Re: DIS: recursion works here?

2012-03-23 Thread Elliott Hird
On 23 March 2012 21:24, Kerim Aydin  wrote:
> there's no way to act-on-behalf of a ruble

Well, then...


Re: DIS: recursion works here?

2012-03-23 Thread Kerim Aydin



On Fri, 23 Mar 2012, omd wrote:
> Or: "The basis of a Golem is the basis of its owner; if this would
> result in circularity, it has an empty basis and is Emancipated.  A
> non-Emancipated Golem with an empty basis is in Storage."

This also doesn't cover the ruble -> golem case, I don't think?

Looked back at the old basis language but it was really framed 
towards the opposite direction of association (one partnership -> 
many FC players) versus (one FC player -> many golems) so wondered
about not confusing concepts (I know, it's not in use, but still...)





Re: DIS: recursion works here?

2012-03-23 Thread Kerim Aydin


On Fri, 23 Mar 2012, Tanner Swett wrote:
> On Fri, Mar 23, 2012 at 1:13 PM, Kerim Aydin  wrote:
> > Trying to define an obvious concept here, does recursion work, can
> > this be written better?
> >
> >     A golem can either have a Boss, be Emancipated, or be in Storage.
> >
> >     If a golem has no owner, or (through circularity of ownership)
> >     owns itself, it is Emancipated.  Otherwise:
> >     1. If a golem's owner is a first-class person, that person
> >        is the golem's Boss; otherwise,
> >     2. If the golem's owner is an Emancipated golem or a non-golem,
> >        then the golem is in Storage; otherwise,
> >     3. the golem's owner's Boss is the golem's Boss.
> 
> If I may make a suggestion:
> 
> {A golem's first guardian (if any) is its owner. A golem's nth
> guardian (if any), for n greater than 1, is the owner of its (n-1)st
> guardian.
> 
> If a golem has no owner, or has itself as a guardian, it is
> Emancipated. If a golem is not Emancipated, but has an Emancipated
> golem as a guardian, it is in Storage. If a golem has a first-class
> player as a guardian, that player is the golem's Boss.}

I like the phrasing.  One difference:  Golem ownership isn't restricted.  
A golem can be transferred to a ruble, for example. If the ownership 
chain is: 
   F.C. person -> Ruble -> Golem
then my version stops at "Ruble" ("non-golem") and calls the golem in 
Storage.  Your version sees all the way through to "Boss".  But also
if the chain is Lost&Found -> Ruble -> Golem then yours makes it 
undefined.

I suppose desired behavior depends on what I'm using this for.  I think
"In storage" is a little more appropriate because once the golem is
owned by the Ruble, the F.C. person can't get at it (because there's
no way to act-on-behalf of a ruble).

-G.







Re: DIS: recursion works here?

2012-03-23 Thread omd
Or: "The basis of a Golem is the basis of its owner; if this would
result in circularity, it has an empty basis and is Emancipated.  A
non-Emancipated Golem with an empty basis is in Storage."


Re: DIS: recursion works here?

2012-03-23 Thread Pavitra
On 03/23/2012 03:52 PM, Tanner Swett wrote:
> On Fri, Mar 23, 2012 at 1:13 PM, Kerim Aydin  wrote:
>> Trying to define an obvious concept here, does recursion work, can
>> this be written better?
>>
>> A golem can either have a Boss, be Emancipated, or be in Storage.
>>
>> If a golem has no owner, or (through circularity of ownership)
>> owns itself, it is Emancipated.  Otherwise:
>> 1. If a golem's owner is a first-class person, that person
>>is the golem's Boss; otherwise,
>> 2. If the golem's owner is an Emancipated golem or a non-golem,
>>then the golem is in Storage; otherwise,
>> 3. the golem's owner's Boss is the golem's Boss.
> 
> If I may make a suggestion:
> 
> {A golem's first guardian (if any) is its owner. A golem's nth
> guardian (if any), for n greater than 1, is the owner of its (n-1)st
> guardian.
> 
> If a golem has no owner, or has itself as a guardian, it is
> Emancipated. If a golem is not Emancipated, but has an Emancipated
> golem as a guardian, it is in Storage. If a golem has a first-class
> player as a guardian, that player is the golem's Boss.}

Possible shenanigans if a player is ever owned, but that's likely to
cause all sorts of trouble anyway.


Re: DIS: recursion works here?

