[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: memberships
From: |
Marcus G. Daniels |
Subject: |
Re: memberships |
Date: |
10 Jan 2000 15:13:35 -0800 |
User-agent: |
Gnus/5.070084 (Pterodactyl Gnus v0.84) Emacs/20.4 |
>>>>> "CS" == Stumpo, Cody (C M ) <address@hidden> writes:
CS> In one of those freaks of complex
CS> dynamics we all love so much, I believe the money people here did
CS> not want to join until the future of Swarm was clearer.
The main technical goal I have for Swarm is to excise wierdness.
This is constructive in the sense of making Swarm work with popular
technology (e.g. Windows, Java, XML), and deconstructive in the sense
of removing non-portable dependencies (e.g. phase splitting based on
features specific to the GNU Objective C runtime), inefficient
implementation approaches (e.g. Tcl interpreter GUI bottlenecks), and
compressing baroque interfaces (e.g. integrating reflection features
in defobj and objectbase into defobj and objectbase/activity into
activity).
None of this is fun work, and all of it is intended to make Swarm
easier to learn and easier to maintain. Of course, it is work that
any competent software engineer could do in a stable environment, and
increasingly, I think, with less hassle. If the SDG didn't exist, and
a company like Ford had important models that depended on Swarm, Ford
could at moderate cost hire their own engineer to maintain the parts
of Swarm they cared about.
However, the SDG has several advantages:
1) The SDG is a legal entity, a 501c3 non-profit. The SDG president,
Irene Lee, has management experience in several Silicon Valley
companies.
2) The SDG is efficient. Releases and support are taken seriously,
even with our small budget. New code snapshots with pre-built, drop-in
DLLs for Windows are available almost every week, sometimes daily.
3) The SDG has organizational experience and an advisory board with
almost all all the people that have worked on the project, including
Chris Langton, Roger Burkhart, Nelson Minar, the designers.
4) The SDG has directors that are invested. Chris Langton and Glen
Ropella have a commercial modeling company that uses Swarm. I
work on Swarm much more than full time, and have a strong
desire to see through everything from visual programming to
automatic load balancing on MPI and multithreaded SMP.
5) The SDG has ties to the SFI and leading-edge complexity research.
6) The Swarm community is not that small or even that volatile. With
each release there are several hundred CD-ROM orders and thousands
of downloads. The mailing lists membership is above 600 and
growing. Given that Swarm is software for a particular kind of
research (unlike GCC, Linux, or Emacs) these are comparable or
larger numbers to other projects I know about.
7) The Swarm project has funding to continue one year with one full-
time developer, buffered and via SFI. There is no immediate
crisis that is likely to result in a disruption from which we
can't recover. Also the SDG now has independent offices from SFI.
CS> I think some key
CS> metrics would be number of users, number of serious users, number
CS> of developers, frequency of releases. These need to grow (with
CS> integrity) by a factor of some small-integer multiple.
I don't know how to quantify `serious users', but there are positive
indications. For example, on a recent visit to Santa Fe, Nelson Minar
recently remarked to me that he was at a conference and two
researchers (that he had not already known) presented work done with
Swarm. There are certainly people that use Swarm for models that
don't subscribe or participate on this list. Finally, being a free
software project, there is no restriction on how expertise or
distributed development may occur. (And I put as high or higher
priority on questions about Swarm internals as I do on basic Swarm
usage questions.)
I think these points adequately address any stability, sole source,
or industry-buy-in worries a company like Ford might have.
==================================
Swarm-Support is for discussion of the technical details of the day
to day usage of Swarm. For list administration needs (esp.
[un]subscribing), please send a message to <address@hidden>
with "help" in the body of the message.