finite element line with
start point
end point
id
cross-section (optional)
local coordinate system (optional)
some property (optional)
some other property (optional)
additional settings (optional)
etc
Now in 99% of the cases, users will only specify the first 4 parameters and leave the others blank. I'm not a huge fan of to many inputs so to clean up the canvas/components, I thought about exposing the optional parameters only upon zooming in on the component.
So far I've sometimes added a secondary component with more inputs to specify a list of additional settings (similar to the "settings" panel that exists/existed in Kangaroo), but this I find rather messy.
Alternatively I guess I could quite happily live with exposing the additional parameters at the click of a button. This I can do with the ZUI as it is written? I still need to get my head round what's what in this happy world of the canvas' third dimension...
…
ut only one list of indices(i).
Equal numbers of lists of items and indices.
Which I think probably covers 99% of usage. However for the optimizations to kick in the wrap input has to be either 1 value, or a single list of values, or equally many lists of values as the i input.
I can't just upload the maths.gha file for you to test as this optimization required a change in grasshopper.dll (I needed a way to inform the component base class to stop to try and solve it for every item in (i) whenever I treat the (i) input as a datatree. I can send you an official installer later though (after lots and lots of internal testing) so you can see if it works for you.
--
David Rutten
david@mcneel.com…
Added by David Rutten at 5:29am on February 15, 2014
ve an adequately low value for vibration. Script all runs fine and matches expected results for a few test cases. Phew. Note that in excel form this calculation runs over 6000 cells of look up's and arrays, it's not a trivial calc that can practically be rearranged to make it work back from an acceptable value.
But in my results I have 99 sets of vibration-acceptable outcomes, which I'd now like to examine to see which one uses the least material in its particular combination of beam thicknesses, slab thicknesses and other stuff to find the overall most efficient system. The plan at the moment is to copy the solver Record over into Excel, extract the %'s for the different variables and post-process the info to sort the acceptable outcomes by weight.
It feels like there must be a better way that avoids taking the data out and having a gap in the parametric thinking. Is there a way to ask Galapagos to give me all (or at least, loads of) combinations for which R < 8 and then test those for the minimum weight? Can I automatically take the winning results from fitness test 1 out into an array of data that feeds into a second fitness test in the same grasshopper space? …
gain profits falls in the latter category.
The challenge here is to do the job (up to a point) using "anodyne" ways at the cost of providing slow, incomplete and quite inefficient solutions.
That said this specific vault case requires addressing 4 "classes" of problems (for instance: regions due to ccx events or alternatively circuits in graphs etc etc).
Back to business:
Creating a realistic "random" W truss of that type is one of the most challenging tasks in parametric adventures (in fact ... is the top dog by some miles). One of the many issues is an approach to manage "on-the-fly" clash situations by individually modifying nodes (without been sure that you can arrive to an all overall valid solution). Since one "path" tried may yield dead-end(s) this means keeping track of your corrective actions in a hierarchical manner and been able to follow a different "path". Another (obvious) issue is to use instance definitions for all the "components" thus achieving almost real-time response (try to manage 100K++ "solids" [sleeves, cones etc etc] to see what I mean) ... etc etc.
The big thing is: what are you going to tell to your instructors about the required code part? (that 99% mentioned) And if a "complete" solution is primarily based on "black boxes" could - in the instructor's eyes - your Master Thesis qualify as yours?
That said Vaults_V1 is achievable solely via components.…
ostly via C# because ... er ... the remaining 99% (how to do some real-life canopy and/or a real-life truss out of the relaxed line graph) is only doable via code - no ExoW/IL (so ... the 1% is indeed doable).
At first ... just double click the Kangaroo1 engine, halt the simulation AND ONLY THEN redirect the resulting line list to the ExoW/IL. As delivered neither is active.
Note: ExoW and/or IntraLattice MAY or MAY NOT work (each one has his own issues, but ExoW despite the glitches yields way better looking liquid stuff). So the liquid root may or may not be the holly grail that you expect (life sucks).
Note: As is delivered this only does a liquid node load bearing structure (ideal for Planet Utopia). Paint the thing black, do some proper pavement, populate with birds of pray, wait for the envelope def (that's freaky), put humans inside, lock the doors > massacre.
