Talk:Universal Constructor

Unsorted talk
Hey my dear ferkel, since you are researching von neumann automata, can you put up a better set of references? Maybe some of his articles are online somewhere...


 * Don't you wish. Actually, don't I wish, too.  Finding the material online is not easy.  I think that some notes from the Hixon Symposia are available, in the form of a PDF.  Otherwise, the best material for von Neumann cellular automata (the 29-state kind) is contained in Burk's 1970 volume Essays on Cellular Automata.  The 1966 volume, Theory of Self-Reproducing Automata is good background but, it is not as easy to understand as the material in the 1970 volume.  William R. Buckley 18:02, 18 June 2006 (UTC)


 * The PDF mentioned above is available at the URL http://www.cs.caltech.edu/courses/cs191/paperscs191/ in a file named VonNeumann56.pdf to which I refer you. William R. Buckley 18:20, 18 June 2006 (UTC)


 * Further review suggests the above document is not for Hixon. Also, the following URL seems to provide access to the Hixon documents, given fee payment, of course.  William R. Buckley 19:57, 18 June 2006 (UTC)

I noticed that some details were not correct on this page when I viewed it on July 13, 2005, so I correct those details. As I am currently involved in ongoing research efforts with von Neumann cellular automata, I am well qualified to make these corrections.


 * Hello William R. Buckley. (Please remember to sign your comments on Talk pages.) Two queries: firstly, from what you've written it looks like the Mange et al paper is based on your work - is this correct? Secondly, I'm surprised you've said that the constructor is expected to soon become practical, since there would have to be room for several machines plus their tapes, and this would require a vast amount of CA space - could you explain this? Thanks, Ferkel 15:50, 14 July 2005 (UTC)


 * William R. Buckley, I'm having more concerns about your edits. Reading Pesavento's papers (eg. ) it is clear that they did follow JvN's original state rules for the machine shown on the Universal Constructor page. They also presented an extension of it that allows the machine to be smaller and more efficient. In their software they give both .jvn (original ruleset) and .evn (extended ruleset) automata, and the image on the article page is the .jvn one, and thus is the one using JvN's original ruleset. I agree that the exact details of the layout are theirs and not JvN's but it is not correct to say that they deviated from JvN's ruleset. Please respond to these issues - thanks! Ferkel 10:12, 28 July 2005 (UTC)

Dr. Hutton, with regard to the design published in the journal Artificial Life, see the section 11.1, where it says

“11.1 The Cell States

In this transition rule each cell can assume one of 32 possible states.“

So, the Artificial Life version is not a von Neumann 29-state self-replicating cellular automaton.

The design presented in the "Artificial Worlds and Urban Studies" volume is based upon only the 29-state rule set of von Neumann. However, that design is not capable of engaging in self-replication, because it cannot provide the triggering signal from outside; i.e. the parent cannot trigger the child into operation. Further, this assertion of mine ignores trivial changes, like alterations to signal timing, and the extension, retraction, and re-routing of signal lines. Non-trivial changes will involve configurations that operate upon signals; i.e. convert their form or content.

It is only upon application of non-trivial changes to automaton configuration that the 29-state design (given in the book, as mentioned above) will be able to effect a cycle of self-replication.

I would argue that one should study the design of Nobili-Pesavento in detail, before arguing that their work satisfies the insights of von Neumann. After all, it is upon the assertions of our learned few that societies come to believe facts.

William R. Buckley

To editor at IP Address <68.251.145.22>

Please cite the reference to von Neumann's recognition of the nature of Open-Ended Evolution. Also compare this statement with the work of Myhill, and his proof of the existence of "machine-building machines from which the successive generations of descendants become in a certain sense more and more intelligent." William R. Buckley 23:25, 19 October 2005 (UTC)

Finally, also comment on the source of mutation suggested as a component of Open-Ended Evolution. Is this the only source of such mutation, errors in copying the genetic message? William R. Buckley 23:29, 19 October 2005 (UTC)

The Wikipedia needs a spelling checker. William R. Buckley 23:30, 19 October 2005 (UTC)

von Neumann specificity
Most of the Universal Constructor page is specific to von Neumann's machine, or variations thereof. Perhaps that content should be moved to a page with a more specific name (e.g., Von Neumann's Universal Constructor), and this page be left to deal with the general topic (with, of course, a link to the new page). Jmdyck 21:59, 15 May 2006 (UTC)


 * I would agree with such a move, the creation of a general category, and links therein to the various forms which have been developed. However, the concept of universal construction originated with von Neumann, and so the present article is reasonably the root of the general topic.  On the other hand, the notion of a general category is in keeping with recent topical developments: the founding of the field of Constructor Theory.  This new field is the chief result of the workshop Machine Self-Replication, held adjunct to the Tenth International Conference on the Simulation and Synthesis of Living Systems (ALife X).  So, my suggestion is, should such a general category be created, it should be called by the name of this new field (Constructor Theory) with the variants (von Neumann, Codd, ...) being identified as particular examples under this field.  William R. Buckley 20:51, 8 August 2006 (UTC)


 * You might want to check out a discussion I'm hoping to start over at Talk:Self-replicating machine. I think there are several articles that inter-relate that could use a bit of coordinated editing.  Sort of a mini-WikiProject, if you will.  But I'm not sure where to start or how to go about it.  Ideas welcome.  --DragonHawk 00:47, 1 September 2006 (UTC)

Universal Constructors
A link to a musical act, it is not relevant to the topic of Constructors, such as that of von Neumann's dream. That the names are similar does not justify inclusion of this musical act within the external links of the article. I would argue the proper place for this link would be in musical categories. Further, the present article (Universal Constructor) should at most include a link to a separate category, which would serve such context switching as is suggested by the (improper) inclusion of a link to the musical act within the external links of the present article. William R. Buckley 20:28, 8 August 2006 (UTC)

Merging with Article: Self-Replicator

 * Please see discussion at Talk:Self-replicating machine

This should not be done. Indeed, Self-Replicator needs to include a list of various forms. This would allow the various specific examples (to include the von Neumann general solution - Universal Constructor) to be cross-linked. For instance, it should be possible to move from Universal Constructor to Ribosome. The article: Universal Constructor is about a general and abstract solution, and the essential features of a problem. As such, it deserves special treatment. William R. Buckley 04:20, 18 August 2006 (UTC)
 * I actually agree with you. Please see my commentary at Talk:Self-replicating machine.  I should have put a link from this talk page to that one originally -- sorry about that! --DragonHawk 12:43, 18 August 2006 (UTC)

Wikification
I went a little bit janitorial and wikified the article as much as I could without altering the information. Also, the Java Applets link was broken, so I removed it. I was bold and removed the tag as well, since it didn't seem neccesary anymore. I didn't know what exactly to do with the Demonstration image. It wasn't very 'wiki' sitting there in the middle of the page, so I put it to the right of the fesiability and demonstration section. I think it looks wiki now, hope you like it. -- Amenzix 23:20, 31 August 2006 (UTC)

Merging with von Neumann universal constructor
The problem is that a universal constructor is a general concept, and von Neumann's expression of same is a specific example. The two articles are meant to express this relationship. Also, Codd, Langton, Byl, and a good deal of others, including myself, have designed and constructed machines that are capable universal constructors. Yet, they are not von Neumann's universal constructor. There is an hierarchical structure that is needed in order to make the topic intelligible. William R. Buckley 07:41, 12 January 2007 (UTC)