What a busy couple of months we've had! Nothing brings me more pleasure than to push a bunch of code to an open community, and today I get to do exactly that. I am proud to announce the release of the first of the Dojo 1.3 series: Dojo 1.3.0. Buen Provecho.

This release is monumental. Though the release notes only cover the additions and possible changes from 1.2, the overall implications of the fixed issues and new functionality are staggering. The focus of 1.3 was Internet Explorer 8 and WebKit (Chrome, Safari) compatibility and general performance, though the overall scope of Dojo invariably states a slew of new widgets, functions and API cleanups have landed, and are available now on two quality content distribution networks out of the box. While I could carry on about the new exciting features in 1.3, I'd rather you just try them ...

To load Dojo 1.3 from the Google CDN add:

to any existing webpage. Or from the AOL distribution:

Both are identical, and both are lightning fast ways to add a whole host of concise JavaScript API's to a page. The public APIs provided by dojo.js are in the download directory as a handy cheat sheet covering all that is "Base Dojo". No summaries or other text at the moment, a full "base reference guide" is in the making for that -- the default print styles in the cheat sheet export to a simple PDF on two pages.

New Base API's

Without going into detail on features and APIs new in 1.3, I'd like to take a moment to expose some new base DOM API's we've put in place to make simple tasks even easier: dojo.create, dojo.destroy, dojo.empty, as well as added magic to dojo.place.

These new functions are super fast convenience API's and should allow you to write incredibly concise code. We were able to eliminate lines upon lines of code internally by switching to the new APIs, and we're not even 100% complete auditing DojoX. Developer convenience, ftw!

Acme, Sizzle

Dojo now has a new selector engine, Acme, which offers considerable speed improvements to an already-fast dojo.query. Alex's work there is amazing. To summarize: AST-like inspection of queries to always find the fast-path, before even starting. The Sizzle engine from John Resig is also available as a build-time option and works transparently as the underlying selector engine for dojo.query.

Not only is dojo.query faster than ever, that which you get back from a query (a dojo.NodeList) has been similarly quickened. Eugene's work there both improved overall performance and provided a series of developer API's to map base "single-node" functions directly into NodeList. the _adapatAs* series of functions are hot, and should make advanced plugin-making even easier and eliminate a lot of code duplication.

Documentation!

Documentation has been seriously updated for Dojo 1.3.0 -- The Uxebu folks have generated a client-side API parser and viewer covering Dojo, Dijit, and Dojox (and even your own code, provided you follow our syntax). The output from the internal doc parser (powered by the inline documentation) is available from the 1.3.0 download section as well, though is just pure XML. Call to arms: If you have uber-XSLT skills, we'd love to see what you can generate from api.xml.

The user documentation from the "live" DojoCampus docs are being tagged for the first time, and despite having some underlying usability issues (we're working on it!), the content is plentiful and accurate. We'll place them live on dojotoolkit.org as we deprecate the "old book", though they are provided already as an archive for offline viewing. Thanks to all the Uxebu team for getting all this together. The Dojo Toolkit, when you consider the coverage and scope of Dijit and DojoX, is a huge documentation effort -- coherent and unified documentation is a huge win (and will only get better). Thanks to Remi, Nikolai, Jared, Klipstein, Bill, Alex, Peller, Gary, Chris, Ben, and all the other many contributors assisting with the doc migration and writing. It is not an easy undertaking.

Additionally, I have updated the "unofficial" Komodo Editor code-completion file. The official plugin version should appear from ActiveState shortly.

Thanks to all the developers and contributors for all their hard work/long hours/general rockin' that went into making this yet another solid Dojo Toolkit release. Huge thanks also go out to the AOL and Google teams behind the CDN deployment -- they are a great service to many developers and users worldwide.

As usual, the 1.3.1 follow-up will be coming shortly -- addressing anything missed by the beta/release-candidate cycle as well as the updated translations for Dijit's localization/internationalization efforts.

On a somewhat related note, I'd like to take a moment to mention/announce plugd 1.3.0. The plugin is available as a standalone "dojo.js" replacement (just throw it in your project on top of dojo.js!) or as a 2k adapter to load a bunch of solid additional base API's. We'll be continually working to merge the best parts of plugd back into Dojo land, but for the time being they are provided as a separate entity, and are fully documented.

Trunk has once again been re-opened to enhancements, but contributors remember: absolutely no new enhancements or functionality without documentation, unless the module is clearly marked pre-beta/experimental (and even that only can happen in DojoX). We've caught up to the retroactive documentation needed and can stay ahead of the game if we all follow that one, simple rule.

Looking forward to a the 1.4 development cycle. We already have lots of goodies lined up and ready to land and be documented, and lots of ideas for performance gains across the rest of the toolkit.