Talk:L-system

This paragraph is unclear to me:

As an L-system these tilings are called Penrose's rhombuses and Penrose's tiles. Above pictures were generated for n=6 as an L-system. If we properly superimpose Penrose tiles as an L-system we get next tiling:

It doesn't define what n is. It doesn't define what "properly superimpose as an L-system" means. It doesn't say which L-system was used to generate those images. I can't guess these from context. --LC

in the canter dust example,it contains a line of "constant" about some 60 degrees. what's that line doing there? I think it is wrong.

in the Koch snowflake example, shouldn't the plus and minus sign be part of the alphabet?

i think the penrose tiling example is a bit forced. Penrose tilings are aperiodic. So, i doubt it can be reduced to a string replacement system.

also, i'm not sure L-Sytem has a place in serious math. If it does, it is or relates to formal logic. Should there be a mention?

thanks.

Xah P0lyglut 11:18, 2004 Jan 15 (UTC)

Constants?

In Example 3, Cantor Dust, what are the constants for? (I'm not even sure if that's one, two, or three constants you're trying to list.) You never seem to use them.

Concerning the Koch curve shouldn't + and - be constants (clearly there're no productions for them). According to the definition stated, the alphabet consists of variables, hence I disagree with the poster asking that they should be included in the alphabet.

Koch curve

I would like to see one higher order iteration (n=20 ?) of the Koch curve. IMO, the result is illustrative and pretty impressive especially for people new to the subject.

Need new Penrose time images

The three images of Penrose tilings are all untagged, and they don't have authorship info (dating from before "the big conversion"). I guess when the next untagged-image jihad gets underway these images may well meet with infinite justice, so we need to start thinking about replacements for them ASAP. It's a bummer, as they're really attractive images. -- Finlay McWalter | Talk 15:17, 7 October 2005 (UTC)

These images were made by me with Winfract version 18.21, and we can tag them as public domain. I'll do that right the way. --xJaM 14:02, 21 October 2005 (UTC)

Example 1 seems strange to me

Lindenmayer's original L-system for modelling the growth of algae.

variables : A B

constants : none

start  : A

rules  : (A → AB), (B → A)

which produces:

n=0 : A → AB

n=1 : AB → ABA

n=2 : ABA → ABAAB

n=3 : ABAAB → ABAABABA

I would expect it to look like this more in keeping with how the others are shown:

n=0 : A

n=1 : AB

n=2 : ABA

n=3 : ABAAB

n=4 : ABAABABA

Am I missing something? Hogan 02:11, 29 April 2006 (UTC)

Yes, you are right - the format of Example 1 was not consistent with the format of other examples. I have fixed it. Gandalf61 08:16, 2 May 2006 (UTC)

As I had problmes to follow the intial example I would suggest this (only in the first example, to make the principle more clear):

   n=0:         A              start/axiom/initiator
               / \
   n=1:       A   B            the single A spawns into an A followed by a B
             /|    \
   n=2:     A B     A          the former A again spawns into AB, B turnes into A
           /| |     |\
   n=3:   A B A     A B        note all A's producing a copy of themself in the first place, then a B, which turns
         /| | |\    |\ \
   n=4: A B A A B   A B A      into an A one generation later, starting to spawn/repeat/recurse then

which is inspired by the (existing) external link [1] (13 MB!). Maybe an image instead of the ASCII graphic would be even more apropriate?

Also the Fibonacci example 2 refers to example 1. Then ex. 1 should have a start/axiom/initiator of B (and the above diagram extended accordingly), else the reference doesn't hold true! - Deerwood (talk) 05:02, 1 July 2008 (UTC)

Example 1 seems clearly wrong. Also, it's unclear what the terminology (A → AB), (B → A) means. Does (A → AB) mean "A gets followed by AB"? Or does it mean "A gets replaced by AB"?

If (A → AB means "A gets followed by AB" then example 1 is clearly incorrect. In that case, we would have

n=0 : A

n=1 : AAB

n=2 : AABA

n=3 : AABAAB

n=4 : AABAABA and so on. In short, you'd always get nothing but a repetition of AAB over and over again. Since that's obviously not what an L-system does, your explanation doesn't seem to have made clear what's going on.

If, on the other hand, (A → AB) means "A gets replaced by AB," then your first example is also clearly incorrect. In that case, we would get

n=0 : A

n=1 : AB

n=2 : ABA

n=3 : ABAB

n=4 : ABABA and so on. In short, a mere repetition of AB over and over again. This is also clearly not what an L-system does.

So either I'm completely missing something, or the rules you've given are being used in some way that's not at all clear from your explanation, or Example 1 is simply incorrect.

Either way, some more clarification of exactly what the symbols mean and precisely how the rules are supposed to operate would be a big help.

Another example: Heighway dragon

Another nice example is a dragon curve. It is described by the following L-system:

variables : L R

constants : + −

start  : R

rules  : (R → R+L), (L → R−L)

meaning, respectively:

L, R — step forward

plus — turn right by 90°

minus — turn left by 90°

Some initial strigns are:

n=0 : R

n=1 : R+L

n=2 : R+L+R−L

n=3 : R+L+R−L+R+L−R−L

n=4 : R+L+R−L+R+L−R−L+R+L+R−L−R+L−R−L

CiaPan 20:53, 6 October 2006 (UTC)

Merge with graftal article?

