[ka-Map-users] Re: [Dojo-interest] Proposal to merge AJAX framework
libraries
Alex Russell
alex at dojotoolkit.org
Wed Jul 6 13:56:21 EDT 2005
On Tuesday 05 July 2005 2:03 pm, Cameron Shorter wrote:
> Hello AJAX projects,
Howdy,
> As you can see from the list below, there is a large number of AJAX
> libraries that a new developer can choose from and the task of
> reviewing all of them to make a selection becomes a daunting task.
Agreed. This has pretty much killed most attempts at web app UI
templating in the Python community.
> I assume all projects share similar functionality
I'm not sure this is a valid assumption, and even if it is, I think that
there is a broad spectrum of views on how generic such a kernel
can/should be constructed. Dojo, for instance, has a completely generic
JS "core" which is specialized for different host environments
(browser, command line JS, SVG, etc.), but even the small amount of
machinery required for this is likely to strike many people who are
just starting out as "too complicated" (who knows, they might just be
right). Prototype, OTOH, is browser-only from the ground-up and bakes
many assumptions into it's core as a result. These kinds of things may
turn out to be irreconcilable differences.
> and it would make
> sense to pool resources and factored out the common elements into a
> core library (library set). All projects could be based on one well
> maintained core project.
Licensing is challenging in this regard. A lot of people have chosen the
GPL either because it's a low-effort decision or because for one reason
or another they want to retain the exclusive ability to license a set
of wholly owned code differently (say, for a commercial customer, i.e.,
the Trolltech model).
OTOH, many organizations would prefer something that's licensed in a
BSD-style or is clearer in the JavaScript context (what does "linking"
mean in a web page?), which makes the "mix and match" conjoining
proposed here difficult. Are you prepared to reject or revert patches
w/ dubious or unclear licensing and lineage?
To organizations looking to build applications on top of any such
platform, being able to ensure a "clean" lineage is tremendously
important.
> I've set up mail list/group at
> http://groups-beta.google.com/group/ajax-web-technology to discuss
> the development of a collaborative project. Please join and
> introduce yourself if you are interested.
>
> I have not been able to contact some projects who probably should be
> involved. In particular, Ruby on Rails, Google, Yahoo. Feel free to
> forward this message on to interested parties.
>
> I've been motivated to write this email because I want to factor AJAX
> functionality out of http://mapbuilder.sourceforge.net and
> incorporate it into an existing library, then use that library.
> We have designed a Model/View/Controller design pattern into the
> browser where widgets are updated whenever an XML model is updated
> (by user or serer refresh).
> Widgets are rendered by the browser using XSL.
> We limit the number of widgets downloaded to the client to those used
> by using a configuration file.
Why not just use a package system that satisfies the dependencies for
you so you only ever get what you need?
Regards
--
Alex Russell
alex at dojotoolkit.org BE03 E88D EABB 2116 CC49 8259 CF78 E242 59C3 9723
alex at netWindows.org F687 1964 1EF6 453E 9BD0 5148 A15D 1D43 AB92 9A46
More information about the ka-Map-users
mailing list