[OpenAjaxIDE] Evaluation criteria
jferrai at us.ibm.com
Thu Oct 4 08:53:28 PDT 2007
Hello IDE WG,
Bertrand and I had an email discussion in August about evaluation criteria.
Here are the two emails:
I did what I could but I’m not sure what we’re trying to achieve here.
We have a list of requirements which encapsulate where we want to go, and
we have a number of existing formats, none of which is the one we want
exactly. We know it’s going to take work to get where we want to go. The
different requirements are not weighed except for the MUST/SHOULD/MAY, but
some MUSTs are much more important than others. There is a number of other
things the list doesn’t capture:
· Is the format already being used and tested against
real-world, large-scale applications and how well does it do?
· Is the scope just Ajax applications or is it broader (other
languages and environments)?
· How many of the requirements could be satisfied through a
simple extensibility feature? Should the most Ajax-specific features
be simple extensions?
· How flexible is the format? I mean by that flexible as a
format but also is each owner completely open to modifications?
What we need to find is which format will be the easiest to steer into our
goals, which is not necessarily the same as asking which format satisfies
the most requirements today.
My thinking is that the comparison table is informational and educational
but not deterministic about which technical direction we would go. The
table will help us understand the metadata formats that people are using
and it will help us see if we left out any key requirements.
In terms of how we might proceed, we could:
(a) Use one of the existing industry metadata formats as a starting point,
from which we (undoubtedly) make changes
(b) Invent something new from scratch
Option (a) seems more attractive to me because we probably can move more
quickly with (a) both in terms of completing a specification and in
achieving at least one implementation.
I recommend that we have discussions soon about the criteria we use in
evaluating existing metadata formats for suitability. You list four
criteria below. I would add three others:
* Good workflow with Ajax libraries that use JSDoc - There are a good
number of Ajax libraries, particularly some of the most popular ones, that
use JSDoc to document their APIs. The whole point of our initiative is to
allow IDEs to do good things with these Ajax libraries, so compatibility
with what is used today is important.
* Open source reference implementation(s) - At OpenAjax Alliance, if at all
possible, at the same time we deliver a specification, we also make sure
that there is open source available so that (most of) the world can work
with things immediately. For example, with the OpenAjax Hub, we have both a
specification and our own open source.
* Can be supported reasonably by the IDEs that have been represented in
Will you be available for the phone call this week or are you still on
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the IDE