I've added graftal to the "See Also" section, but does it really need a separate article? There doesn't seem to be a whole lot of difference between L-systems and graftals. Perhaps the articles should be merged together. What do you think? -- Sakurambo 桜ん坊 13:53, 15 May 2007 (UTC)

This is a good suggestion, since, as you noted, there is not much difference. In fact, a graftal is an L-System. So I definitely support this idea. Kwvan (talk) 18:29, 12 October 2009 (UTC)

Better examples

Some of the examples given in this article are a bit rubbish. The first example provides no explanation of what is going on. What does the string "ABAABABAABAABABAABABAABAABABAABAAB" have to do with algae? If the only important information is the length of each sequence, then how does this differ from Fibonacci's modelling of rabbit populations? The second example is completely redundant — the first example also generates a Fibonacci sequence, in case you hadn't noticed, and in any case, there's a much simpler algorithm for calculating this series, so in what way is this a useful example of the capabilities of an L-system? Examples 5 and 9 are quite useless without the rules used to produce these figures. -- Sakurambo 桜ん坊 14:10, 15 May 2007 (UTC)

The Penrose tiling algorithm is given at the article of the same name, though the syntax is different. I could try translating between notations myself, but I am loathe to introduce errors. The modified Koch curve I'm not sure about, from the Koch curve page the standard curve uses

   Alphabet : F
   Constants : +, −
   Axiom : F++F++F
   Production rules:
   F → F−F++F−F 

I am guessing that the variant would alternate the production rule between F → F−F++F−F and F → F+F--F+F though this is pure speculation from the description, I have no idea how it was actually generated. As to the criticism of the first examples, algae was merely what he was historically trying to model the growth of, I don't think it is meant to represent the useful capabilities of the system, more to provide an example that correlates it with a well known algorithm. Nazlfrag (talk) 06:27, 13 June 2008 (UTC)

Explanations

I added an explanation ASCII art to the first example and hope this doesn't get deleted immediately without discussion.

Visitors/readers of an encyclopedia ain't expected to be mathematicians, programmers or specialists in any way, are they? Readers should at least be able to understand the concept/essentials ... may be, one or the other reader then gets inspired to read on, learn and understand? And, may be, contribute? - Deerwood (talk) 03:38, 3 July 2008 (UTC)

IFS

Is there a reference for the statement that "L-systems can also be used to generate self-similar fractals such as iterated function systems"? Richard Pinch (talk) 21:17, 16 July 2008 (UTC)

Examples...

Example 1 and 2 are practically the same system; having both is somewhat redundant.

Example 3(?) the dust example, should be changed to use conventional notation namely F for draw forward and f for move forward.

Example 5(?) the penrose tilings example, doesn't supply the rules for the system; the rules need to be added or the example scrapped.

All the examples really need reworking; the term 'constants' isn't one that's really used in any of the published papers on L-systems.

The article could use an example of the various common types of L-system, stochastic (random), parametric and context sensitive.

87.194.144.173 (talk) 15:36, 3 May 2010 (UTC)

Cantor example is nothing like the diagram

The rules give:

n = 1 : ABA
n = 2 : ABABBBABA
n = 3 : ABABBBABABBBBBBBBBABABBBABA
n = 4 : ABABBBABABBBBBBBBBABABBBABABBBBBBBBBBBBBBBBBBBBBBBBBBBABABBBABABBBBBBBBBABABBBABA

now only if each line is scaled to the same lenghth might you get something like the picture. --Paddy (talk) 23:33, 14 May 2011 (UTC)

The scaling is so typical, perhaps it was overseen to mention it. The L-system itself is just the evolution of the character strings from one generation to the next. On top of that one may feed the character string of each generation as command sequence into a drawing machine like the (LOGO) Turtle. To obtain convergence of some kind one has to care how the characters relate to commands. This means adjusting the angles and step sizes, and also the orientation of the turtle at the start point, if in 2D. In the Cantor example this is simple -- three times as many characters, each a move forward, means one third of the step size from one generation to the next.--LutzL (talk) 13:19, 15 May 2011 (UTC)

Chomskys "Ultra-conservative" views

Without wanting to make any statement about the facts, could the sentences "yes: in language theory, Chomsky is ultra-conservative" please be reworked or removed? This does not sound very encyclopedia-like. In general, the whole paragraph "Chomsky's conservative...refer to Chomsky.[2]" should probably not in the introduction of the article but maybe at some later point. 95.117.217.100 (talk) 11:24, 25 July 2011 (UTC)

I have removed that paragraph from the article's lead. I am not completely sure what the editor who wrote this was trying to say, but in any case it has no relevance to the subject of this article. Gandalf61 (talk) 11:40, 25 July 2011 (UTC)

Context-free systems

I don't know L-systems so well but in general formal languages the context-free ones are a strictly larger class that the mentioned equivalents of regular languages. It seems to apply here as well. Can someone competent look at it and eventually fix it? Neználek (talk) 13:17, 11 April 2012 (UTC)

This Article Needs A ReWrite

As it is now, I can't tell if it's supposed to be about some type of mathematics or what. The article has tons of images of random plants and is talking all about biology and other non-math things. I'm pretty sure the concept of "L-System" is purely mathematical.... please remove these ancillary or tangential "connections" or at least dont make them the focus of the article anymore. — Preceding unsigned comment added by 71.201.95.139 (talk) 21:02, 8 May 2013 (UTC)

L-systems are a mathematical concept that have applications in modelling plants and other organisms. We can't write about L-systems without including these applications. However, I have modified the lead paragraph so that it says what an L-system is before it mentions applications in biology. Gandalf61 (talk) 08:48, 9 May 2013 (UTC)