Re: [ba-ohs-talk] open source question
On Mon, 2002-02-18 at 10:36, Johannes Ernst wrote:
> I'm trying to understand the reasoning why some/many/most/few(?)
> people believe that a broadly adopted / adoptable OHS
> should/could/must(?) be licensed under an Open Source license, and I
> would appreciate your insights, either to the list or to me
> individually. (01)
The arguments you refer to deal with questions like "should I
open-source my project" or "should I use an open-source software
component" and they are questions designed to appeal to business
people concerned with security, stability and predictability.
In discussing this, we need to remember that "open source" was
a term invented by a group of people who were trying to make a
point about the practicality of "open source" software with business
users. This was in opposition to the "free software" people, who were
largely trying to make a political point about software as property.
Our issues are a little bit different. (02)
Ultimately the real goal of an OHS system depends on universality -- we
want to prvoke a revolution. The OHS works only if it can become a
universal infrastructure upon which any interested individual/group can
innovate. I always look back at the examples of the Internet and the
Web as models of how we want the OHS to be able to grow and inject
itself into every aspect people's lives. A strong argument can be made
that there were three fundamental components that made the Web able to
explode exponentially after 94-95: (03)
1) Very simple protocols (HTTP/HTML) -- allowed virtually anyone with
any level of technical savvy to come in and create web sites,
servers, and client libraries and applications.
2) Open source software (CERN httpd, libwww, Mosaic, Apache) -- allowed
virtually anyone with a little more technical savvy to extend the
capabilities of the systems w/o reinventing the wheel. It was
important that it was possible for students, reserachers and
hobbyists to develop ideas freely and then (possibly) take them
easily into the commercial world (this actually argues for BSD-style
open source...)
3) Free, high-quality end-user application(s) (Netscape) -- enabled a
massive, uneducated audience for all of the publishers, and became
a de-facto standard for testing and distributing extensions to the
environment. (04)
The real technical requirements are down there at the bottom:
dirt-simple open standard protocols and business-friendly open-source
infrastructure. Without those, I don't think you can hope for
universality and you probably won't seed a revolution. (05)
BTW. There was one final piece that was essential to continued growth,
but only after the others ensured that the system had grown too big, too
quickly already... Any guesses? (06)
--
-------------------------------------------------------------------------------
Lee Iverson
leei@ece.ubc.ca Dept of ECE, 2356 Main Mall
http://www.ai.sri.com/~leei/ Vancouver BC Canada V6T 1Z4
Office: (604) 822-3381 (07)