…
思った感じになりません。
balls の代わりにplanarカーブを直接入れてみましたがエラーが出ます。
ファンクションにしてみたところ、forループので作った数値が反映されていません。
ファンクションのインスタンス?を出力していないと思い上記のようにしましたがエラーが出てしまいます。
以上の事から自分の認識が正しいのかよくわからなくなりました・・・
python自体の深いところをわかっているわけではないので余計こんがらがりました。
そこで、for b in ballsはどのような条件または使い方であれば使えるのでしょうか?
そして、上記のように別のオブジェクトに対しての使い方はどのようにすればできるのでしょうか?
2:同じファンクション内のdist = rs.Distance(self.pos,b.pos)についてですが
この文章も for b in balls によってbはBallのインスタンスであると定義?されたためb.posがbの位置であると分かるのでしょうか?
pythonは定義しなくても動いてしまうのでどのような時に使えるのか文章見ただけではよくわかりません・・・
大変細かいことかもしれませんが、よりpythonをしっかりと理解するためにも、どなたかわかる方ご教授いただけると幸いです。…
.0004. [1 of 7] Writing simulation parameters...5. [2 of 6] No context surfaces...6. [3 of 6] Writing geometry...7. [4 of 6] Writing materials and constructions...8. [5 of 7] Writing schedules...9. [6 of 7] Writing loads and ideal air system...10. [7 of 7] Writing outputs...11. ...... idf file is successfully written to : c:\ladybug\unnamed\EnergyPlus\unnamed.idf12. 13. Analysis is running!...14. c:\ladybug\unnamed\EnergyPlus\eplusout.csv15. ......
Done! Read below for errors and warnings:
16. 17. Program Version,EnergyPlus, Version 8.2.7-777c1f8d79, YMD=2015.02.28 16:09,IDD_Version 8.2.718. 19. ** Warning ** IP: Note -- Some missing fields have been filled with defaults. See the audit output file for details.20. 21. ** Warning ** Version: in IDF="'8.2.7'" not the same as expected="8.2"22. 23. ** Warning ** ManageSizing: For a zone sizing run, there must be at least 1 Sizing:Zone input object. SimulationControl Zone Sizing option ignored.24. 25. ** Warning ** ManageSizing: For a plant sizing run, there must be at least 1 Sizing:Plant object input. SimulationControl Plant Sizing option ignored.26. 27. ** Severe ** GetHTSubSurfaceData: Surface Openings have too much area for base surface=F73533B3C6894C67936B_GLZP_1228. 29. ** ~~~ ** Opening Surface creating error=F73533B3C6894C67936B_GLZP_12_GLZ_1230. 31. ** Severe ** GetHTSubSurfaceData: Surface Openings have too much area for base surface=F73533B3C6894C67936B_GLZP_2532. 33. ** ~~~ ** Opening Surface creating error=F73533B3C6894C67936B_GLZP_25_GLZ_2534. 35. ** Severe ** GetHTSubSurfaceData: Surface Openings have too much area for base surface=F73533B3C6894C67936B_GLZP_2836. 37. ** ~~~ ** Opening Surface creating error=F73533B3C6894C67936B_GLZP_28_GLZ_2838. 39. ** Severe ** GetHTSubSurfaceData: Surface Openings have too much area for base surface=4BDFD67E6D0E486796CC_GLZP_940. 41. ** ~~~ ** Opening Surface creating error=4BDFD67E6D0E486796CC_GLZP_9_GLZ_942. 43. ** Severe ** GetHTSubSurfaceData: Surface Openings have too much area for base surface=4BDFD67E6D0E486796CC_GLZP_1044. 45. ** ~~~ ** Opening Surface creating error=4BDFD67E6D0E486796CC_GLZP_10_GLZ_1046. 47. ** Severe ** GetHTSubSurfaceData: Surface Openings have too much area for base surface=4BDFD67E6D0E486796CC_GLZP_1148. 49. ** ~~~ ** Opening Surface creating error=4BDFD67E6D0E486796CC_GLZP_11_GLZ_1150. 51. ** Severe ** GetHTSubSurfaceData: Surface Openings have too much area for base surface=4BDFD67E6D0E486796CC_GLZP_1552. 53. ** ~~~ ** Opening Surface creating error=4BDFD67E6D0E486796CC_GLZP_15_GLZ_1554. 55. ** Severe ** GetHTSubSurfaceData: Surface Openings have too much area for base surface=4BDFD67E6D0E486796CC_GLZP_2456. 57. ** ~~~ ** Opening Surface creating error=4BDFD67E6D0E486796CC_GLZP_24_GLZ_2458. 