Skip to Content.
Sympa Menu

sm-discuss - Re: [SM-Discuss] SCM tools...

sm-discuss AT lists.ibiblio.org

Subject: Public SourceMage Discussion List

List archive

Chronological Thread  
  • From: Andrew <afrayedknot AT thefrayedknot.armory.com>
  • To: sm-discuss AT lists.ibiblio.org
  • Subject: Re: [SM-Discuss] SCM tools...
  • Date: Sat, 11 Mar 2006 11:15:23 -0800

On Sat, Mar 11, 2006 at 09:02:15AM +0100, Pieter Lenaerts wrote:
> Op vr, 10-03-2006 te 14:18 -0800, schreef Andrew:
>
> > Theres no reason we can't just use searchable keywords to help users
> > find spells, and possibly to organize them into a multi-level symlinked
> > hierarchy. That would accomplish the goals of option 3 *without* having
> > to move spells around in the repository. Also, as evidenced from past
> > grimoire re-org discussions, categorial organizations that make sense
> > are different for every person. Trying to come up with one that satisfies
> > everyone isnt going to happen. Its a loosing proposition.
>
> while I of course understand and acknowledge the advantages of 1 and 2 I
> think option 3 is the simplest way for users and we should stick with
> that.

You make the assumption that we're currently using option 3, when in
fact we're not even close. This is demonstrated by the quite long and
unresolved discussion about grimoire re-org that took place many months ago.

We actually have some hybrid of the three options. We have some sections
that are setup almost purely based on code reuse (devel-python, pear-php,
etc.), some sections based on one particular vector of commonality (such
as a group of software, xfce or e for example), and some sections mostly
for convenience of ownership.

Looking at the bigger picture, everyone has their special ideal
interpritation of how the grimoire should be organized (you and I included).
What Im trying to do is point out where those views come from and
the underlying motivations behind them. Whats happening now is everyone
has their own idea and they're trying to run that direction with it,
but not everyone is running the same direction, so imagine vectors in
physics going different directions and canceling each other out.

>
> imo the keyword solution is just making things more complicated without
> giving a real advantage.

How does it make things more complicated?

>
> I think the code reuse can (and is atm?) achieved in a different way
> than throwing away our sections.

Lets be clear. I *NEVER* said we should throw away our existing sections.

Theres plenty of code re-use in the grimoire right now, Im not alluding
to some magical new technology, this is something we do right now. What
I was pointing out was that we should do more of it, because by doing
less of it, we're shooting ourselves in the foot.

Why? Well, because i think its easier keeping our existing code-reuse
structure, and adding alternative views on top of it, seen by users,
than it is to try and setup new code-reuse structures on top of a
(hypothetical) view that everyone agrees on, which in reality probably
wont be very conducive to code-reuse. Assuming of course, we can agree on
some way to categorize spells ala option 3 (demonstratably so, we cant,
we tried, we couldnt get a consensus).

In other words, lets not try and solve that problem here, solve a
different one with our native grimoire layout, and solve the searchability
problem elsewhere. If we solve the searchability issue some other way,
we can avoid having discussions like this. Its also much easier (i would
argue) adding keywords for new categories than it is trying to move lots
of spells around in p4, getting people to agree on whether or not we
should have a section or not etc.

The discussion would just be "i want to add a new keyword 'scm' to these
spells, any objections, deletions, or additions?" That seems a lot more
managable to me, maybe I'm wrong, i donno.


>
> As of 2), I think per spell maintainership as a general rule will end us
> up with only a couple spells that have 5 candidates for maintainers, and
> all other spells with no maintainer. if I just look at editors section I
> can point out that vim won't have a trouble finding a maintainer, but
> who uses cooledit? who would maintain yudit?

And thats different than how it is now? Most sections are
unmaintained. Why? well the main thing I hear is that either the section
is too big, or the developer doesnt want to maintain stuff they dont use
(how many editors do you really use?). People are here to have fun,
we cant make them maintain stuff they dont want to maintain. So, why
force people to maintain whole sections when we know it doesnt work
except for with the most dedicated people.

Also, we've got around 4000 spells, and maybe 20 people maintaining
them. 200 spells per person. Theres going to be unmaintained stuff no
matter how you stack it. Putting faith in the "people maintain a section"
paradigm, when we've got limited *volunteer* resources doesnt seem like
its working.

In anycase, we dont have to do anything about changing maintainership
structure, I was just pointing out that its something to think about,
and things dont have to be the way they are. As I said, we can set things
up so its more obvious whats not maintained, and who is actively maintaining
what.

>
>
> > In other words, lets not bother trying to find the one true categorial
> > organization scheme. Organize by code-reuse first, then by
> > maintainership. Then use keywords to accomplish categorizing things (as
> > opposed to picking one of a half dozen categories something could fit in).
>
> keywords will be subject of discussion then

In the past it was agreed that keywords are a good idea.

>
> the argument that you can't ever find "the one and only correct section"
> for a spell simply points out that this is impossible for all spells: in
> real life things aren't all black or all white, so we shouldn't make a
> priority out of it iyam.
>
> so far for my rant, I think we should stick with our simple,
> hierarchical solution as it is now, and not complicate things.
>

My suggestion was actually for keeping our current grimoire setup mostly
intact. It mostly lends itself to code reuse as the path of least
resistance. Keywords were a generally accepted idea in the past. And
lastly, Im missing a lot of support behind the argument that things will
be more complicated. I'm not sure where that idea comes from...

-Andrew

--
_________________________________________________________________________
| Andrew D. Stitt | acedit at armory.com | astitt at sourcemage.org |
| irc: afrayedknot | Sorcery Team Lead | ftp://t.armory.com/ |
| 1024D/D39B096C | 76E4 728A 04EE 62B2 A09A 96D7 4D9E 239B D39B 096C |
-------------------------------------------------------------------------




Archive powered by MHonArc 2.6.24.

Top of Page