2012-03-23 Thread Tanner Swett
On Fri, Mar 23, 2012 at 1:13 PM, Kerim Aydin  wrote:
> Trying to define an obvious concept here, does recursion work, can
> this be written better?
>
>     A golem can either have a Boss, be Emancipated, or be in Storage.
>
>     If a golem has no owner, or (through circularity of ownership)
>     owns itself, it is Emancipated.  Otherwise:
>     1. If a golem's owner is a first-class person, that person
>        is the golem's Boss; otherwise,
>     2. If the golem's owner is an Emancipated golem or a non-golem,
>        then the golem is in Storage; otherwise,
>     3. the golem's owner's Boss is the golem's Boss.

If I may make a suggestion:

{A golem's first guardian (if any) is its owner. A golem's nth
guardian (if any), for n greater than 1, is the owner of its (n-1)st
guardian.

If a golem has no owner, or has itself as a guardian, it is
Emancipated. If a golem is not Emancipated, but has an Emancipated
golem as a guardian, it is in Storage. If a golem has a first-class
player as a guardian, that player is the golem's Boss.}

—Tanner "Becquer" L. Swett


Re: DIS: recursion works here?

2012-03-23 Thread Kerim Aydin


On Fri, 23 Mar 2012, Elliott Hird wrote:
> Looks OK to me, except that your self-ownership thing is borked: your
> recursion rules only apply if that /doesn't/, so it will never be
> found to own itself per that process. You'll have to work it into the
> process itself, keeping track of the names seen along the way, like
> manually maintaining a stack.
> 
> Although it might suffice simply to say "If this process would not
> terminate, the golem is instead Emancipated."; the courts can solve
> the halting problem.

I thought about defining the circularity "if following the ownership
chain you get back to yourself..." and then just figured the self-
evident common definition would suffice for the starting condition.
So yes, it's a purposeful mix of common definition followed by procedure.

Interestingly, this exact issue of how Courts deal pretty easily with 
halting problems (or procedural circularity in general) was the subject 
of an exchange of letters between Hofstadter and a law professional 
(republished in Metamagical Themas):

Law professor:  "it's trivial and uninteresting to the court; a judge
can easily see what is meant and determine that there's circularity."

Hofstadter:  "yes, but isn't it interesting that humans can do that
so adeptly!"





Re: DIS: recursion works here?

2012-03-23 Thread Kerim Aydin


On Fri, 23 Mar 2012, Elliott Hird wrote:
> On 23 March 2012 17:58, Pavitra  wrote:
> > I don't see the problem. It seems to me to work correctly.
> 
> Oh, silly me; I thought the procedure also defined a golem's owner,
> due to the circularity remark, but in fact it doesn't. (Right?) So
> this looks fine to me.

Yah, ownership is defined under assets.





Re: DIS: recursion works here?

2012-03-23 Thread Elliott Hird
On 23 March 2012 17:58, Pavitra  wrote:
> I don't see the problem. It seems to me to work correctly.

Oh, silly me; I thought the procedure also defined a golem's owner,
due to the circularity remark, but in fact it doesn't. (Right?) So
this looks fine to me.


Re: DIS: recursion works here?

2012-03-23 Thread Pavitra
On 03/23/2012 12:53 PM, Elliott Hird wrote:
> On 23 March 2012 17:13, Kerim Aydin  wrote:
>> Trying to define an obvious concept here, does recursion work, can
>> this be written better?
> 
> Looks OK to me, except that your self-ownership thing is borked: your
> recursion rules only apply if that /doesn't/, so it will never be
> found to own itself per that process. You'll have to work it into the
> process itself, keeping track of the names seen along the way, like
> manually maintaining a stack.

I don't see the problem. It seems to me to work correctly.

> The courts can solve the halting problem.

If I were the sort of person to have an email signature, I would put
this in it.


Re: DIS: recursion works here?

2012-03-23 Thread Elliott Hird
On 23 March 2012 17:13, Kerim Aydin  wrote:
> Trying to define an obvious concept here, does recursion work, can
> this be written better?

Looks OK to me, except that your self-ownership thing is borked: your
recursion rules only apply if that /doesn't/, so it will never be
found to own itself per that process. You'll have to work it into the
process itself, keeping track of the names seen along the way, like
manually maintaining a stack.

Although it might suffice simply to say "If this process would not
terminate, the golem is instead Emancipated."; the courts can solve
the halting problem.