59. ** Severe ** GetHTSubSurfaceData: Surface Openings have too much area for base surface=4BDFD67E6D0E486796CC_GLZP_2560. 61. ** ~~~ ** Opening Surface creating error=4BDFD67E6D0E486796CC_GLZP_25_GLZ_2562. 63. ** Severe ** GetHTSubSurfaceData: Surface Openings have too much area for base surface=4BDFD67E6D0E486796CC_GLZP_3064. 65. ** ~~~ ** Opening Surface creating error=4BDFD67E6D0E486796CC_GLZP_30_GLZ_3066. 67. ** Severe ** GetHTSubSurfaceData: Surface Openings have too much area for base surface=4BDFD67E6D0E486796CC_GLZP_3268. 69. ** ~~~ ** Opening Surface creating error=4BDFD67E6D0E486796CC_GLZP_32_GLZ_3270. 71. ** Severe ** GetHTSubSurfaceData: Surface Openings have too much area for base surface=4BDFD67E6D0E486796CC_GLZP_3472. 73. ** ~~~ ** Opening Surface creating error=4BDFD67E6D0E486796CC_GLZP_34_GLZ_3474. 75. ** Warning ** GetSurfaceData: Very small surface area[2.94495E-004], Surface=F73533B3C6894C67936B_GLZP_076. 77. ** Warning ** GetSurfaceData: Very small surface area[3.84753E-004], Surface=F73533B3C6894C67936B_GLZP_178. 79. ** Warning ** GetSurfaceData: Very small surface area[9.16905E-004], Surface=F73533B3C6894C67936B_GLZP_380. 81. ** Warning ** GetSurfaceData: Very small surface area[4.96186E-004], Surface=F73533B3C6894C67936B_GLZP_482. 83. ** Warning ** GetSurfaceData: Very small surface area[2.37373E-005], Surface=F73533B3C6894C67936B_GLZP_684. 85. ** Warning ** GetSurfaceData: Very small surface area[6.35824E-004], Surface=F73533B3C6894C67936B_GLZP_786. 87. ** Warning ** GetSurfaceData: Very small surface area[5.86549E-004], Surface=F73533B3C6894C67936B_GLZP_888. 89. ** Warning ** GetSurfaceData: Very small surface area[7.63765E-004], Surface=F73533B3C6894C67936B_GLZP_1090. 91. ** Severe ** GetSurfaceData: Zero or negative surface area[-8.09566E-004], Surface=F73533B3C6894C67936B_GLZP_1292. 93. ** Warning ** GetSurfaceData: Very small surface area[1.51701E-004], Surface=F73533B3C6894C67936B_GLZP_1394. 95. ** Warning ** GetSurfaceData: Very small surface area[9.29917E-004], Surface=F73533B3C6894C67936B_GLZP_1596. 97. ** Warning ** GetSurfaceData: Very small surface area[2.94451E-004], Surface=F73533B3C6894C67936B_GLZP_1698. 99. ** Warning ** GetSurfaceData: Very small surface area[8.03294E-004], Surface=F73533B3C6894C67936B_GLZP_17100. 101. ** Warning ** GetSurfaceData: Very small surface area[6.83026E-004], Surface=F73533B3C6894C67936B_GLZP_18102. 103. ** Warning ** GetSurfaceData: Very small surface area[9.29917E-004], Surface=F73533B3C6894C67936B_GLZP_20104. 105. ** Warning ** GetSurfaceData: Very small surface area[3.19851E-005], Surface=F73533B3C6894C67936B_GLZP_21106. 107. ** Warning ** GetSurfaceData: Very small surface area[7.63765E-004], Surface=F73533B3C6894C67936B_GLZP_23108. 109. ** Severe ** GetSurfaceData: Zero or negative surface area[-4.05899E-004], Surface=F73533B3C6894C67936B_GLZP_25110. 111. ** Warning ** GetSurfaceData: Very small surface area[6.35824E-004], Surface=F73533B3C6894C67936B_GLZP_27112. 113. ** Severe ** GetSurfaceData: Zero or negative surface area[-9.91146E-004], Surface=F73533B3C6894C67936B_GLZP_28114. 115. ** Warning ** GetSurfaceData: Very small surface area[2.70158E-004], Surface=F73533B3C6894C67936B_GLZP_29116. 