[gentoo-dev] Re: Two-level portage tree is irrelevant.

[Duncan]

Andrew D Kirch posted on Mon, 24 Aug 2009 12:14:56 -0400 as excerpted:

> Dmitry Grigoriev wrote:
>> http://bugs.gentoo.org/show_bug.cgi?id=282491
>>
>> The idea is that package tree physical structure must correspond to
>> logical structure. E.g. package kde/games/tactics-and-strategy/knetwalk
>> instead of kde-base/knetwalk, and kde/games/all instead of manually
>> managed meta-package or set @kde-games (kde/all == @kde,
>> kde/games/arcade/all == @kde-games-arcade, ...).
>>
> I don't see a problem with this per-se other than that the massive
> amount of re-organization which would be required, which could otherwise
> be spent on fixing bugs, adding enhancements, and other cool stuff.  I
> think the price is too high in the manpower catagory.

General observation:  In the FLOSS community it is often said (correctly) 
that for something to be done, it normally must scratch an itch that 
someone with the skills to do it has, an itch bad enough to motivate the 
dedication of the necessary time and intellectual energy.

It's thus that there are all sorts of otherwise impractical little 
projects going on, some to  eventual usability, some to eventual full 
maturity, some dying on the vine, as it were.  It's the incredible 
broadness of the community, and thus the incredible broadness of 
selection of all those little projects, that continues to drive FLOSS, 
generally far more broadly and effectively than it could ever be driven 
in an unshared or charge-to-share primarily cost/payment driven 
proprietary system.

You see this put to great effect in the firefox extensions setup.  
There's dozens of browser choices, but really only one with the 
extensibility of firefox, an extensibility that many users quickly find 
indispensable, thus making firefox itself indispensable for those users.

The same applies to some Gentoo projects.  Realistically, how many of 
those exotic archs we support, if only in -prefix or experimental form, 
would even exist at all, if they had to be cost and time justified?  But 
they are someone's hobby, Gentoo is a volunteer organization, and those 
devs have volunteered to make their hobby yet another Gentoo project.

Thus we get to the point.  I agree that it's not particularly practical 
to think about how the Gentoo tree might be better organized if we were 
to do it today.  However, if someone with the skills and the drive to 
make it so can be found, that either has that itch bad enough, or can be 
/given/ that itch bad enough, to actually /make/ it so, well then, it's 
likely to happen.  Otherwise, no, it's not, as however nice it might be 
in theory, there's always higher priority more effective ways for those 
with the skills and the access to make it happen, to spend their time.

Thus, the OP's mission, should he choose to accept it, is to either 
develop the skills and become a Gentoo developer himself, thus giving 
himself access to do it, or to effectively enough spread that itch to 
someone who has the skills and the access, thus giving them the 
motivation necessary as well.

Otherwise, I agree, it's simply very unlikely to ever happen, because the 
solution we have now is "good enough" and the cost of changing it and 
taking care of all those loose ends to make it work is high enough, that 
there are always better ways to spend that time and energy.

-- 
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman