Permaculture Database

[Paul Cereghino <paul.cereghino@comcast.net>]

/Scott Pittman:
/

/       The idea is that every plant evolved with a collection of associates 
that
        provide for its needs ie minerals, fertility, pollination, protection,
	procreation and etc.  This collection would be classified as a guild 

/what collection? -- all those species coexisting within the wild range of a given species 
population? -- then is guild a network of nodes unique to each species, and each species is the 
center of its own universe.  Some needs are common to all plants... is this part of the guild?  
This seems like a kind of uber-clemetian uber-facilitation perspective on vegetation assembly, 
that all plants in there "edenic state" are part of a wholisitically complete 
"meta-being".  It seems to me like evolutionary dance is played out at a scale larger 
than a 100 x 100 foot assemblage... in any given patch of vegetation some populations are 
expanding while others are retreating.  However, any given 100x100 plot may work for better or 
for worse for the gardener and the yields they seek./

        would include other plants, animals, and micro flora and fauna.  Over 
time
        most plants that have been domesticated have been stripped of their
        associates and have become dependant on us to maintain them in 
relative
        health./

So perhaps the guild concept really just the axiom: "provide each 
species with a range of mutually beneficial interactions".  It starts as 
a conceptual model, and ceases to be a 'guild' when it is installed, 
whereupon it becomes reality.  And so can you really identify a "guild" 
in concrete terms?  Or are we just cataloging an infinity of 
'guild-based' experiments? 

When we say apple guild, we are talking from the perspective of the 
apple, but if we add the dandelion or the bulb, do we then begin 
designing the dandelion and the bulb guild?  I think not.. the dominant 
long-lived woody elements are the focus of the design and so take a 
position of logical dominance in the assemblage.  I believe that 
understory communities are somewhat beholden to their long-lived woody 
neighborhood.

I am starting to think that a database of guilds is simply a catalog of 
guild-based vegetation experiments based on a network of purported or 
hypothetical interactions.

What if each experimental 'guild-like' assembly is an 'observation', not 
unlike a  vegetation plot sample.  You could compare all guilds in the 
data set and look for similarness/differentness in terms of species 
composition (you'd run some calculation to identify the compositional 
'distance' between two assemblages, a multi-variate cluster analysis or 
even ordination).  You could cluster guilds that are most similar, 
clusters could be defined by their common composition.  In line with the 
idea of long-lived woody dominance, you would weight those species more 
heavily.. differences in woody composition would be more important in 
measuring differentness then would herbaceous species.

A core structure?

[species_table]
species_code

[species_guild_link_table]
guild_member_type (3=primary, 2=secondary,1=tertiary)
species_code
guild_code

[guild_table]
guild_code

[guild_observation_table]
gobservation_code
gobservation_family (strength, weakness, dynamic)
gobservation_text
guild_code

[attribute_table]
attribute_id
attribute_family
attribute_name
attribute_value (number)

[species_attribute_link_table]
species_attribute_id
species_id
attribute_id


to do the analysis you would just roll up the table

                     species_code >>>
guild_code   
\/\/\/\/\/\/               0   1  2  1  3  0  1  0  0  0 
                          1   0  1  0  0  1  3  0  1  0 << 
guild_member_type

and run it through a cluster analyis algorithmn

this way your data is clean, a huge catalog of  'guild experiments' can 
be reduced to clusters of 'guild experiments'.  Within a cluster of 
'guild experiments' some guilds would get more traffic, more information 
would accumulate... other recessive guild experiments could be mined for 
ideas or eventually deleted.

and then the environment?

[tbl_place]
place_id
lat
long
non-derivable place attributes


places could also be clustered based on similarity of attributes, and 
guilds from similar places identified.

So three core functions for a DB?:

1. assist in identifying componenets that might fit in a guild
2. record and collect information about individual guilds from places
3. allow for data simplification of large numbers of guilds/places