117. ** Warning ** GetSurfaceData: Very small surface area[3.22781E-004], Surface=F73533B3C6894C67936B_GLZP_30118. 119. ** Warning ** GetSurfaceData: Very small surface area[4.67821E-004], Surface=F73533B3C6894C67936B_GLZP_33120. 121. ** Warning ** GetSurfaceData: Very small surface area[3.22737E-004], Surface=F73533B3C6894C67936B_GLZP_34122. 123. ** Warning ** GetSurfaceData: Very small surface area[2.65634E-004], Surface=4BDFD67E6D0E486796CC_GLZP_0124. 125. ** Warning ** GetSurfaceData: Very small surface area[4.70736E-004], Surface=4BDFD67E6D0E486796CC_GLZP_1126. 127. ** Warning ** GetSurfaceData: Very small surface area[3.42507E-004], Surface=4BDFD67E6D0E486796CC_GLZP_3128. 129. ** Warning ** GetSurfaceData: Very small surface area[5.89276E-004], Surface=4BDFD67E6D0E486796CC_GLZP_4130. 131. ** Warning ** GetSurfaceData: Very small surface area[1.91146E-004], Surface=4BDFD67E6D0E486796CC_GLZP_6132. 133. ** Warning ** GetSurfaceData: Very small surface area[9.71205E-004], Surface=4BDFD67E6D0E486796CC_GLZP_7134. 135. ** Warning ** GetSurfaceData: Very small surface area[4.34494E-004], Surface=4BDFD67E6D0E486796CC_GLZP_8136. 137. ** Severe ** GetSurfaceData: Zero or negative surface area[-3.60159E-004], Surface=4BDFD67E6D0E486796CC_GLZP_9138. 139. ** Severe ** GetSurfaceData: Zero or negative surface area[-1.11946E-004], Surface=4BDFD67E6D0E486796CC_GLZP_10140. 141. ** Severe ** GetSurfaceData: Zero or negative surface area[-3.41257E-004], Surface=4BDFD67E6D0E486796CC_GLZP_11142. 143. ** Severe ** GetSurfaceData: Zero or negative surface area[-8.21483E-005], Surface=4BDFD67E6D0E486796CC_GLZP_15144. 145. ** Warning ** GetSurfaceData: Very small surface area[2.65716E-004], Surface=4BDFD67E6D0E486796CC_GLZP_16146. 147. ** Warning ** GetSurfaceData: Very small surface area[4.84044E-004], Surface=4BDFD67E6D0E486796CC_GLZP_17148. 149. ** Warning ** GetSurfaceData: Very small surface area[7.12297E-004], Surface=4BDFD67E6D0E486796CC_GLZP_19150. 151. ** Warning ** GetSurfaceData: Very small surface area[6.14324E-004], Surface=4BDFD67E6D0E486796CC_GLZP_22152. 153. ** Warning ** GetSurfaceData: Very small surface area[8.88887E-004], Surface=4BDFD67E6D0E486796CC_GLZP_23154. 155. ** Severe ** GetSurfaceData: Zero or negative surface area[-9.89060E-004], Surface=4BDFD67E6D0E486796CC_GLZP_24156. 157. ** Severe ** GetSurfaceData: Zero or negative surface area[-1.14849E-003], Surface=4BDFD67E6D0E486796CC_GLZP_25158. 159. ** Warning ** GetSurfaceData: Very small surface area[4.00479E-004], Surface=4BDFD67E6D0E486796CC_GLZP_27160. 161. ** Warning ** GetSurfaceData: Very small surface area[6.63061E-005], Surface=4BDFD67E6D0E486796CC_GLZP_28162. 163. ** Warning ** GetSurfaceData: Very small surface area[1.09018E-004], Surface=4BDFD67E6D0E486796CC_GLZP_29164. 165. ** Severe ** GetSurfaceData: Zero or negative surface area[-2.49326E-005], Surface=4BDFD67E6D0E486796CC_GLZP_30166. 167. ** Severe ** GetSurfaceData: Zero or negative surface area[-3.17446E-004], Surface=4BDFD67E6D0E486796CC_GLZP_32168. 169. ** Warning ** GetSurfaceData: Very small surface area[8.60686E-004], Surface=4BDFD67E6D0E486796CC_GLZP_33170. 171. ** Severe ** GetSurfaceData: Zero or negative surface area[-2.48515E-005], Surface=4BDFD67E6D0E486796CC_GLZP_34172. 173. ** Fatal ** GetSurfaceData: Errors discovered, program terminates.174. 