Post-OOPSLA Meeting, Seattle, November 2002
 

Shepherding/Submission Quality group – Neil Harrison

Our discussion focused mainly on the pre-conference work, namely shepherding and screening papers for acceptance.  We talked about the following:

- The quality of shepherding is inconsistent: some shepherds are really good, but the worst ones do little or nothing.

- Many of the most experienced pattern people have drifted away from being shepherds.  One reason may be the next point.

- Some shepherds (Bobby brought this up) don't get anything out of being a shepherds.  Of course, this is a big deal, since we don't pay any money to shepherds; they have to get non-monetary compensation (i.e. they have to feel good about it.)  Part of the problem may be that the authors don't respond well; they don't make much progress. Part may be that the papers themselves aren't in the shepherd's area of expertise, and thus the shepherd can't do much better than comments on form.

Both these may be caused in part by the following situation, which happens at every PLoP:  The first batch of shepherds picks papers to shepherd, leaving a few "undesirable" papers.  The program chair has to scramble and twist arms to get shepherds for the final papers.

There is a reason those papers are left to last. They may be particularly bad, or may be on a topic no one else knows anything about. Or they might be papers _about_ patterns, by people who really don't know anything about PLoP conferences.  This happened with two papers at VikingPLoP, and shepherding was a bad experience in both cases. (Well, one author never responded at all, so the shepherding was easy...)

Things to do to help:

1. Keep building up the shepherd pool. Besides the usual recruitment and shepherding class at PLoPs, send names of candidates to Neil.

2. Program committees need to take a more active role. First, they should pre-screen papers.  In particular, papers that are not pattern papers should either be rejected out of hand, or handled differently in any case. They shouldn't be shepherded.  Second, PC members should be more active in overseeing shepherding. Third, they may need to be more thorough in accepting/rejecting papers after shepherding.

3. It must be made clear to authors what is expected of them, i.e. if they don't respond to their shepherd, their papers will be rejected. There should be a standard letter outlining expectations from the program chair (different from the letter of introduction from the shepherd.)

4. We probably need to make it more clear what is expected of shepherds.

Software Time Capsule - Dragos Manolescu

    Library of Congress Digitizing program
    Storage formats
    Recording folk history
    Folk history event
    Collaborate with historians for ideas, of what, how, why
    Find ways to obtain resources ($, space, volunteers)
    How to deal with IP, etc.
    How to organize what is important
    Example
    Manifesto
    Criteria
    What related artifacts do you keep for s/w?
    Artifacts associated with the individuals

Patterns Conferences – Doug Lea

I suggested that Hillside sponsor an academic-style refereed conference on patterns, colocated with OOPSLA and/or ECOOP, and with a submission rule that the patterns must have already been workshopped at a PLoP.  The main motivation is to provide a "certification process" and publication outlet for the best patterns coming through the PLoPs.  Most people seemed to like the idea, but there were enough snags that we ultimately didn't take any action on it. Among the problems are that ACM is unlikely to publish the proceedings because of SIGPLAN rules.  And that colocating with ECOOP probably wouldn't work because it is a single-tracked conference. And enough decisions have already been made about OOPSLA 2003 that we don't think it could be done then.

We later joined in with the Publications discussion group. The consensus seems to be that some publication venue would be better than a conference. One possibility that got weakly positive response was to publish a reviewed pattern per issue of the online Journal of Object Technology (http://www.jot.fm/).

Cool Conferences – Dirk Riehle

A cool conference or event or whatever is something that immediately makes you want to go there.

We identified four main dimensions why a conference might be cool:

- cool (or hot) topic
- cool people
- great location
- cool style

Cool topics

- an event with "best-of" tutorials
- a workshop on "getting real" that is admitting and talking about fear (Martine)
- interdisc. (for the uninitiated: something interdisciplinary)
- a metaphor manifesto
- a three-legged event: complexity science, business, software (Dirk)
- a mixer: techies meet artists
- a future of patterns retreat

Cool locations

- something floating, for example, canoeing in Colorado to your next day location
- something in the Alps, in the summer, or for skiing in the winter
- hook up with international organization that knows cool locations
- standard locations: Allerton, Irsee, etc.
- another suggested location: www.wolfsberg.com

Cool event-styles

- summer camp-style, vacation-time University---combine with vacation
- open-space, that is, nothing predetermined, just free time if so desired

Pattern Community and Conference of Japan – Mimpei Morishita

We—Mikio Aoyama, Takako 'Tina' Nakatani, Hironori Washizaki, Terunobu 'Terry' Fujino, and Mimpei Morishita—discussed issues of Japanese pattern communities and pattern conferences of Japan at the post-OOPSLA meeting.

