Thursday, August 30, 2007

New Fast JSON-processor: Jackson!

Developers, meet mr. Jackson: possibly the world's fastest JSON parser! (and at very least, fastest one written in Java).

You heard it here first folks: a new streaming light-weight, and VERY fast JSON-processor (parser+generator) package written from scratch, in Java, is now officially out in the wild. Check out Jackson Hatchery Page for details.

So what's in it for me? This new release will eventually develop to be a similar high-quality, easy-to-use and widely used building block as Woodstox is. And that will happen fastest when there are fearless Early Adopters who check out new things.

Stay tuned: I will try to cover more ground with my next entry. For now, I just wanted to get the word out.

Oh, and please do send me feedback if you end up using it. While basic fundamental unit test coverage exists, there are bound to be some rough edges.

Monday, August 20, 2007

More fun integration: Antlr with Stax!

Here's another interesting-looking article, toying with the idea of generating a parser that takes xml tokens (elements, attributes) as its lexical tokens, to parse higher-level constructs. ANTLR is a fine parser generator in its own right, so this could be something akin to peanut butter and jelly meeting to make a great sandwich (even if analogue here may be rather cheesy and naive). :-)

Anyway, in addition to coolness of the almost-recursive use case of "Using parser generator with a parser", this actually seems potentially useful: to me, state machine based systems seem like an obvious way to generate higher level abstractions on top of xml structure. So hopefully someone has time to explore this approach in more detail (I'll have to keep this mind as a potential mini-project myself).

Tuesday, August 14, 2007

Efficient XML processing with ColdFusion, Woodstox, XOM

Here's yet another interesting article regarding applications of Woodstox: XML StAX Processing with ColdFusion. It's not exactly new, but I just now found it. I remember original work by Jim Collins, when he was integrating Stax implementation with Cold Fusion. It sounded like it could be useful for situations where resource usage matters, such as when dealing with sizable (multiple megabyte) xml documents. Alas, it sounds like this wasn't fully finished, or at least not fully published.

Using XOM (and Nux library as the glue in-between) may also make sense, although it may remove many benefits of streaming processing. However, given the high-quality xml processing API XOM offers, as well as its customizability -- for example, it is quite easy to make it only build sub-trees of the source document, not the whole document -- this should be a nice addition to a CF developer's toolbox.

Related Blogs

(by Author (topics))

Powered By

About me

  • I am known as Cowtowncoder
  • Contact me
Check my profile to learn more.