David Holland dholland at hcs.harvard.edu
Thu May 18 00:35:13 EDT 2000

 > >> My suggestion to the membership - pick one thing, and do it well.  Start
 > >> up _a_ project in the HCS, and be the force behind it, and get other
 > >> people involved, and don't back down.  If you think it's cool, and you
 > >> keep at it, so will other people.  But you have to get the momentum 
 > >> going.
 > >
 > >I think this is the best advice we've seen all evening ;-) A major
 > >"problem" with the HCS, as Rick has pointed out, is that it tends to
 > >get drawn in too many directions and, as a result, projects tend to
 > >peter out.  It also means that people with limited time to devote to
 > >the HCS tend to shy away as there aren't any organized, ongoing
 > >projects that they can work on and still not be sucked into a major
 > >commitment.
 > Would people prefer to have say only one project opportunity at any given
 > time, such that those that are interested get really involved and those
 > that don't just wait until the next one rolls around?

I doubt it. Different things interest different people.

I have two comments to make (having been around even longer than Rick,
scary as that is):

first: it's tempting to try to organize a project and get other people
to do all the work. or even a lot of it. This doesn't work. You have
to be prepared to do much of the initial work yourself or with one or
two other people. Once things are rolling and you have some momentum,
then you can get more people to help. Sometimes. Sometimes, the four
people who show up for the second meeting _are_ the project - but
that's fine.

second: historically, the HCS has been very bad at programming
projects. There have only been a handful that succeeded in any form,
and, as far as I can remember, nearly all of those only did so because
there was some kind of external deadline or agreement and got done at
nearly the last minute by one or two people. But there have been
plenty of failures. There are lots of reasons for this, but ultimately
it's the same problem as starting an open-source project: you can't
design by committee and you can't delegate work to volunteers who have
no real commitment.

The best kind of HCS project is the kind that has a few people doing
the planning and setup and then more people doing a little bit each
later on, stuff that doesn't require legwork or much planning or
commitment - maybe an afternoon in the office together or the like.
acctserv has sort of worked this way in the past, after a fashion; so
has manager. So have other things as well. Programming projects can't
be done this way, which is I think part of why they tend to fail.

Anyhow, I don't actually have anything substantive to say, so I'll
leave it at that...

   - David A. Holland
     dholland at hcs.harvard.edu

More information about the hcs-discuss mailing list