A problem of Japan is there is no appropriate local forum to discuss software patterns.  JapanPLoP -- is a Japanese voluntary association modeled on the Hillside group and a base of MensorePLoP -- had played such role for several years, but it has stopped its activity since a half year ago. The trigger was the failure of MensorePLoP 2002.  MensorePLoP 2002 was cancelled because of some problems:

-          Financial problem

       A company underwriter of MensorePLoP 2001 couldn't continue their financial support for MensorePLoP;

  - Copyright problem

    Authors of MensorePLoP 2001 papers was confused as copyright issues wasn't clear;

  - Language problem

    Conference organizers planed MensorePLoP 2002 as an international conference using English, but actually, there were many demands for using Japanese.

We have a plan of organizing another software pattern community as a working group under SIGSE(SIG of Software Engineering) of IPSJ(Information Processing Society of Japan).  It will be open to everyone who wants to discuss software patterns. It seems that the working group has some merits; we expect both industry and academia join it and work together; the IPSJ brand will be helpful for calling participation and financial sponsors.

We hope that a successor to MensorePLoP will be held based upon the working group activities.

We need not only new patterns but also more studies on existing patterns, then the conference will be writers' workshop + alpha.

We discussed the evolution strategy as follows:

(1) Conference for Japanese engineers

   - One software pattern track for the Symposium on Object-Orientation(counterpart of OOPSLA)

   - Workshop in Japan

(2) Conference for Asia / Pacific

   - In conjunction with APSEC(Asia-Pacific Software Engineering Conference) annually held in December around Asia-Pacific

Publications – James Noble

We discussed the vacuum left by the demise of the PLOPD book series as a venue for material that has been workshopped but is now worthy of a wider audience. This disappearance has caused a range of forces to manifest themselves:

 * pattern writers (and shepherds) can't find material relevant to patterns they are writing or shepherding.

 * some patterns are better than others, but looking through a PLOP proceedings, there is no way to tell which is which

 * programmers do not get the benefit of expertise described as patterns

 * material from conferences can be lost forever, making future historical work impossible.

 * authors do not get public recognition for their work (unless you are slashdotted)

 * authors (in universities or research companies) do not get academic recognition for their work

 * the patterns community becomes more fragmented

The group considered that a pattern collection book series, or a patterns journal could address many of these forces.  In the current economic climate, however, it was considered unlikely that a commercial publisher would support or adequately resource such a journal. For example, most pattern authors would be unwilling to pay page charges for their papers to appear.  The core of the patterns community is too small to make this a viable proposition, and the Hillside group does not (currently) have the financial resources to support this kind of venture itself.

On the other hand, a number of smaller efforts are emerging that address these forces individually. For examples, the PLoPs are establishing their own individual archives to maintain their own material. By publishing proceedings with ISBNs or ISSNs, PLoPs can ensure some level of academic credit for their work.

The group also identified some other small proposals, after merging with the Patterns Conference discussion group.  In particular, the patterns home page could present a "pattern of the month", and over time build a list of widely accepted patterns. Hillside could negotiate with existing journals to publish one peer-reviewed, typeset pattern per issue, or with existing conferences to provide an academic-style refereed patterns stream. 

Summary Business Meeting Post-OOPSLA, November 9, 2002, 1h – Dirk Riehle

Part I: President's Note (Richard Gabriel)

We finished the new mission statement that states Hillside's direction.

The mission of the Hillside Group is to improve the quality of life of everyone who uses, builds, and encounters software systems-users, developers, managers, owners, educators, students, and society as a whole.

Developing software is one of the most difficult human activities, and it affects every aspect of modern life. Understanding and helping the human element is critical for achieving success. The Hillside Group believes that software systems and software development can be made more humane by paying attention to real people and existing practices.

The Hillside Group promotes the use of patterns and pattern languages to record, analyze, and improve software and its development, and supports any new practices that help achieve its mission.

The Hillside Group sponsors a variety of activities to achieve this mission-organizing workshops, conferences, and publications for discussing, recording, and documenting successful software practices.

The mission statement was confirmed by the attending community members.

Part II: Membership Rules (James Noble)

Membership Rules

• Hillside keeps a list of members, which includes each member's name, address, and email address.

• Candidates must be nominated for membership by a Hillside member.

• Candidates are admitted to membership by the board—this may be delegated to a designated membership secretary.

• Members may resign, or choose to become emeritus; resigned or emeritus members can rejoin at any time.

• Benefits of membership:

  • you are on the members' mailing list

  •  you are invited to the Hillside Annual Meeting (post-OOPSLA every year)

• It is expected that new board members will have been Hillside members for at least 1 year before they join the board.

The rules were confirmed by the attending community members.   Joe Yoder will be the membership secretary.

Part III: Financial Report for 2001 (Dirk Riehle)

The financials for 2001 are in and it doesn't look rosy: we are down from an all-time high to about $40K in the bank. 2002 promises to have stopped the losses, but we will need come up with future opportunities.

For a more detailed discussion, please see the appended files:

- Hillside Financial Report 2001.doc
- financials-2001-2002-preview.doc

The report was confirmed by the attending community members.