[Dojo-interest] The state of dojo

[Alex Russell]

On Thursday 13 January 2005 12:10 pm, Dror Matalon wrote:
> Hi folks,

Hi Dror,

> We provide DHTML tools which you can see at
>
> http://www.zapatec.com/website/main/products/prod1/
> a calendar
> 	and
> http://www.zapatec.com/website/main/products/prod2/
> a DHTML tree (Beta).
>
> We have our own utils library
> http://www.zapatec.com/website/main/zpcal/doxygen/utils_8js.html, but
> would consider using dojo down the line.

What kinds of features would make Dojo palletable to you for those uses? I 
could list off a set of things that we'll be providing, but it seems more 
useful to me if you could articulate the what and the why of your potential 
decision.

> As far as dojo goes ...
>
> I looked at the website, I looked at the developers mailing lists,
> and I looked (some) at the cVS repository, and I still don't have a good
> picture of where dojo is and where it is going.

Yeah, we're not doing as good a job as we should about documenting all of our 
discussions. I'm trying to correct some of that through "requirements" 
documents, but my success rate there is pretty low.

> So could one of the developers could give us some of the following info:
>
> o What has been already developed

A "bootstrap" for multiple script interpreters. One of the major goals of Dojo 
is to provide a toolkit that is independent of HTML as a rendering context 
(although that is the primary target). We want Dojo to allow authors to 
construct mixed HTML/SVG applications using only Dojo widgets. Or even 
SVG-only apps. Likewise, the same declarative DojoML (the XML declaration 
syntax) should allow you to render the same widget in different context, so 
long as there is a rendering available.

Unit testing system, including fake DOM implementation. This mainly comes from 
Mark Anderson, who wrote the Burst toolkit after much (sometimes painful) 
exposure to netWindows. We've integrated these into the build system, 
allowing for the first time our project to do command-line unit testing of 
not only pure JS, but also DOM-specific features. Tom Trenka is helping us 
out here right now by hooking this up to an XML parser so that we can test 
the DOM-parser portions of the toolkit.

The beginnings of a markup language, and a first-pass parser for it. This XML 
dialect will (optionally) be used to declare Dojo widgets, and it'll have 
translations down to HTML and SVG so that, like netWindows, you can use Dojo 
inline with other HTML and not have to worry about backward compatibility or 
accessability.

An event system. The new Dojo event system is an improvement over the systems 
provided by f(m), netWindows, and Burst. Those familiar with Aspect Oriented 
Programming will find it interesting and useful. Everyone else will just 
think it's really handy = )

Custom modifications to Rhino to support packaging and documentation. More on 
that in another mail some other time.

> o Where would you like to be in 6-12 months

A shipped version of the toolkit with a reasonably complete set of widgets, 
tools, documentation, and back-end support infrastructure. I'm fortunate in 
this because my employer (Jot) is funding me to work on Dojo full time.

> o how many people are actively involved in dojo development

I guess I don't have a number offhand, but let me think about that a little 
bit more. It's surely less than 10, and more likely in the range of 5.

> o anything else that you think would give us a better picture of dojo.

Yeah, Dojo isn't necessarily about doing new things, but rather getting the 
DHTML community all pointing in one direction and backing a single set of 
widgets, tools, and core code which is liberally licensed. It's my firm 
opinion that webapp authors need an easier way to build web apps that are 
still true WEB APPS (i.e., not flash). Dojo is being developed to give us 
better tools at a high quality level without the cost of tying yourself to 
one of the proprietary DHTML toolkits. This stuff should be a commodity, and 
in order for apps to become more responsive, it needs to be.

That's what we're all about. Great tools for people building responsive apps. 
And we need your help in defining what "great tools" actually means.

Regards

-- 
Alex Russell
alex at jot.com
alex at dojotoolkit.org