On Sun, 2005-11-13 at 22:34 +0000, Stuart Herbert wrote:
> On Sat, 2005-11-12 at 10:26 -0500, Chris Gianelloni wrote:
> > If users are interested in non-critical information, there's already a
> > mechanism in place for them to get such things.  They can join the
> > mailing lists.  Do we not already have a gentoo-events list?  We also
> > have a gentoo-releng list, or gentoo-announce.
> 
> At this point, I think you're suggesting that we different news carried
> by different mediums.  If so, I think that's very different from the
> proposal I'm putting forward.

I thought your proposal was to get critical information to our users,
not force every user to read that $dev is going to be in $country from
$date1 to $date2.  As I understood it, the portage-delivered news would
be 100% tree-related and not filled with nonsense.  If I am mistaken in
this, then I change my opinion on supporting this proposal, as I surely
don't give a damn about some dev meet in the UK that I would never be
able to attend and *definitely* don't want that *shoved* down my throat
by the tree.  I also noticed how you lost context in my quote by the way
you quoted it.  Thanks.

> > > I'm not hoping for a 100% perfect technical solution straight away.
> > 
> > I am.  Anything less at this point is a half-assed design.  The *design*
> > should be 100% from the start.  While implementation can occur in
> > stages, you should not design as you go.
> 
> I think that's a worthy goal, but looking around, it looks to me that
> software design just doesn't work like that in real life.  Designs have
> to adapt and change as time passes, not just implementations.

Really?  I work with quite a few developers where I work.  We have
meetings.  During these meetings, requirements are hashed out to cover
the scope of the project.  The code is then written to the
specifications.  If a later change is made into the requirements, then
another meeting takes place, and a change request is agreed upon and
scheduled.  They sure as hell don't let the requirements slip otherwise,
or they would end up in the ever-developing and never-completing world.

We're talking about a *very* simple set of things that need to be
developed here.  Why *would* we even consider not laying out the
requirements up front?

-- 
Chris Gianelloni
Release Engineering - Strategic Lead
x86 Architecture Team
Games - Developer
Gentoo Linux

Attachment: signature.asc
Description: This is a digitally signed message part

Reply via email to