[dojo-contributors] 1.3 direction

Peter E Higgins dante at dojotoolkit.org
Sun Oct 5 09:33:22 EDT 2008

To be honest, your list sound frighteningly involved for the time frame
we're looking at.  We were "way overdue" on 1.2, so time is limited ...
1.3 seems to be decidedly driven by IE8 support / release. Given the
limited window of coding we're given for this release, I have the
following comments (inline):

Alex Russell wrote:
> My big items for 1.3 are:
>     * unifying abstractions for dijit.WidgetSet/dojo.query/etc.
This is cool, but would require a lot of testing and would touch a lot
of the codebase. To ensure we adequately address any potential
backwards-compatibility issues and our own use throughout the source
receive proper attention, I would suggest punting this idea to 1.4
>     * dojox.dtl moved into Core and DTL becoming the default template  
> system for Dijit
I'm not sure it is ready for that?  Granted, I'm admittedly a dojox.dtl
noob, and the whole thing has recently excited me (and others) a lot,
but it seems to lack full documentation.

This is a great guide:
http://dojotoolkit.org/book/dojo-book-0-9/part-5-dojox/dojox-dtl but it
needs to be migrated to the wiki. The API docs seem painfully slim as
well: http://api-staging.dojotoolkit.org/jsdoc/dojox/HEAD/dojox.dtl --

There is also a caveat regarding dtl's use in the x-domain case, where
you need to explicitly require the tags and filters you are using which
imho prevents it from being used immediately to replace the Dijit
_Templated mixin.

I agree after the docs and examples are beefed up it would be a great
addition to Dojo Core, but it is currently marked as experimental, and
clearly has some shortcomings as it stands. In following the spirit of
the basis of dojox.dtl (django), I would urge we get all the docs,
examples, and unit tests in place before we commit to Core.

>     * IE 8 support
This was agreed upon as one of the main driving factors for 1.3.
Allegedly, it will be mid-to-late Novemeber, which only gives us two
months to implement and test.
>     * Editor (ugg)
ugg, yes. But I have faith in you.

>     * package tools (make it simple to distribute extensions,
> incremental  
> re-builds, etc.)
decoupling DojoX has always been on the agenda, afaik, which is how I
interpret "distribute extensions" -- the uxebu folks had a novel idea in
that regard, which might warrant investigation, but I'll let them put
the idea out for discussion, as I'm a little fuzzy on the details.
Incremental rebuilds is an interesting idea, but again, frightening
"changing" for such a short timeframe.

So, to follow up, my 1.3 bullet-list would include:

* IE8 compatibility across projects
* Fixing serious bugs, and going light on features and enhancements
* Getting some of the low-hanging DojoX components promoted to Dijit
(expando comes to mind, Grid hopefully?)
* Getting some of the previously discussed enhancements into Core:
    * dojo.createElement / whateverItWillBeNamed (with full unit tests,
docs, and examples)
    * dojo.parser.instantiate - we discussed a more simple way to grep
attributes off of nodes and create widgets
       based on existing nodes.
     * others?
* DOCUMENTATION! We've come a long way in this effort, though still have
much work to do. This is still one of my top priorities, and the only
place we even come close to being out-done by the other libraries.  I
suggest we spend a lot of the manpower going into 1.3 to ensuring all
the docs are in place, working, readable, API info available, etc,
before putting new items into the code base. It is much easier to
document something before it is released than to play 'catch-up' with
older functionality as we have done in the past. We take a lot for
granted because we know how a lot of the core and base functions work,
but there is still much to be explained. I'd like to be able to point
someone to a documentation page for about any question we can field.

Peter Higgins

> Regards
> On Sep 25, 2008, at 11:27 AM, Bill Keese wrote:
> > Can we contributors reach some consensus on the two or three top
> > priorities for the 1.3 release that we'll all work on?  What do you
> > think they should be?  (I won't be there on Sunday so I thought I'd
> > start the thread here.)
> > Basically, I'm looking for consensus/buy-in from the people actually
> > writing the code, so we can work together towards that common goal.   
> > We
> > did this once before back in the 0.9 development days, when we decided
> > to rewrite everything.  But I don't think we've really done it since
> > then.   The 1.3 delata should be a lot smaller than 0.9 delta, but
> > should still set a general direction.
> > I have some ideas but I figure I'll let others answer first.    
> > Although,
> > BTW, I started a discussion about Dijit 1.3 priorities in
> > http://dojotoolkit.org/2008/09/17/dijit-1-3.
dojo-contributors mailing list
dojo-contributors at dojotoolkit.org

More information about the dojo-contributors mailing list