Horst Herb wrote:
> If you are still in doubt whether browser based web-apps can be as responsive 
> and rich as desktop apps, check
> http://www.google.com/webhp?complete=1
> 
> (that is, if you are not already used to full featured Ajaxian web apps like 
> gmail etc.)
> 
> After you have been impressed, rest assured that the performance within an 
> intranet is instantaneous, just as classic desktop GUI
> 
> Now, observe the power of RoR as development platform:
> http://www.rubyonrails.org/screencasts
> 
> Still in doubt?

No doubt that Ruby-on-Rails (RoR) is good stuff, and absolutely no doubt
that sensibly AJAX-enabled Web frameworks based on dynamic languages are
the way to go, but choosing a development platform is a pretty
fundamental thing and I urge everyone to proceed carefully and not just
take Horst's recommendation of RoR as Bible truth. I have enormous
respect for Horst's abilities and intelligence, but technologically he
can be a bit bull-at-a-gate and over the years he does seem to have
started but not quite completed (at least not to the point where they
can be released to others) a string of software projects which he has
told us about on this list, each using slightly different technologies,
and Horst is the first to admit that he is time-poor. Thus, I would urge
several independent test drives of RoR and of its competitors in the
open source dynamic language Web app framework marketplace of ideas -
those are, I think, Django and Turbogears (both Python-based) - in a
more relaxed and languorous fashion than Horst might be able to do,
given the enormous competing demands on his time.

The main Django site is at http://www.djangoproject.com/ and has lots of
documentation and introductory material, but there is also a
introductory  screencast at
http://www.throwingbeans.org/django_screencasts.html

Django has grown out of an industrial-strength content management system
for a series of newspaper web sites, and is attracting a lot of
attention - there are several large-scale sites using Django listed on
teh home page of its web site. It's architecture is elegant and is
readily extensible. It doesn't feature a lot of spiffy AJAX stuff out of
the box, but lots can be added.

A serious alternative is TurboGears, also Python-based, mixing in teh
best-of-breed Python subcomponents. See http://www.turbogears.org/ -
there is a link to a 20 minute screencast from the front page which
demonstrates the construction of a wiki from scratch... TurboGears
includes more AJAX browser widgets out-of-the-box, using the very
well-written and rather Pythonic MochiKit Javascript library (by
contrast RoR uses the Dojo Javascript library, which although capable of
some slick client-side browser tricks and effects, is more than a bit
horrifying when you start to examine the source code for it and the
unspeakably horrible abuses of the browser document object models to
which it resorts to achieve its apparent magic). With Turbogears you can
even substitute alternative object-relational mapping layers (it
supports two very good one, SQLobejct and SQLalchemy, but others can be
used or these can be tweaked - with Django to some extent and definitely
with RoR you are completely stuck with the object/data persistence
mechanism provided) - but all support multiple back-end storage engines.

RoR is written in Ruby, which is a nice language which is a cross
between Python and Perl in many respects. However, Ruby is much younger
than Python and doesn't have such a large (and smart) community behind
it, nor an incredible depth of third party packages and add-ins for it.
It is also a fair bit slower than Python- which may or may not matter -
but Python is highly optimised and there is a multi-million Euro
EU-funded project to make Python orders of magnitude faster again
through some super-clever compiler tricks. Nor does Ruby have both
Google *and* Microsoft backing it, as Python does.

So, I am not asserting that Django or Turbogears are better than
Ruby-on-Rails, or that RoR is not the framework/language of choice for
an open primary care EMR/ER project. I am just saying that it might not
be wise to just take Horst's word for it.

Finally, for those with fast broadband connections and lots of download
overhead, this 400MB 35-minute screencast from NASA on rapid Web app
development frameworks is both amusing and instructive:
http://oodt.jpl.nasa.gov/better-web-app.mov and reinforces the view that
Horst is barking up trees in the right copse with RoR, if not exactly
the right tree.

Tim C
_______________________________________________
Gpcg_talk mailing list
[email protected]
http://ozdocit.org/cgi-bin/mailman/listinfo/gpcg_talk

Reply via email to