Re: On the union of two communities - The GNU Octave atmosphere

[Daniel J Sebald]

On 09/14/2012 02:45 PM, Júlio Hoffimann wrote:
> Dear all,

> I'm here to purpose the union of these two important communities (Core
> and Forge), to eliminate this strong separation that makes devs/users
> lives much more difficult. We can make the GNU Octave atmosphere even
> richer and minimize the losses, as nature does.
>
> I don't see reasons to maintain two mailing lists, two disconnected
> repositories, two communities. After all, we're all doing our best to
> provide what we think is useful to us and to others.
>
> Being concrete, what do you think in gradually absorb Octave Forge
> packages into Octave main repo as subrepos? This would require an
> effective collaboration to review the packages, to purge redundancies
> (reimplementation) and i'm here to help, just to mention!
>
> People interested in contribute to the Forge packages would just do it
> inside the main Octave repo, inside a subdirectory. Forge maintainers
> would have total access to the subrepo contributing exactly the same way
> they contribute today, with the difference they would be closer to the
> core, which is great.
>
> The builds would be completely independent, we would just add targets to
> the packages (make forge, make forge-optim, ...).
>
> Please, put all the individual feelings aside and think as a community.
> I'm glad to be part of it.
>
> Best regards,
> Júlio.

Júlio,

The OctaveForge repositories was a discussion item at OctConf 2012. 
Opinions are varied, but I'll try summarizing.

First, I don't think there is the great divide that you are suggesting. 
 There is an OctaveForge repository separate from the core of Octave 
for a few reasons.  Probably the most salient reason is the fact the 
whole project becomes difficult to manage if the core of Octave and all 
the packages are combined into one.  I think there may have been a time 
when scripts were organized and combined with Octave, then it became too 
much and people thought a packaging system was needed.  Mailing lists 
re-arranged quite a bit as well, at first being narrow, then splitting 
into specific function which became irrelevant as some of the mailing 
lists got little traffic.  Development has sort of evolved into the 
current arrangement that seems to function well enough for the time being.

There are three, maybe four levels of Octave code:

1) Core Octave written in C++ (i.e., compiled code)
2) Commonly-used, moderately-general m-scripts (i.e., interpreted code)
3) Compiled or scripted code related to user interface, whether that be 
a graphics engine, GUI/IDE, etc.
4) Voluminous packages of field-related m-scripts

The first level needs to be extremely efficient and well organized code. 
 Being part of GNU software, it must also ensure that licenses are 
observed properly.  Up to this point, the main Octave maintainers (i.e., 
those most active and productive...and those most active has changed 
over the history of the project with people coming and going on the 
basis of available time) had to focus on levels 1 and 2 with blinders to 
levels 3 and 4, otherwise they wouldn't have been so productive.  If one 
were to look, I think they'd find that the most active people on the 
maintainers email list (emails and coding) are the least active on the 
OctDev list, aside from an occasional post to clear up some detail.

At level four the packages are associated with diverse fields with 
contributions from many.  Efficiency, correctness and such just doesn't 
get the robust scrutiny that the code maintained in the core does.


> Before entering the atmosphere, i want to thank you all from outside, as
> an external admirer. You all did a great job providing a free
> alternative (pick any interpretation) for scientists explore the
> boundaries of knowledge with no constraints on what they can or cannot
> do with the tools in hand. Thank you.
>
> Now let's dive in with the minds open, forget the past because it
> doesn't alter anything...
>
> I can smell two types of air. One is more dense, full in substratum, in
> fact it occurs in deeper layers, close to the core where we can find a
> rich environment with beautiful (parser) trees, birds, ... It's
> refreshing. The second can only be smelled on high mountains, and lucky
> are the ones who had the opportunity to feel this sensation.
>
> As a adventurer, i would like to share this sensation with my friends,
> but due to the altitude and faraway places they always give up in
> joining this journey. I insist to them the air found in the mountains is
> sublime, but it's hard to convince people when they are under the effect
> of toxic gases produced in urban centers.

A bit off topic, but on the matter of toxic gases, society can make 
urban centers that are livable, which helps preserve natural spaces.  It 
doesn't have to be one and the other.

Dan