Latest changes to Edge Waves

[Daniel Yoder]

Hi Team,

I think we're getting close to the 0.9.0 release, at long last.

I've continued to make some relatively dramatic changes. The three
biggest things recently are:

1. Pete's Hoshi is now the default way to do views for classic apps.
2. Automatthew's Cassandra is now the default way to do views.
3. My Hive gem is now used to run the server.

The classic template also has a default start-up page so that it is a
bit less mysterious to work with.

Regarding item (2), since Cassandra has been made obsolete (because
Matthew graciously gave up the RubyForge project to the authors of an
admittedly very cool distributed database gem of the same name), I've
simply added it to Waves (as views/cassy.rb).

I've updated all the dependent gems (autocode, functor, etc.) so that
the current gems on RubyForge match the ones that Waves will be
expecting. Yes, this breaks the current gem, but it is broken anyway.
Unless you happen to be using a very old version of Rack.

Of course, there has been a variety of smaller changes as well. At
the moment, the tests are simply not running, due to some refactoring
I did to the runtime. I would imagine that none of the tutorial apps
run at the moment, either.

However, the good news is that I think I am finished with the big
changes. Mostly now I will be working on 3 things, in preparation for
the 0.9.0 gem release:

1. Updating and adding to the tests. I will post more on this
separately.
2. Updating the documentation, including RDoc and tutorials. I'll post
on this separately as well.
3. Fine-tuning specific features, particularly some of the newer ones.

I am also considering reversing the precedence for functor
declarations so that the request matching is FIFO instead of LIFO. My
reasoning in using LIFO was that I was trying to mirror the way Ruby
works when you define methods, so that Functor was basically adding
method overloading. And I still think that was a reasonably sound
approach.

That said, not only does the LIFO approach confuse people, but you can
argue that that we're implementing pattern-matching rules, rather than
overloading, in which case FIFO makes sense. So why confuse people if
you can argue it either way? I'd love to hear your thoughts on this.

I am actively opening up tickets on GitHub, so please feel free to
jump in on any of the three above priorities, or one of the tickets,
or whatever. I'd like to get a stable gem out as soon as possible.
We've squandered a LOT of interest (including a great write-up in a
recent issue of Rails Magazine by Carlo Pecchia) because we don't have
a stable release that works. No sense crying over spilled milk, but I
think this is going to be a really compelling release, so I am excited
about getting it out. I've played around building prototype apps (more
on this later) using edge and it is like driving a Ferrari compared to
using any other framework.

Regards,
Dan

[seneca]

Thanks Dan for all your efforts in bringing 0.9 to life.
Although I didn't manage to play with previous versions of Wave
(couldn't run it) I am looking forward to have a go at it now.

[Daniel Yoder]

Thanks for your patience and keeping an open mind!

Regards,
Dan