home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!olivea!spool.mu.edu!sgiblab!ultra!jerbil
- From: jerbil@ultra.com (Joseph Beckenbach {Adapter Software Release Engr})
- Newsgroups: comp.programming
- Subject: Re: collective program design
- Message-ID: <1992Dec18.214939.16716@ultra.com>
- Date: 18 Dec 92 21:49:39 GMT
- References: <fkaiser.724352429@Physik.TU-Muenchen.DE> <MAYERS.92Dec17114526@yak.syr.EDU>
- Organization: Ultra Network Technologies
- Lines: 99
-
- As always, this is one man's opinion, based on my experiences.
-
-
- >> Frank Kaiser <fkaiser@Physik.TU-Muenchen.DE>
-
-
- >>The best public domain programs known were created by many programmers
- >>working together. Everyone of these did a little share of work.
- >>But the begin and development of such a workgroup was a result of randomness.
-
- Hmm, most of the freely-redistributable and public-domain programs
- I've seen began with a few people solving a problem. Others laid additional
- code atop it, tweaked designs, and expanded the operating model along the
- same lines as the original operating model. "Random" is very much the wrong
- word to use here; I think "informal" better captures the flavor.
-
- ("Random" implies that there is no intelligence or purpose behind
- an action; "informal" implies that it wasn't necessarily done so that it
- could be done again.)
-
-
- >>One could think of rules for finding the need for a new program, for
- >>example a newsgroup with suggestions for new programs, together with
- >>a sort of democratic decision to start a new project.
- >>For the development of such a project, one could thing of a bbs, where
- >>everyone willing to take a part of the burden can find a task to solve.
-
- We already have these. :-) Someone complains about something,
- and he or someone else figures out a quick way to solve the problem. Word
- (and software) leaks out, and pretty soon others are playing with it,
- making it more portable, more robust, and more usable. Some of those changes
- make it back to the original author, and POOF the ball is rolling.
-
-
- >>This procedure should have to be higly formalized in a top-down-manner.
- >>I don't believe this is a new idea. Please comment it.
-
- Not sure what you mean by "top-down" in this context. And no, it's
- not a new idea at all: it's ad-hoc group software development. However, you
- need motivated people to work on these projects, and continuity in the form
- of documentation and (often) regular social contact.
-
-
-
-
- > Michael Ayers <mayers@yak.syr.EDU>
-
- > The democratic process thing sounds like the weakest link - "Okay,
- > we've all agreed on what to do, now who's going to do it?" [silence]
-
- Wrong question, Michael. "Okay, we've all agreed on what to do, now
- who's most interested in making sure the design stays intact? Who's keeping
- track of what's happening? And what do each of you want to start with?"
-
- A usual title for the person answering the first question is
- "principal architect" or "principal designer": he is the final word on
- this group as a whole believes should be the content of the software.
- He also is the one to divide up the work, with the help of everyone else.
-
- A usual title for the second person is "project manager": he is
- the final word on getting the resources and organization necessary for the
- project to progress. He also is the one to track that the work continues,
- and takes care of the process.
-
- Depending on the group, the principal architect or the technical
- manager makes any assignments. With smaller, ad-hoc groups where everyone
- works as peers and could do any task nearly as well as anyone else, the tasks
- go to those most interested in doing them.
-
-
-
- > I do agree that a newsgroup or mailing list for thinktanking
- > prospective projects would be a good idea, I just think that a relatively
- > anarchic framework is mandatory - project control must remain entirely
- > within the group which does the work.
-
- And someone within the group needs to handle the process, at least to
- the extent that everyone can build with everyone else's pieces.
-
-
-
- > I can see myself working on my own baby, and then giving it away. I
- > can picture myself working on someone else's project, so long as they were
- > nice to me, so they could give it away. Working for a committee? Sorry,
- > I just can't picture myself doing that for free.
-
- Ah, but working for (and with) a peer group is something else entirely.
- I infer that "committee" means a group which emphasizes process over content.
- Other climates for working within a group exist. I for one think that this
- would be a neat and logical extension of GNU and other freely-redistributable
- work.
-
-
- Joseph Beckenbach
- toolsmith and releaser
- --
- ----
- Joseph Beckenbach jerbil@ultra.com 408-922-0600 x246
- (speaking from, but not for, Ultra Network Technologies)
-