Skip to Content.
Sympa Menu

pcdb - Re: [pcdb] Modelling RDF in relational databases

pcdb@lists.ibiblio.org

Subject: Permaculture Database

List archive

Chronological Thread  
  • From: jedd <jedd@progsoc.org>
  • To: Permaculture Database <pcdb@lists.ibiblio.org>
  • Subject: Re: [pcdb] Modelling RDF in relational databases
  • Date: Mon, 22 Feb 2010 23:10:50 +0000

Matt,

Very glad to see you're not one for brevity.

I think you'll fit right in.

On Monday 22 February 2010 20:44:46 matt wrote:
>First question - are there any resources other than the archive that
>I should be aware of? I saw mention of a wiki, but also that it had
>been overrun by spam. Would it be possible for people more up to

I don't think there's any sites. Mail archives aren't much fun to
read unless you can pull them into a local reader - but happily
(regrettably?) there aren't a huge number of messages in here,
and in-message noise aside, the s:n ratio is pretty high.


>1) Highly searchable database of organisms with extendable
>attributes for each entry, like PFAF but improved for permie use.

Correct. Though rather than needing to be re-deployed for
Malaysia, say, it'd contain enough information about location (from
each of the contributors) to give context for all the observations
about different organisms.

This raises interesting questions about zones (USDA, etc), and
I'm particularly sensitive to this subject given I'm from AU.

For example - each person in the DB records their lat/long/alt
and some (though I haven't worked out how) climate information.

Consider this page - that summarises the local Meteorological
bureau's recordings for a town near my place. It'd be fantastic
to have some consistent way of recording this *type* of data,
so that other people could easily identify like-climated people,
and consequently the correlated and presumably most-useful
observations.

http://www.bom.gov.au/climate/averages/tables/cw_061086.shtml

But I digress.




>discussing the wiki he had set up for his land. Jedd - does it
>still exist? Can I see it? This kind of tool would be for

I have some very basic stuff at:

http://dingogully.com.au/trac/pdb

git repository is .. well .. hmm. I got excited a while back, and
then got overwhelmed, and then did some work on a forum, and then
got overwhelmed again, and then got a job. Sadly just before the
last step I got very excited about jquery, but sadly I'm pretty woeful
with things js, so I lost interest very quickly, so if you do grab
the git repo (and really, there's no reason to - there's nothing of
substance in the thing) you'd probably want to go back about 10
commits to the pre-Mess stage.

Having said that, I have some schema stuff in there - a collection
of .mysql files - that gives some insight into what I was thinking
of doing, though it's *way* far from being interesting insofar as
addressing the really complex problems. And for that matter,
you can browse that via the trac browse-source web interface.

Brutally honestly - I'm not likely to get much time to get back onto
this stuff in the next several months. The gig I've taken on is
a bit heavy. On the upside, I might get sacked or resign in
frustration in a few months ..



> a) A suite of simple tools that accomplish basic tasks but
>can be combined as needed by the user/gardener. Think
> handheld - like an iphone app tied back to a
>database somewhere.

One of my design goals is the whole thing will be portable - anyone
can download the entire set of data - I don't want to be precious
about this stuff at all. I rate the problem of feeding stuff back
into the primary db quite low - it's more facilitating the running
of a complete system on your laptop kind of thing.



> b) Create a very flexible system that is much like (1) where
>you enter just your projects and organisms, and relationships
> you notice between them. Make this very searchable
>so that you can keep track of what's going on and more
> importantly pick up on trends over the past years as
>your experiments/projects progress. This is basically a data
> management tool where the skill of the user in
>querying the history will result in pattern discovery. The pattern
> discovery itself is too difficult to automate
>(because life is beautiful) but the skilled gardener would pick up
>on things immediately given the right interface.

> c) Extend (b) so that each practitioners personal findings
>can be automatically cross referenced with other's information.
> This is a network database model, with a wealth of
>information available but reasonably parse-able.

This very much sum up what I want to do, too.



>3) Full scale recording and tracking of activity/productivity within
>an environment.

I understand your concerns and warnings on this front.

However, for something like this:

> 3) Searching for fundamental processes that push an organic
>system towards success as opposed to failure

... I have no idea how you model the stuff that you'd need to
model in order to try to identify these processes, if you see
what I mean.



> 1) Using energetic analysis to scientifically show the huge
>efficiency of natural and permaculture systems.

- would imply we have comparable data for non-Pc systems
(which I don't think we'd be likely to get..?) Would we want to
have a system that let us record typically non-Pc stuff - broadscale
cropping, inorganic pesticide use, etc?

Not that I don't think this is is a laudable goal - just that I think
there are better ways of making this point - controlled
experiments, f.e. - and I tend to think of (my view of) the
system as being one that helps Pc types track Pc type
activity. Yeah, I realise this means all my biases and prejudices
(BD500 anyone?) are going to get in the way.



>And other things too. The first step in something like this might
>be just to set up a better way than a wiki to record all sorts of
>data and relationships without an endgame in mind. Once the

I was really resistant to a wiki before, but .. the idea is growing
on me. I think early on we were talking about putting some of
the actual data into the wiki, rather than treating it as a meta or
design tool only.

Now we can have a brief argument about which wiki to use ;)


>An a nibble for the tech heads: You are on the wrong track thinking
>about this in RDBMS systems. What you want for this kind of thing
>is full scale OODBMS - yeah that's right, Object Oriented Database.

OODB's .. yeah .. postgres is one such beast, or even some ORM
over the top of an RDBMS - there's several that get thrown around
on the CodeIgniter web site where I do a fair bit of lurking.

But ... (!) ... I'm finding it hard to get excited about ORM's, or
even native OODBMS. I tried to get into db4o, for example, but
just couldn't summon up the requisite level of interest. ;|

I know that the bulk of this mindset is stemming from the 'you like
what you know, and that's usually what you played with first' thing.

Are you suggesting that what we're talking about *can't* be done
with an RDBMS, or merely that it can't be done *easily*?

j.





Archive powered by MHonArc 2.6.24.

Top of Page