175. ...Summary of Errors that led to program termination:176. 177. ..... Reference severe error count=24178. 179. ..... Last severe error=GetSurfaceData: Zero or negative surface area[-2.48515E-005], Surface=4BDFD67E6D0E486796CC_GLZP_34180. 181. ************* Warning: Node connection errors not checked - most system input has not been read (see previous warning).182. 183. ************* Fatal error -- final processing. Program exited before simulations began. See previous error messages.184. 185. ************* EnergyPlus Warmup Error Summary. During Warmup: 0 Warning; 0 Severe Errors.186. 187. ************* EnergyPlus Sizing Error Summary. During Sizing: 2 Warning; 0 Severe Errors.188. 189. ************* EnergyPlus Terminated--Fatal Error Detected. 41 Warning; 24 Severe Errors; Elapsed Time=00hr 00min 1.51sec190.…
y to heaven (or hell) is full of pain,frustration and tears. In plain English: if you are not totally committed (and willing to pay the heavy price) ... well ... what about forgetting all that freaky stuff? (the best option, trust me)
Note: 99% of beginners dream to learn programing in order to make geometry. But the truth is that this is the least (and rather the most insignificant) that you can achieve especially when working in teams with lot's of CAD/MCAD apps (and verticals) in the practice of tomorrow (bad news: tomorrow is already yesterday).
Anyway: How to go to Hell in just 123 easy steps
Step 1: get the cookiesThe bible PlanA: C# In depth (Jon Skeet).The bible PlanB: C# Step by step (John Sharp).The bible PlanC: C# 5.0 (J/B Albahari) > my favoriteThe reference: C# Language specs ECMA-334The candidates:C# Fundamentals (Nakov/Kolev & Co)C# Head First (Stellman/Greene)C# Language (Jones)Step 2: read the cookies (computer OFF)Step 3: re-read the cookies (computer OFF)...
Step 122: re-read the cookies (computer OFF)Step 123: Open computer > burn computer > computers are a bad thing (not to mention the Skynet trivial thingy).May The Force (the Dark Option) be with you.
…
ore simplest way, for example i need to declare min distance (minDist) like huge number to avoid overlapping values. may be find way do not use second loop in script at all . In the neoarchaic's script has line 99 (j = pts.ClosestIndex(pt)) but i have no idea how to do it in C#.
2. In a file i prepared GH+Hoopsnake and GH+Anemone solution for this script . Can i use the same principle in C# (shipt list with wrap values )? This solution I need for studying list operations in C#.
Thank you for helping .
ps..My script:
int num = x.Count; double minDist,dist; int minI = 0; int i = 0;
Point3d pt; List<Point3d> z = new List<Point3d>();
while (i < num) { i++; z.Add(x[minI]); pt = x[minI]; minDist = 1000000000000; x.RemoveAt(minI);
for (int n = 0; n < x.Count; n++){
dist = x[n].DistanceTo(pt);
if ( dist < minDist) { minI = n; minDist = dist; } } } A = z;
…
ng is deciding how and where to store your data. If you're writing textual code using any one of a huge number of programming languages there are a lot of different options, each with its own benefits and drawbacks. Sometimes you just need to store a single data point. At other times you may need a list of exactly one hundred data points. At other times still circumstances may demand a list of a variable number of data points.
