babel
>
> So I think there have been enough proof of concepts of using async.el
> and inserting results at a later date. I think what would be really
> useful is some feedback from the org-mode maintainers about the various
> approaches and if something generic could be included with org-mode
> itself.
>
> Any thoughts?

I never used any of these, so please take this with a grain of salt.

AFAIU, these solutions are too limited at the moment. They suffer from
the same problem as current ":cache" parameter, i.e., they seem unable
to cope with blocks that refer and execute other blocks. 

IMO, for anything serious, we need to implement something that is able
to capture the "closure" of a block. In this case, all blocks belonging
to that closure would be marked as read-only during the process, so as
to avoid race conditions. We also need to be able to retrieve all the
results from all the blocks involved.

Regards,

-- 
Nicolas Goaziou