---476953799-2119418668-1226771162311
Content-Type: MULTIPART/MIXED; BOUNDARY="-476953799-2119418668-1226771162=:2311"

  This message is in MIME format.  The first part should be readable text,
  while the remaining parts are likely unreadable without MIME-aware tools.

---476953799-2119418668-1226771162311
Content-Type: TEXT/PLAIN; charset=ISO-8859-1
Content-Transfer-Encoding: 8BIT

On Sat, 15 Nov 2008, Luis Lavena wrote:

> On Nov 15, 1:36 ¨Âí¬ Äáöå Ôèïíáó ¼ä®®®Àðòáçðòïç®ãïí÷òïôåº
> > On Nov 14, 2008, at 4:32 PM, Luis Lavena wrote:
> > > Plans for 1.9 are contemplated but not for the current One-Click
> > > codebase. [...]
> > What should I recommend to PickAxe readers? I'm planning to ship the > > book when 1.9.1 is stable, and I'd like to give them good advice.
> 
> As you may know One-Click bundles almost the world in Rubyland. Some
> of the things bundled has not been updated (by the project developers)
> since 2006 and are still bundled "for compatibility with previous
> releases".
> 
> Doing a release for 1.9 series requires review of all those packages,
> some not being updated for 1.9 and others that lacks complete set of
> testing setup that can ease the process of review.
> 

Can I suggest you make more "noise" about this, please?  It needs to 
get out on the Rubyforge site, on the Wiki, and elsewhere.  This is
a serious problem affecting open source and community based software,
and those who object to this model of development have a strong case
when they speak of "abandonware".
The positive things that will help, I think, are:

List the affected projects.
   People may have forgotten their stuff got bundled, and if it 
   got bundled because of something else, they may not know.

For each, list what needs to be done
   Update for new language features, compatibility with new library
   versions, build of test harness, probably other stuff, (docs?).
   Maybe you need people to test stuff on Windows Vista Home, Windows
   Vista Business, and Windows Vista Ultimate, on 32 bits and 64, on
   Athlons and on Intels, on versions of windows as old as 95, on
   netbooks and servers, on Wine on Linux.

If you can tell us what tools we need, what versions, that's even better.
   Is it practical to make a One Click developer's kit for the One Click
   installer?  A One Click tester's kit?  This is about lowering the
   barrier to getting help.
   Maybe the Rubicon project can be resurrected, it's descendants
   contacted, (Tattle?) so that the stats and bugs can be collected?

This is probably a smaller task than solving the problem itself
directly.  We need this big task sliced up into small enough bits
so that people feel they can contribute.  Everyone is busy.
People will find different bits trivial to handle, something that
looks like a month long project for one person may be doable in an
afternoon by someone working in that area.
Maybe as you get started documenting what needs to be done, people
will get the idea and contribute to this preliminary stage.
Perhaps adopting this model for a status page would be useful:

http://ruby-doc.org/stdlib/status.html

I wonder if you can turn some of these things into small enough projects
to be deemed fit for the Ruby Quiz? 

We really need some sort of call-to-arms on this.  I wonder if Dave and
Andy would let us adopt the slogan "Don't live with broken Ruby on Windows"?

        Hugh
---476953799-2119418668-1226771162311--
---476953799-2119418668-1226771162311--