In programming jargon, lists and arrays are typically used to store an ordered collection of data points, where each item is directly accessible. Bags and hash sets are examples of unordered data storage. These storage mechanisms do not have a concept of which data comes first and which next, but they are much better at searching the data set for specific values. Stacks and queues are ordered data structures where only the youngest or oldest data points are accessible respectively. These are popular structures for code designed to create and execute schedules. Linked lists are chains of consecutive data points, where each point knows only about its direct neighbours. As a result, it's a lot of work to find the one-millionth point in a linked list, but it's incredibly efficient to insert or remove points from the middle of the chain. Dictionaries store data in the form of key-value pairs, allowing one to index complicated data points using simple lookup codes.
The above is a just a small sampling of popular data storage mechanisms, there are many, many others. From multidimensional arrays to SQL databases. From readonly collections to concurrent k-dTrees. It takes a fair amount of knowledge and practice to be able to navigate this bewildering sea of options and pick the best suited storage mechanism for any particular problem. We did not wish to confront our users with this plethora of programmatic principles, and instead decided to offer only a single data storage mechanism.*
Data storage in Grasshopper
In order to see what mechanism would be optimal for Grasshopper, it is necessary to first list the different possible ways in which components may wish to access and store data, and also how families of data points flow through a Grasshopper network, often acquiring more complexity over time.
A lot of components operate on individual values and also output individual values as results. This is the simplest category, let's call it 1:1 (pronounced as "one to one", indicating a mapping from single inputs to single outputs). Two examples of 1:1 components are Subtraction and Construct Point. Subtraction takes two arguments on the left (A and B), and outputs the difference (A-B) to the right. Even when the component is called upon to calculate the difference between two collections of 12 million values each, at any one time it only cares about three values; A, B and the difference between the two. Similarly, Construct Point takes three separate numbers as input arguments and combines them to form a single xyz point.
Another common category of components create lists of data from single input values. We'll refer to these components as 1:N. Range and Divide Curve are oft used examples in this category. Range takes a single numeric domain and a single integer, but it outputs a list of numbers that divide the domain into the specified number of steps. Similarly, Divide Curve requires a single curve and a division count, but it outputs several lists of data, where the length of each list is a function of the division count.
The opposite behaviour also occurs. Common N:1 components are Polyline and Loft, both of which consume a list of points and curves respectively, yet output only a single curve or surface.
Lastly (in the list category), N:N components are also available. A fair number of components operate on lists of data and also output lists of data. Sort and Reverse List are examples of N:N components you will almost certainly encounter when using Grasshopper. It is true that N:N components mostly fall into the data management category, in the sense that they are mostly employed to change the way data is stored, rather than to create entirely new data, but they are common and important nonetheless.
A rare few components are even more complex than 1:N, N:1, or N:N, in that they are not content to operate on or output single lists of data points. The Divide Surface and Square Grid components want to output not just lists of points, but several lists of points, each of which represents a single row or column in a grid. We can refer to these components as 1:N' or N':1 or N:N' or ... depending on how the inputs and outputs are defined.
The above listing of data mapping categories encapsulate all components that ship with Grasshopper, though they do not necessarily minister to all imaginable mappings. However in the spirit of getting on with the software it was decided that a data structure that could handle individual values, lists of values, and lists of lists of values would solve at least 99% of the then existing problems and was thus considered to be a 'good thing'.
Data storage as the outcome of a process
If the problems of 1:N' mappings only occurred in those few components to do with grids, it would probably not warrant support for lists-of-lists in the core data structure. However, 1:N' or N:N' mappings can be the result of the concatenation of two or more 1:N components. Consider the following case: A collection of three polysurfaces (a box, a capped cylinder, and a triangular prism) is imported from Rhino into Grasshopper. The shapes are all exploded into their separate faces, resulting in 6 faces for the box, 3 for the cylinder, and 5 for the prism. Across each face, a collection of isocurves is drawn, resembling a hatching. Ultimately, each isocurve is divided into equally spaced points.
This is not an unreasonably elaborate case, but it already shows how shockingly quickly layers of complexity are introduced into the data as it flows from the left to the right side of the network.
It's no good ending up with a single huge list containing all the points. The data structure we use must be detailed enough to allow us to select from it any logical subset. This means that the ultimate data structure must contain a record of all the mappings that were applied from start to finish. It must be possible to select all the points that are associated with the second polysurface, but not the first or third. It must also be possible to select all points that are associated with the first face of each polysurface, but not any subsequent faces. Or a selection which includes only the fourth point of each division and no others.
