Rails 2.3 Works With Ruby 1.9.1 (stable)
Many people don’t know that Ruby 1.9.1 is actually considered stable (due to the odd numbering). Ruby 1.9.1 is based on the exciting (yarv-based) branch of ruby, which promises to be several times faster than the old version of ruby, and offer new features with respect to threading. Here is What’s new in Ruby 1.9.1
As you can see from the Ruby on Rails 2.3 Release Notes , all of the rails tests now pass with ruby 1.9.1. So what’s the problem? Well, many basic gems and plugins are not compatible with ruby 1.9 yet. Really, what should amount to only a few hours work for each gem maintainer is holding everyone back here, it seems.
This guy actually got Rails to work with ruby 1.9.1, but man, that looks like a hassle: Ruby 1.9.1 and Friends
I’m thinking of putting together an amazon ami to use rails with ruby 1.9.1. Would anyone be interested in that?
The Stale Paradox
My friend, Chris Gorman, came up with this one:
Soft things, like bread, get hard when they get stale.
Hard things, like cereal, gets soft when it gets stale.
So, say you take something that’s already stale, like stale bread, and
repackage it as croutons, they get soft when they get stale, right?
How many times can this cycle repeat itself?
Answer:
from Adam GANDERSON
It’s actually a function of the food reaching equalibrium with the ambient humidity. As such, how often you can “cycle” your food depends on how much/how often the humidity changes. Note that temperature changes can effect flavor chemistry, and bacterial and fungal growth rates are increased in different parts of the cycle. I could go on to water activity versus water content (which takes into consideration concentration of compounds like salt or sugar that “tie up” water.) But that’s beyond the scope of this course.
The short answer? Entropy is a bitch my friend. And she always increases.
Love and felches,
Tappan
Sent from my iPhone
Color Scheme Designer
cool tool
http://colorschemedesigner.com/
HTTP “Server Push” Options Reviewed
So, http has this problem where the server can’t push things to the
client. There are a few ways to get around this if you want to make
something like google chat, or facebook chat.
1. Polling. Swoopo.com does this. Every second basically, the client
does an ajax request for data, and updates the page.
2. “Comet” Long lasting connections. Basically, javascript creates a
long lasting socket connection to a broadcast server that is built to
handle a lot of connections. Rails (or django) passes a message to
the broadcast server, and everyone “subscribing” to the channel, gets
the message.
3. “Juggernaut”, like comet, except that it uses flash to create the
long connection. 64squar.es uses this (chess site)
Grockit.com uses comet for users to be able to interact.
Google and Facebook spent millions on in-house comet solutions to make
the chat work. Facebook claims to have done a lot of the technology
in c++ and erlang. Orbited is an open-source broadcast server, and
RabbitMQ is an example of a messaging service.
All of the options have problems. Normally, I am anti-flash anything.
However, I think I’m going with Juggernaut for a new project, because
I’m willing to trade a few crashed browsers so that 22% of my user
base (ie 6) doesn’t have to hear constant clicking ala comet or
polling.
Apparently, this is a huge barrier holding back sites at the moment,
and google and facebook have spent millions to temporarily overcome it
for a first-mover advantage.
I don’t think chat is really essential to most sites, unless you’re
facebook or gmail. Lack of reliable server push, is holding back a
genre of web applications though. All 3 solutions are hacks and not
very reliable. Google and Facebook have done a lot of work gracefully
degrading the problems so chat gets archived when it’s not delivered,
etc. Chat is one thing, and there are a lot of pluggable chat
applications now, but if server push were simpler, and robust, it
would open up a lot of games and things like group drawing boards.
Right now, there is no barrier to making a digg clone, or a social
network. A framework exists to make these things, and there is no
longer room to create an innovative product based on those frameworks
alone. Server push is something that, right now, if you can iron out
the bugs to still make your application work, creates a barrier to the
product already existing, and other people cloning it.
Still there is plenty of demand for the “old” style web 2.0 apps, and
if you have a project with goals that a web 2.0 app can deliver, I
don’t think server push is worth it.
Just set up wordpress on slicehost
Now, i need to figure out how to plug emacs into this baby.
Sweet, weblogger.el did the trick: weblogger.el patched to support tags
