pcplantdb

[Chad Knepp <pyg@galatea.org>]

Hi there ibiblio folks,

I'm part of the group of people that is working on a yet unnamed
permaculture database currently organized around the
pcplantdb@ibiblio.org mailing list.  We truly appreciate your generous
offer to host our project and we would be honored to be a part of the
ibiblio collection.  The complication is that we would like to use
Zope products to develop this database.  Lawrence London has informed
us that Zope is currently not available on ibiblio but that you would
consider installing it if we supplied you with more information on its
requirements.

After some discussion, we are still fairly uncertain as to real world
metrics but we are willing to hazard some guesses based on pfaf.org
and galatea.org/eden.  

Disk space required for additional and updated software packages
(namely Zope 2.7.2 and python >= 2.3.3) should take less than 100
Megs.  Our current dataset in MySQL is around 45 Megs but we are not
sure if it will be larger or smaller in Zopes object database.  We
have also acquired an image set which will probably increase this
significantly.  My best guess in terms of disk usage to start with is
200 Megs.  

We hope to grow our dataset by soliciting user contributions but this
will probably not significantly impact growth during the development
stages (6-9 months).  It is also uncertain as to what level of
adoption we will have by the permaculture community.  If our project
succeeds as we hope/expect it to, our dataset could grow into the
Gigabyte range esp. considering the inclusion of photos and other
digital media with its larger size.  This should be several years in
the future though.

Bandwidth should also start relatively low perhaps 1-5 Gigs/month and
possibly growing 10-20 Gigs/month in full use.  This is based on PFAFs
approximately 10 Gigs/month throughput.

CPU usage is also hard to estimate and although Zope is not exactly
know for its performance we are confident that we can maximize its
potential.  If your server supports logical partitioning it might be a
good idea to apply constraints to CPU usage.

Our team happens to include two individuals with Unix sysadmin
experience.  If you would be interested, we could begin development by
doing a local install of the tools we need and run Zope on a high
port.  This would allow us to begin development on ibiblio early on as
well as get a better understanding/estimate of the sort of system
resources necessary.  Later, if desired, Zope could be
installed/integrated globally or simply started as root during init.
Like any good server Zope will setuid() to 'nobody' or selected uid
after binding the low port.

The development of our project is currently being funded by a grant
from Threshold Foundation as well as a significant amount of volunteer
effort.  We would like to continue development beyond the parameters
specified in the current grant and if applicable, future solicitations
will include requests for support of our ongoing hosting costs.  This
way we could, hopefully, help offset some of the costs we incur.

Thanks for your time,
Chad

-- 
Chad Knepp
python -c 'import base64;print base64.decodestring("cHlnQGdhbGF0ZWEub3Jn")'