[Rubygems-developers] What plans now?

[Gavin Sinclair]

On Tuesday, May 18, 2004, 2:19:36 AM, Ryan wrote:

> --- Gavin Sinclair <gsinclair at soyabean.com.au> wrote:
>>  
>> Everything I've seen with continuations has been rather difficult to
>> understand at first.  I get it after a while but can never imagine
>> coming up with the code myself.

> I spent several hours a few weeks ago trying to wrap my head around
> continuations and how they work in Ruby, but they still confuse me as well.
> They just require a different way of thinking, and that can take a little while
> to develop in one's mind.

> Regarding the overall topic, I'll put in my two cents and some of my original
> thoughts about RubyGems.  My basic concept of gems was that they would be
> self-contained packets of code, with some amount of meta-data so they were a
> little "smarter" than a normal container.  But they would still primarily be
> acted upon by outside agents, whether those agents were installers, servers,
> documentation browsers or what have you.

> So I'm not sure I like the idea of gems being able to install themselves, etc.
> I consider that outside their realm of responsibility, and it overly
> complicates the concept of a gem.  If we develop some complex API where gems
> can install themselves, resolve their own dependencies, serve themselves to
> other machines, etc., we may put people off with the complexity.  I know I can
> get a bit overwhelmed when trying to understand some new API and I come across
> some "uber-object" that can do everything.  Plus it really isn't OO, is it?
> Sounds more procedural since you are putting everything in one place.

> Maybe I just misunderstand what you are saying here, but I do think it requires
> more thought.

You're quite right.  I think we're on our way to finding the right
balance.

I suppose, though, that if a gem can't install itself, then it can't
know whether it's installed.  But perhaps the repository can answer
the question for us.

Cheers,
Gavin