WICSA 2005 Talk:BoFs

From WICSA Conference Wiki

Bredemeyer provides a nice framework for identifying the aspects of what an architect knows, what an architect does, and who an architect is (see [1]). It would be a good basis for discussion. Also, this discussion is indirectly related to the Education Working Session. Jeff Tyree

---

Results from BOF - A FORUM FOR DISCUSSING THE KEY CHALLENGES FACED BY PRACTICING ARCHITECTS

Attendees

1. Jeff Tyree
2. Art Akerman
3. Andrew Mellinger
4. Eoin Woods
5. John Klein
6. Kong Posh Bhat
7. Chris Corbell
8. Eltjo Poort
9. Paul Clements
10. Teemu Makela
11. Vikram Jamawal

The BOF attendees would like to make the following requests and offers to the WICSA Conference Organizers:

Requests [more TBD]

• Keep a strong distinction between research papers and practitioner papers.
• Solicit case study papers explaining (in detail) architectures for challenging systems
• Do a better job getting more practitioners to attend. Advertise at conferences such as SD East/West. Additionally, target specific companies recruiting people for working sessions. There are many architects that would probably jump at the chance to collaborate informally with other architects.
• Consider working with IASA and/or WWISA to attract practitioner participants.

Offers [more TBD]

• A page has been started for Practitioners interested in participating in case-studies and similar research
• Review practioner submissions for WICSA6 (John Klein, Eoin Woods)
• Organize workshop at WICSA6 to bring together Masters and PhD students and practitioners to discuss the realities of practicing architecture, with the goal of focusing new research. (John Klein)

Discussion Summary:

Observations about the conference and the bridge (or lack thereof) between academics and practitioners:

• From those that have attended previous WICSA Conferences, the conference has made positive progress toward providing papers and sessions that are more approachable, from a pacticing architecture perspective.
• That said, a significant majority still feel that there is substantial progress to be made to make the conference more relevant to those that use architecture and the associated techniques in the field.
• The concepts and techniques as presented will be difficult to take "back to the office" and directly applied. Often considerable engineering efforts are required to fit the ideas into practice. In addition, the concepts and techniques need to be "sold" to the organization -- a task which would much benefit on quantified evidence on the expected gains for the approach.
• Presentations should all start with "What problem are we trying to solve". Someone mentioned that what is needed are the stakeholders and the key concerns to presented before going into the presentation.
• The concept of tutorials made need to be reworked. They may be better replaced with brown-bags. They can be somewhat less formal, with a different set of expectations. OTOH, some of the practitioners coming to WICSA are really looking toward the quality tutorials instead of paper sessions -- they help draw 'normal' practitioners from companies around the conference site to participate.
• Perhaps the SEI should make its architecture courses available in conjunction with WICSA.
• There was not a clear separation between research track papers and pratitioner track papers. It seems that once a submission happened, the distinction vanished. Practitioner papers should be reviewed by practitioners -- if a practicing architect sees research-based comments on his paper, he won't be likely to submit again.
• Efforts should be made to contact leading architects directly and get them to attend.

Observations about the Practice:

• Although the majority of field architecture is "soft-skill" related, not technical, the academic researh is predominantly focused on the technical aspects. That is, acadamics are focused on the "harder science" and have not focused on the social science aspect. As a note, much of an architect's focus is on managerial and political issues.
• It was noted that ~1% of those practicing architecture in the field are aware of the concepts and techniques covered at the conference.
• There is some disagreement among the group as to who needs to be trained on architcture concepts. Is it (1) new, young developes; (2) older, and perhaps stoggy, developers who have been in the same application space for 20 (or more) years; or (3) managers.
• Architecture, as practiced, doesn't lend itself well to algorithmic-based processes. Many of the techniques presented seem to require more rigor than can be expected (and tolerated) from field architects.

Individual observations and contributions (made after the BOF session)

Clements:

• After a research presentation, perhaps we could ask practitioners in the audience to judge it for applicability, usability, understandability, practicality, etc. We could have a standing panel of practitioner reviewers in each session. Research presenters will know of the panel beforehand.
• After a practitioner presentation, perhaps we could ask an in-place researcher panel to comment on the presentation for use of available techniques.
• We could have a standing session at each WICSA: "What practitioners need from research and what resarchers need from practitioners"
• WICSA may no longer be able to bridge the gap between practitioners and researchers merely by "stacking" the organizing and program committees with people from both, although this is vital. We may need to make this bridging happen actively (not passively) in the paper and working sessions.

Krikhaar:

• I added an idea in the closing sessions (WIKI): Let's organize a working session where architects from industry draw their software architecture: no preparation, free format, etc. We can build up the database Grady was talking about. (see WIKI closing session). It will become a real working session when researchers related the probably more ad-hoc method in terms of state-of-the-art. In this way practitioners can learn from researchers and vice versa. It will also fill a gap between the in industry and academia. It will need very good moderators, and it will take time, but anyhow worth to work put I think.

EoinWoods:

• I'd suggest that papers are organised into three groups:
  • early research (mainly of interest to researchers);
  • transferrable research (research that is now ready for technology transfer and practitioner feedback); and
  • practice (practitioner papers that will, inevitably, be written differently to researcher papers).
• Case studies from many companies (such as my employer) will be difficult due to standard confidentiality constraints. Still, they're very valuable and should be solicited if possible, even perhaps if this involves asking some of the vendors (although safeguards need to be placed to avoid the flawless sales-pitch story).