W3C home > Mailing lists > Public > spec-prod@w3.org > April to June 2010

ReSpec 2 repository is up

From: Robin Berjon <robin@berjon.com>
Date: Tue, 4 May 2010 17:26:56 +0200
Message-Id: <497334F6-5F16-41E2-8C62-6C9D867590D0@berjon.com>
To: Spec Prod <spec-prod@w3.org>
Hi all,

thanks to the W3C Team, development on the next version of ReSpec can now be done using Mercurial instead of CVS. Woo!

You can clone a copy:

  hg clone https://dvcs.w3.org/hg/respec2

And browse it online from :


Any one with a W3C account should be allowed to push changes.

This is *NOT* intended as an immediate replacement for ReSpec v1, it is not even alpha code yet. That being said the general architecture has been worked out enough that it can already do useful stuff (in the lab). Of particular interest:

  - It is now completely modular, in fact it is essentially a plugin system that happens to grok specs. This means that it can be used to produce specifications outside W3C (as has been requested) and also that different WGs don't need to share the same settings but can add their own plugins without needing Politburo intervention.

  - The references system is a lot less brittle than previously (no more collective editing of a long JSON file).

  - It uses a templating system, you can say goodbye to endless lines of string concatenation.

  - It builds on jQuery, which does make some of the code a lot simpler.

  - IE-compatibility is now a goal (which doesn't mean it currently works, it just means that you can scream at me when it doesn't). For people just viewing ReSpec drafts I have hope that it'll work down to IE6; for editors who preview their work from the local file system it probably won't work below IE8 (due to using postMessage()). I guess that that is acceptable (note that you can always run a local web server if you need to).

  - Performance will be more variable, but it should feel at least slightly better. The reason is that before the CSS and the references DB (all of it) were loaded synchronously, not matter how many references you actually used. Now they load asynchronously which means that the document starts normally looking fine pretty quickly even though the total load time might be longer (with one HTTP request per reference).

  - I haven't checked any of it in yet, but I actually have some QUnit-based unit tests brewing.

Things that need doing before it ships (for more detail look at the source):

  - Browser compatibility testing (it's only been tested in Firefox 3.6 so far).

  - Port over: WebIDL processing, several of Shane's features, syntax highlighting, and saving the document.

  - Documentation will need updating to describe the new import format and architecture.

As always, patches, comments, etc. welcome!

Robin Berjon - http://berjon.com/
Received on Tuesday, 4 May 2010 15:27:25 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 19:55:15 UTC