find an alternative concept to SF
Nick Hockings
s96121272 at tuks.co.za
Thu Nov 15 10:33:21 UTC 2001
Quoting "Georg C. F. Greve" <greve at gnu.org>:
> Or we should find an alternative concept that works out better.
It seems:
THE MAJOR COST
is Bandwidth to allow the world's developer community to access the
depository.
THE GREATEST RISK
is one organization can control the code and the most efficient means of
collaboration.
THE TECHNICAL DISADVANTAGE
Bandwidth bottle necks between the depository and the developers make downloads
very slow and unreliable in many countries/regions. (This is a problem for me
in South Africa, and worse in neighbouring countries.)
A possible solution:
There should not be a single depository in a single place. If there are many
local depositories (like one in every major city) the demands will be much less
on bandwidth per depot and on the long distance connections of the net.
Each program would have its home depository. The records of that development
team would then be mirrored by those local depots that recieved requests or
wanted to promote each development project.
The FSF (and anybody else) could then produce directories of where to find the
home and mirrors of each development project. Each depot could then use these
directories to find home depots of the projects it chose to mirror, and point
requests for downloads to the nearest/most accessible mirror.
Who would keep all these local depositories? Any body who wanted to, but my
first choices would be universities. Who in the universities? Staff/Student
Freesoftware Associations would be good. Universities allow a high degree of
freedom to their Societies and often provide significant facilities.
(Esentially a student computing society is aking to connect its server to the
university's network and website.)
What do we need to make this happen? A freesoftware depository management app
that's not too dificult to set up, and lots of student Freesoftware societies,
or other supporters with significant webspace.
(There are lots of other reasons to set up student freesoftware societies, but
this is probably the most significant contribution they could make to the
movement as a whole.)
Nick Hockings
<s96121272 at op.up.ac.za>,
<s96121272 at tuks.co.za>,
<nickh at nupedia.com>
More information about the Discussion
mailing list