The algorithm of plants: After the fire
On Lindenmayer systems, architectural models, and a licensed picture that still has to carry the record.
A rewriting system
In 1968 Aristid Lindenmayer described a plant as a string that rewrites itself. Each symbol is a module: an internode, a leaf, a bud. Each rule replaces that module with a short string of new ones. The rules run together, not one branch after another. Prusinkiewicz and Lindenmayer later showed the consequence. A few productions, applied again at every scale, make a crown that looks branched because it is the same instruction, smaller. That self-similarity is what gets called a fractal. It is not an ornament added to the tree. It is the growth.
The Archive does not grow every plant as a classical L-system. Grasses use a culm generator of that family. Trees use the architectural models of Hallé, Oldeman and Tomlinson (1978): twenty-three recorded ways a trunk and its branches may repeat, each with a branching pattern, a canopy type, a trunk axis, an orientation, and a growth rhythm. Radius follows a pipe model, in the line from Leonardo’s observation to the West, Brown and Enquist relation, so a limb is as thick as the limbs it must feed. Branching is nodal. A child leaves the parent along the internode, on a spiral, rather than only at the tip. Three solvers sit under that: a recursive skeleton, space colonisation toward a crown of points, and a self-organising pass in which vigour and shade compete. The L-system is the formal ancestor. The working method is an architectural model plus a measured rule for thickness and attachment.
What the early record actually is
The first stills are the rule with almost nothing else on it.

A leafless scaffold is then set beside a tree that exists.

The comparison is numerical, and it is allowed to fail. A review sheet of 16 November 2026 states the test in the caption: stage 1, bare branches, no leaves, work in progress, not approved. For Cinnamomum camphora the sheet records a skeleton 17.7 m tall and 30.6 m wide, an outline overlap of 0.95 against the photograph, six main limbs, and a clear trunk to the first fork at 11 per cent of height. It also records the miss. The fine twigs fork too little, and the height of 18 m is assumed, not measured. Ulmus americana and Salix × sepulcralis are marked the same way. Where the fork height or the curtain of twigs does not meet the photograph, the model is parked.

That page is the craft. A person still decides whether six limbs and a fork at a stated height are the plant. The method is what became larger. The judgement did not leave.
A depiction under licence
Later stills are a different object, and they say so. On the lemon-scented gum sheet the same skeleton is shown as a full crown, a shaped crown with sky between the clumps, and a hero: taller trunk, balanced crown, bark carried as a picture. A field photograph sits in the row. A single twig is built from photographed leaves. The mesh is no longer only the rewriting. It is a licensed reading of it: daylight, bark, and a crown opened so the architecture can be seen.


TLA-185 is the record those pictures have to keep. It is 185 fields: the 169 of the base standard, plus 16 for climate screening. The modules cover taxonomy, traits, environment, provenance, licensing, the product, the project, site context, and the botanical asset. The open specification is published under CC BY-NC-ND 4.0. That licence is for citing the grammar. It is not a licence to the model, the still, or a filled catalogue. Licensing is itself one of the modules, five fields inside the 185.
A creative still may change the depiction. It may choose a hero bark, a shaped crown, or a studio sky. It may not drop the name, the architectural comparison, or the rights line. The picture is allowed to be a picture. The record stays the record. Nothing on these sheets is a conformance seal. The sheets mark themselves unapproved where the work is unfinished.
The inventory of 17 November 2026 counts 1,451 plants: 981 with a thumbnail on the main pages, and 470 still listed in the models-in-progress appendix. Daylight presentation, work in progress, unapproved. Craft has not been replaced. It has been asked to cover a field of that size, one parked verdict at a time.

After the fire
A crown fire removes the derived string. Leaves, fine twigs, and much of the scaffold go. In many eucalypts the axiom does not. Epicormic buds sit under the bark. A lignotuber holds a bud bank at the base. Some banksias and hakeas hold a canopy seed bank that opens with heat. What grows back is not a memorial of the crown that burned. It is the same productions, started again from what survived, at a smaller module. The black trunk with green along it is a rewriting system that was not erased.
The ash bed is the honest landscape for this work. It is wide, it is burnt, and it is not empty. The early stills are the bare scaffold after the foliage has been taken off so the rule can be checked. The later stills are the first flush: a licensed green, still attached to the same trunk, still required to meet the record. If the word phoenix is kept, it is not a finished tree appearing at once. It is the capacity of the axiom to be run again.
That return is slow. A bud does not replace a crown in a season, and a parked model does not become a hero by being wished there. The interval is the point. A record that can hold a missing twig, an assumed height, and a photograph on the same page is a kind of patience. It is as close as a specification gets to a regeneration of the person doing the work: not a repair performed on them, but a glimmer that the rule is still in the ash, and that the next module can be written.
The stills are on archivestudioplus.com. Models on that gallery may be incomplete or still in production.
Next Steps
Open Foundation materials for one studio deliverable are on the adoption guide. Citing the grammar stays free under CC BY-NC-ND.
Licensing enquiries: [email protected].