The only way such selection sets can be defined, is if the data structure contains a record of the "history" of each data point. I.e. for every point we must be able to figure out which original shape it came from (the cube, the cylinder or the prism), which of the exploded faces it is associated with, which isocurve on that face was involved and the index of the point within the curve division family.
A flexible mechanism for variable history records.
The storage constraints mentioned so far (to wit, the requirement of storing individual values, lists of values, and lists of lists of values), combined with the relational constraints (to wit, the ability to measure the relatedness of various lists within the entire collection) lead us to Data Trees. The data structure we chose is certainly not the only imaginable solution to this problem, and due to its terse notation can appear fairly obtuse to the untrained eye. However since data trees only employ non-negative integers to identify both lists and items within lists, the structure is very amenable to simple arithmetic operations, which makes the structure very pliable from an algorithmic point of view.
A data tree is an ordered collection of lists. Each list is associated with a path, which serves as the identifier of that list. This means that two lists in the same tree cannot have the same path. A path is a collection of one or more non-negative integers. Path notation employs curly brackets and semi-colons as separators. The simplest path contains only the number zero and is written as: {0}. More complicated paths containing more elements are written as: {2;4;6}. Just as a path identifies a list within the tree, an index identifies a data point within a list. An index is always a single, non-negative integer. Indices are written inside square brackets and appended to path notation, in order to fully identify a single piece of data within an entire data tree: {2,4,6}[10].
Since both path elements and indices are zero-based (we start counting at zero, not one), there is a slight disconnect between the ordinality and the cardinality of numbers within data trees. The first element equals index 0, the second element can be found at index 1, the third element maps to index 2, and so on and so forth. This means that the "Eleventh point of the seventh isocurve of the fifth face of the third polysurface" will be written as {2;4;6}[10]. The first path element corresponds with the oldest mapping that occurred within the file, and each subsequent element represents a more recent operation. In this sense the path elements can be likened to taxonomic identifiers. The species {Animalia;Mammalia;Hominidea;Homo} and {Animalia;Mammalia;Hominidea;Pan} are more closely related to each other than to {Animalia;Mammalia; Cervidea;Rangifer}** because they share more codes at the start of their classification. Similarly, the paths {2;4;4} and {2;4;6} are more closely related to each other than they are to {2;3;5}.
The messy reality of data trees.
Although you may agree with me that in theory the data tree approach is solid, you may still get frustrated at the rate at which data trees grow more complex. Often Grasshopper will choose to add additional elements to the paths in a tree where none in fact is needed, resulting in paths that all share a lot of zeroes in certain places. For example a data tree might contain the paths:
{0;0;0;0;0}
{0;0;0;0;1}
{0;0;0;0;2}
{0;0;0;0;3}
{0;0;1;0;0}
{0;0;1;0;1}
{0;0;1;0;2}
{0;0;1;0;3}
instead of the far more economical:
{0;0}
{0;1}
{0;2}
{0;3}
{1;0}
{1;1}
{1;2}
{1;3}
The reason all these zeroes are added is because we value consistency over economics. It doesn't matter whether a component actually outputs more than one list, if the component belongs to the 1:N, 1:N', or N:N' groups, it will always add an extra integer to all the paths, because some day in the future, when the inputs change, it may need that extra integer to keep its lists untangled. We feel it's bad behaviour for the topology of a data tree to be subject to the topical values in that tree. Any component which relies on a specific topology will no longer work when that topology changes, and that should happen as seldom as possible.
Conclusion
Although data trees can be difficult to work with and probably cause more confusion than any other part of Grasshopper, they seem to work well in the majority of cases and we haven't been able to come up with a better solution. That's not to say we never will, but data trees are here to stay for the foreseeable future.
* This is not something we hit on immediately. The very first versions of Grasshopper only allowed for the storage of a single data point per parameter, making operations like [Loft] or [Divide Curve] impossible. Later versions allowed for a single list per parameter, which was still insufficient for all but the most simple algorithms.
** I'm skipping a lot of taxonometric classifications here to keep it simple.…
Added by David Rutten at 2:22pm on January 20, 2015