Re: Implementation of importdata belonging to Octave Core or Octave-Forge?

[Jordi Gutiérrez Hermoso]

On 22 October 2012 11:24, c. <address@hidden> wrote:
>
> On 22 Oct 2012, at 15:01, address@hidden wrote:
>>
>>> If you ask me, I think that including importdata in core so directly
>>> wasn't really good. As I see it one of the roles of Octave Forge is
>>> to provide a time for testing and eventually merge it into core.
>>
>> Is that the intended use case for OF? My reasoning was that importdata
>> is a core Matlab function, so it should be in core Octave. I didn't
>> think quality went into it. We already have tons of crappy code in
>> core, and we have tons of amazing code in OF.
>>
>> Is the problem that you would like to work yourself on improving
>> importdata but you see it as some sort of obstacle that it's in core?
>> If so, we should work instead to make it easier for you and others to
>> contribute to core functions. Supposedly, working with a DVCS is
>> supposed to be one step towards making this less of an obstacle.
>
> I think the problem here is rather different.
>
> Having importdata in core requires some more work (to fix
> dependencies and coding style issues), no-one appears to be willing
> to do that at the moment except the original contributor who,
> though, needs some help and assistance.

I have done the style fixes myself. They weren't that difficult to do,
and the original contributor hasn't said anything else after I put the
function into core.

> On the other hand the function as it is is already usable and it if
> it were committed to Forge or pasted on Agora it would be readily
> available to interested users (much before Octave 3.8 or 4.0 are
> out)

They can also just copy-paste it from here, but I suppose this is more
difficult to find:

    
http://hg.savannah.gnu.org/hgweb/octave/raw-file/e21b21c2a20d/scripts/io/importdata.m

> When some developer will have time to do the work to include the
> function and its dependencies in core, he could then just easily
> pull it from forge and adapt it.

I already did that. It's just that I did it before the function was
any good. I can keep doing it if someone else wants to make it good.
At least Philip has already caught one mistake I did and the thing
actually works, albeit horribly slowly.

> Of course if, on the other hand, you ARE willing to work on
> importdata yourself now or to assist Erik in doing so, then pushing
> to Octave was indeed a good choice.

I think that's the current situation, yes.

- Jordi G. H.