Agenda 2016 07 08

new

Welcome to Jamie Luo from Warwick University

• Jamie's focus the coming months - filaments

General

Progress on iDyno 2 the past few weeks:

• Restructured and cleaned up Collision class
• new ClassRef class
• agent growth reaction diffusion manager fixed
• Rob has made progress on boundaries wellmixed/ connected compartments but mentioned this is not all functional yet.

biomass components now as separate double aspects or as map, replacing internal products

• it is probably easiest to always register all components in a map, otherwise it becomes difficult to get a total mass to calculate the agent volume and now what aspects to split when dividing.

Speed improvements

• Check output level before constructing string
• SplitTree

Gui nterface and xml output

• newBlank is now being replaced by Constructable paradigm
• Bundle and Pile for easy ModelNode management of Maps and Lists

code design issues

Coding general keypoints

• Random number usage
• Units
• well mixed boundary
• steady state reaction diffusion
• full non-cartesian simulation
• Clean-up
• Hunting and eliminating bugs

Unfinished coding parts

agent package

• Agent register death method
• Make body relocate robust for relocations between 2d and 3d domains
• Adding complex body types with gui

aspect package

• Aspect redirerecting stored in output and editable in gui
• Thinkover - radii and agent size in 2d representation
• Remove or finish filial linkage methods
• in some scenarios eps cumulative seems to cause a stack overflow error
• investigate the usage of the internal production event now that rob has intergrated it (partly?) in transient diffusion solver
• level pull distance rather than adding up
• clean-up rod division and make sure rods perform tip-tip divide
• internal products are currently ignored for volume calculation

Boundary package

• proper initiation from protocol file
• node construction
• detachment, attachment
• biofilm boudary implemenentation, uses aspects thus should become redirectable or restructured in an other way
• Membrane boundary layers are unfinished

dataIO

• Log: allow for different display and log file log levels

grid

• setting difusivity other than 1 and settable from xml
• dimension min other than 0

guiTools

• A lot of classes are not implementing the model node structure yet and thus are not loaded in the gui yet.
• A lot of classes do not allow for editing (set node method missing or unfinished)

Agent container

• move all aspect related methods out of the agent container they should be in redirectable classes and angent container should remain general
• make rtree / splittree settable

JSBML

• is not used, leave out of release version if not used

nodeFactory

• finish replacing newBlank paradigm with constructable paradigm
• finish Bundle and Pile implementation

ProcessManagers

• upon event processManagers?

Shape

• what is still to be done for non-Cartesian shapes

Spatial registry

• splittree is faster but still has some glitches: artefacts with rods in periodic domains, high amount of overlap in one point will cause stack overflow
• spatial registry interface is biassed towards rTree

surface

• bounding volumes for non cartesian domains?
• Collision and adhesion functions settable from protocol file

Writing: what is our status

• Good progress on general framework description
• case studies are hold back by unfinished code.

Planning