[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[news.eclipse.technology.epf] Re: Work products modified by different tasks

Hello Jorge,

At least part of your issue is that it looks like you're using an older version of OpenUP. You should start using OpenUP 1.0, which as of last week you can download from www.eclipse.org/epf. The tasks you refer to have been modified and renamed, and there is no longer an Architecture artifact. There's an Architecture Notebook thatdescribes important decisions. The architecture itself exists as a build that is flagged when the architecture has been validated (at the end of Elaboration).

The idea is that your architecture is continually growing and being elaborated, mostly during early iterations. There is no "analyzed" or "developed" architecture. Analysis and development are going on in parallel (see the new workflow for the Architecture discipline). We didn't feel that a separate analysis task and separate architectural PoC was necessary or desired in the small projects that OpenUP is designed for. Now we Outline the Architecture in Inception just to have a place to start, and Refine the Architecture from then on. There naturally tends to be more analysis early on and more development later on, but the architecture is just the implementation of the set of important decisions in the Architecture Notebook, and it grows and matures over time.

- Jim Ruehlin

Jorge wrote:
Hello again

I came into a conceptual problem which I wanted to share with you. It is
about how to handle work products which are modified by different tasks.
This has led me to distinguish artifacts which are process dependent (they
are produced by more than one task) and artifacts which are not. Take as an
example the Architecture artifact in the OpenUP.


In principle, there are two task "Analyze Architectural Requirements"
and "Develop the Architecture" producing an architecture artifact. It is
clear that first the analyze task is performed and then the develop task.
The second task, intuitively, refines the initial architecture.


The representation used in OpenUP consists in defining, in the develop task,
the artifact both as input and output. In a sense, this is a kind of
refinement or transformation of the initial architecture artifact into
another different artifact with more content than the initial one. At the
EPF composer right now, I could not find a way to distinguish one from the
other.


The solution I applied is defining the "analysis architecture" and then
the "developed architecture". This solution is process dependent, but it is
more clear and allows defining different criteria for a "good" analysis
architecture and a good development architecture. In the OpenUP solution,
it is not possible to do so.

Is there a way right now in the EPF-composer to deal with this situation or
what I did was the only alternative? If the EPF-composer is not dealing
with this feature, do you think is worth including it in the road map?

Regards

Jorge