pcplantdb

[Chad Knepp <pyg@galatea.org>]

Lawrence F. London, Jr. writes:
 > Chad Knepp wrote:
 > 
 > > The learning curve of Zope is two part.  To write highly customized
 > > products for it (what we need), Zope is pretty complex.  If we were
 > > just setting up a wiki in Zope it wouldn't be much harder than
 > > installing any other kind of wiki.  The use and maintenance of these
 > > products is really quite simple and there is lot of customization that
 > > can be done without getting much more technical than writing HTML and
 > > some Javascript.  This part of Zope also has excellent documentation
 > > <http://zope.org/Documentation/Books/ZopeBook/>
 > 
 > Does this mean that he only wiki you can run in a Zope environment
 > is one written in Zope, i.e. Zwiki or some other. Or does it allow
 > wiki apps that need a mysql database and php to access those
 > libraries on some other server, i.e. on some other ibiblio machine
 > where those apps are available globally?

No, Zope plays nicely with just about anything.  We could totally tap
into the MySQL store of your PHPWiki with Zope and either copy the
info into HG or wrap it in wrapper objects and serve it up direct.
The only challenge would be translating the data schema PHPWiki uses
in MySQL into the HG object model.

 > I assume that in a Zope environment you can only run apps written
 > in Zope therefore you are limited to what's currently available.

Although this is true, consider that there are no apps available for a
platform we write from scratch.

 > I am still skeptical about what can be done with Zope to produce
 > the software we will use for the long term.  With Rich's comments
 > in mind, I wonder what limitations will be imposed on users by
 > accessing all our resources (data, graphics, etc) through an
 > all-Zope platform instead of the usual Apache, Linux, MySQL, PHP,
 > HTML, XML, Javascript, CGI scripts, Perl, PHP, etc.  From the
 > perspective of low-level administrators of the HG platform, only
 > being able to work with tools available in the all Zope environment
 > may (may (?), not will) limit their ability to create the type of
 > resources they want to see exist in HG. 

I don't think that Rich has really explored Zope to the extent that he
can make an objective criticism.  Zope can integrate with every one of
'Apache, MySQL, PHP, HTML, XML, Javascript, Perl, PHP', it runs on
Linux and fills the role of a CGI script.  I don't think that the
tools provided by the Zope platform are limiting in anyway.

 > With the Apache/Linux/MySQL, etc. option all you need are the usual
 > web authoring skills to help maintain the HG website, containing
 > the database which would be maintained by those with the required
 > skill levels.

Not really.  I've thought of this in relation to Eden.  Right now the
only way anyone can make cosmetic changes to the program are to go in
and edit the client portions in python that create the HTML.  This is
definitely a higher skill level task.  In Zope, I imagine that the
appearance of the site would mostly be controlled by ZPTs, by either
editing through ZMI or editing in an XML/HTML editor and ftping.
<http://zope.org/Documentation/Books/ZopeBook/2_6Edition/ZPT.stx/>

 > Look at it this way: you, Rich and others are going to be paid for
 > any/all the work on HG you do. It is very likely that a
 > considerable amount of funding will become available for this
 > project and fairly soon at that. So, as long as 1) you and Rich
 > agree on the development platform to be used 2) the best platform
 > is selected based solely on technical criteria (long term
 > stability, robustness and maximum functionality, portability and
 > expandibility, does the most for users of the database) what
 > difference does it make which platform you choose?

I'm not super confident that unlimited funding will shower us from now
on.  I was totally surprised that we got the funding we did.  In my
head this may be all we get ever and I want to make the most of it.  I
also think we can generate something with this funding that will turn
heads and generate more funding.  I think that using Zope adds a lot
to the possibility of creating something meaningful with this money.

Even if we got what we asked for I still don't see any reason to
rewrite everything from scratch just because we could afford to.  Zope
is my answer to 2) for sure.  Still waiting to hear from Rich.

 > I still must stick to my preference that we do this project on
 > ibiblio and not at some location where we have to pay for what
 > ibiblio offers for free. If we have to provide systems
 > administrator resources, say on a co-located server, we will have
 > long term maintenance issues to deal with as well as paying for
 > bandwidth and equipment and its maintenance. ISP's charge serious
 > money for hosting and HG could get stuck with big costs as it
 > grows. This could be a problem for HG offering its resources for
 > free to the public. Stephanie's statement that our grant (future
 > ones too?) allows only a small percentage to be used for
 > wages. This needs to be changed early on or HG will go
 > nowhere. From the beginning I have felt that most (maybe 90%) of
 > funding shoud be used to create the software product. That could
 > change after the software is created and more money is needed to
 > pay people to maintain and upgrade that software and to buy
 > information resources such as datasets and graphics or
 > multimedia. With ibiblio offering us the most expensive component
 > (over the long term) of our project for free we can get by with
 > less grant funding (good in case grants become hard to get) to make
 > HG happen since nearly all that money can pay for software
 > development and maintenance. After all, we go nowhere without that
 > software.

I agree that a large part of the cost of this project will be in
development.  This will gradually diminish as the project matures.  I
don't think hosting is the most expensive component at all.  I
currently pay about $10/month for hosting of galatea.org and a place
like johncompanies would cost us $45/month for the kind of power we
need (4 gigs disk, $2/month/gig for more, and 40 gigs/month
bandwidth).  Add $15 year for domain name and are actual hosting costs
are $555/year which is currently 5% of our funding.  In the worst case
scenario I'm pretty sure we could cover these costs indefinitely with
a PayPal donation button on the front page.

 > Whether Zope or Eden/MySQL, etc. is chosen I think we should use
 > ibiblio as our host. To me, it just doesn't make sense to go
 > elsewhere, for all the reasons I have already mentioned. I have
 > serious concerns about the future of HG if it doesn't exist on
 > ibiblio for as long as they offer us space and resources. JS
 > mentioned that there are lots of free servers out there. Name a few
 > and tell us exactly what they offer for free and for how long and
 > the limitations are that they impose. I don't really see this as an
 > alternative worth considering.

I agree that we need to work out a long term solution for hosting
permaculture.info.  $555/year is a non-trivial amount of money and
having ibiblio cover our hosting costs would be an ideal solution.  I
don't know anything about free servers other than sourceforge and
savannah and they just provide file storage space, not web app
support.

 > So, at this point, my decision is to continue to participate in the
 > development of permaculture.info as long as it will be hosted on
 > ibiblio, regardless of the software used. If the group decides to
 > use other hosting I will make my contributions elsewhere, probably
 > on my own websites, lists and wikis. If Rich prefers to remain on
 > ibiblio and develop PFAF further there I will contribute data and
 > photographs to PFAF (even if it is only through their existing
 > tools used for submitting resources).

I hear that hosting on ibiblio is very important to you.  I think
everyone in the group holds it as the ideal solution.  I can
understand that you feel cautious about supporting the project when we
may be moving in a direction around hosting that feels dangerous to
you.  I understand what it means to give up your own project and
pursue another similar one.  Having concerns about the new project
doesn't make it easy to give up on the sure thing that is safely in
your control.  I don't share your precise concern that hosting is or
will be a big problem and I don't really know how to express what I
feel in a way that you could relate to.

To try and summarize, I hear your position about this decision is that
if hosting is not solely on ibiblio you will not fully participate in
this project.  Is this correct?  Also, can think of anything that we
can do to reassure you about the hosting issue if it seemed desirable
to the rest of the group to use hosting other than ibiblio in the
short term?  I suggested 'give it a year' previously and I suspect
that this was not really enough.  Any ideas?

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