Nelz's Blog

Mah blogginess

JavaOne - Wrap Up (Entry #2)...

TS-6029 Beyond Blogging: Feeds in Action

This talk was pretty cool. It started by comparing and contrasting RSS and ATOM. It turns out that ATOM is generic enough to deliver all sorts of info, not just publications of blogs (for example).

The speaker also went into why REST is such a basically appealing technology: It capitalizes on the CRUD (create, read, update, delete) capabilities already present in HTTP technology.

To start implementing some of this stuff, ROME is a java library that enables both publishing and consuming feeds.

TS-6836 Amazing Web Interfaces with AJAX

Maybe I’m just jaded after seeing the GWT stuff, but I didn’t pull a lot out of this talk.

I guess these were the guys that initially introduced XmlHttpRequest-centric AJAX approach to JavaOne over the past couple of years, but they said they specifically updated the session because the community in general is beyond needing that instruction any more (and I agree). They also said that the term "AJAX" has come to mean more than just the original "Asynchronous Javascript And XML", but that it now was a generic term being used (and abused) to mean "rich user interfaces".

They showed the usage of a couple of Javascript(/AJAX) libraries such as Scriptaculous and Prototype…

TS-4249 Top 10 Ways to Botch Scalability & Reliability

These slides are available, so I won’t re-hash them.

I did like that the presenter started off by saying EVERYTHING he said should be taken with a grain of salt. "No matter what, at some point these principles will be wrong." It was a good reminder or caution that you shouldn’t blindly apply any principles to your code without understanding their essence.

One of the slides also said something like this: "Do no wrong. If you do need to do wrong, understand the whole context around the wrong as well as the affects of doing the wrong."

TS-6536 Identity 2.0

OpenID is the technology being presented here. It is a pretty cool technology that allows the user to control their data across the web (or at least with the services that use OpenID).

There are a few great user benefits. Instead of silo-ing their identities per service, they can consolidate and/or discriminate their identities themselves. Also, their "reputation capital" can be used across the web. (Example: a well-known and well-ranked wiki editor may be able to join a new wiki system with better-than-lowest permissions.)

I think this is a really good idea, and I will be recommending it to any and all public websites that will listen to me… :-)

An aside: I looked at their website, and I think there’s a lot of good information there for engineers… I don’t know that there a lot there for non-technical folks, so I may have a bit of a chore presenting it as an idea to the "product" folks at my work. I’ll recheck the site, and if what I come up with can help, maybe I’ll try to submit it.