[wtp-proposal] May 15 Summary of Web Tools Project Proposal

["Bjorn Freeman-Benson" <bjorn@xxxxxxxxxxxx>]

(As part of the membership-at-large comment process, I will be summarizing
the discussions and feedback at the end of each week: April 30, May 7, May
14, May 21, and May 27. This is the third of those summaries. I invite you
to correct any errors I have made.)

Over the last three weeks, the discussion has become more lively and a
number of threads have emerged:
(1) Offers of contributions
(2) Discussion about project goals and scope
(3) Critiques about how it has not made clear how to contribute

(1 & 3) There have been a number of offers of contributions, both of
developers, and of code and, best of all, of both developers and code. A
number of these offers have included critiques about how difficult it is to
find information on how to contribute, both to WTP and to Eclipse projects
in general. There have been various email and news posting replies to these
critiques and to the proposals.

** In the next week, I would like to see a revision of the proposal web
pages that includes:
     (a) additional contributors and contributions, and
     (b) clearer and easier information about how to contribute to the WTP

(2a) There have been a number of excellent discussions about the project
scope, especially about what is and is not included. And about whether
something that is currently not included will ever be included. For example,
about how WTP will interact with a J2EE server:
  Patrick Mueller <Patrick_Mueller@xxxxxxxxxx>
       Eclipse is going to produce an open source J2EE Server?
  Christophe Ney <christophe.ney@xxxxxxxxxxxxx>
       [WTP will] deliver an exemplary implementation to plug [into] at
least one open source server.
and about whether "standards" like Struts are included in WTP:
  Christophe Ney <christophe.ney@xxxxxxxxxxxxx>
      Once we have tackled standards, we can always consider expanding the
scope, and until this is the case, on-the-edge innovation could find a home
in the Eclipse Technology project.
  Naci Dai <naci.dai@xxxxxxxxxxxxx>
      The question we should pose is: Should WTP become an umbrella for
additional subprojects like Struts, in addition to the fundamental
subprojects WST and JST, or should these technologies go elsewhere?
  Zack Angelo <zack@xxxxxxxxxxxx>
      Well, I certainly agree that the initial scope of the project should
be restricted to building a solid, modular framework and then from there
building a set of tools that are common to almost all Java web development
projects

The proposal has a good start to the wording of what is in and out of scope,
but based on the feedback we are seeing in this membership-at-large comment
period, I believe more clarity could be added. Perhaps a few paragraphs of
issues and tools that are in and out of scope? Additionally, it is clear
that in order for the project to make any headway, it has to start with a
restricted set of features. Thus even if something is in scope for the
entire WTP, it might be postponed until a later phase - the proposal is not
as clear as it could be on that issue.

Further, I would like to comment on the general scoping issue of web tools:
"web tools" could be a vague and all-encompassing concept. The issue here is
that web tools are at the top of the pyramid: they build upon all the
existing tools and thus where is the dividing line between what is in "web
tools" and what is in the lower layers? Consider for example, Java: because
JSP pages are considered part of web tools, one might consider Java to be
part of web tools. But clearly Eclipse has chosen Java to be a separate
project - WTP builds on top of JDT, just as JDT builds on top of the
Platform.
There will most likely be another Eclipse project in our future that builds
on top of WTP. I earlier understood that the proposal's statement "Outside
the project's scope will be support for ... such as Apache Struts" to mean
that those aspects below the horizontal line
(http://www.eclipse.org/proposals/eclipse-webtools/subprojectsac95.gif) were
outside the WTP scope entirely, but some of the conversation in the
newsgroup indicate that perhaps they are inside the WTP scope, but not part
of the early phases. Thus...

** In the next week, I would like to see a revision of the proposal web
pages that includes more discussion of what is in and out of scope for the
entire WTP, and what is in and out of scope for the first phase of WTP.

(2b) There have been questions about what the initial exemplar tool set will
support. Will it support JBoss? JOnAS? Tomcat? etc. While I had initially
asked the team to keep such details out of the proposal, I now realize that
was a mistake, and thus I ask for a revision. The goal of WTP remains to
"focus on infrastructure for tools used to build applications", but good
exemplars are essential and the proposal will be stronger if it includes
these descriptions.

** In the next week, I would like to see a revision of the proposal web
pages that includes a proposal (not a commitment) of what the initial
exemplar tools will support. This description should be based on the use
cases in the proposal, e.g., "run, test, debug web app on Tomcat".

(2c) There has also been a discussion about the WTP project principles and
what exactly WTP would be delivering, e.g.,
  Ed Burnette <ed.burnette@xxxxxxxxxxxxxxxxxxx>
      Eclipse needs to be competitive in the product space it is in
(competitive = as good or better than products
not built on Eclipse)...
      As a practical matter this means there can be no dumbing-down of
Eclipse technology.
  Stefan Matthias Aust <sma@xxxxxxxxx>
      As a potential user, I want something as useful as the JDT - which is
a state-of-the-art Java IDE
I believe the intent of WTP is to deliver a competitive (better than the
others) framework and exemplar tool set. Please correct me if there is
anything in the proposal that differs from this statement. However, WTP is
not going to deliver everything that a developer might desire. For example,
comparing WTP to JDT: JDT does not include a graphical GUI tool - other
projects and commercial plug-ins provide that. And JDT does not provide an
object-relational mapping tool - other projects and commercial plug-ins
provide that. So on and so forth. Thus there will be parts of the tool set
that any given person desires that will be outside the scope of WTP; I
anticipate that those desirable parts not included in the exemplar tools
will be a smaller set rather than a larger set, but then I don't speak for
all developers. In any regard, the exemplars will be competitive and not
dumbed-down.

(2 & 3) And lastly, a comment about contributions:
  Ed Burnette <ed.burnette@xxxxxxxxxxxxxxxxxxx>
      Would we rather they use (and contribute to) something...
  Stefan Matthias Aust <sma@xxxxxxxxx>
      I want it for free.
I admire Mr. Aust's forthrightness but I stick by my previous statement:
Eclipse, as an open source project, will prosper only if we have an active
group of contributors. There are many avenues of contribution, but simply
taking the free software and vanishing into the ether is not one of them.
Mr. Burnette has stepped up to the plate and is a contributor; I look
forward to many more like him.

Summarized by:
WTP Proposal Shepherd (Bjorn Freeman-Benson)