[Sociam-soton] Re: Is email a social machine?

Tue Oct 29 11:52:50 GMT 2013

That all seems fine in general to me, but I’m not quite sure about the disanalogy.
I can see it more as an analogy of how we might study SMs.

As best I recall (my memory is being challenged today!) and can see from the 1936 paper, Turing had no HALT states explicitly in his a-machine.
He writes about machines that reach a configuration from which there is no possible move, which suggest that the machine description does not contain the HALT states, and in fact I am guessing that they could be innumerable.
It is people like Hopcroft and Ullman who introduce a set of HALT states.
But never mind that, the existence of a set of HALT states in the formal definition of a TM is there precisely (I think) because the purpose is to study halting.
It is not an intrinsic part of the execution mechanism.
In the same way, if we wanted to study SMs and their halting properties (which seems to be what we are doing now!), we have to introduce the concept of being able to recognise when they are finished, so that we can talk about it.
That doesn’t mean that SMs have to ever finish, any more than TMs have to ever reach their halting condition.
And the final point is that the set of HALT states for a TM is just that - a set, and therefore could be the empty set, implying that the machine never halts.

Wot larks!

On 29 Oct 2013, at 11:24, Kieron O'Hara <kmo at ecs.soton.ac.uk> wrote:

> I think a halting condition would be at best sufficient but certainly not necessary for social machineness.
> 
> So, halting is a defined, if undecidable, property of Turing machines. But then we need to consider the disanalogies between Turing machines and social machines.
> 
> One disanalogy, and to my mind (one of) the most important, is that a TM's definition includes a set of final states which we can take as collectively defining the goal of the computation. In that context, it makes sense to discuss halting - can we show that from an arbitrary state the TM will eventually achieve one of those final states or not? (even though we know that that question is undecidable for an arbitrary TM).
> 
> On the other hand, with a social machine, although it is goal-directed, it seems a bit of a stretch to say that it is defined *by* its goal, because its goal(s):
> (a) may be different at different levels of the system (e.g. recaptcha, where the designer has one goal for the system, and the users ('components' of the system) want the system to do something completely different. Each goal has to be achieved for the machine to continue functioning), and
> (b) may change in use, when the 'components' of the system may simply decide to participate in the social machine for a different reason, and
> (c) need not be such as to bring the machine to a halt. Achieving the goal of a good conversation, or some interesting game-play, for instance, is actually a reason to carry on with the machine rather than halt it. The halting condition of a mailing list is that it *hasn't* achieved its goal. The halting condition of Wikipedia is that it has completely failed to tell us anything interesting about the world. Having failed to achieve their goals, people will stop using them.
> 
> So for a social machine, even if it had a set of final states at a time t0, there is no guarantee that the same final states would be in place at a later time t1. Whereas a Turing machine's set of final states is an invariant.
> 
> Kieron
> 
> 
> On 29/10/2013 10:54, Leslie Carr wrote:
>> Does a mailing list have a halting condition? Can it ever finish its task?
>> 
>> Sent from my iPhone
>> 
>> On 29 Oct 2013, at 10:50, "rt506" <rt506 at ecs.soton.ac.uk<mailto:rt506 at ecs.soton.ac.uk>> wrote:
>> 
>> I agrees with Les, however, if we talk about having an agenda/purpose, would a mailing list with a specific purpose be a social machine where email is one if the tools within the social machinery network?
>> 
>> Ramine
>> 
>> 
>> Sent from Samsung Galaxy Note
>> 
>> 
>> 
>> -------- Original message --------
>> From: Hugh Glaser <hg at ecs.soton.ac.uk<mailto:hg at ecs.soton.ac.uk>>
>> Date: 29/10/2013 10:35 (GMT+00:00)
>> To: Kieron O'Hara <kmo at ecs.soton.ac.uk<mailto:kmo at ecs.soton.ac.uk>>
>> Cc: sociam-soton at ecs.soton.ac.uk<mailto:sociam-soton at ecs.soton.ac.uk>
>> Subject: [Sociam-soton] Re: Is email a social machine?
>> 
>> 
>> Agreed.
>> Is a crowbar a machine?
>> Is steam power a machine?
>> They might be loosely referred to as such in use, but to people studying statics and dynamics and dynamics they aren't.
>> 
>> Hugh
>> 023 8061 5652
>> 
>> On 29 Oct 2013, at 10:27, "Kieron O'Hara" <kmo at ecs.soton.ac.uk<mailto:kmo at ecs.soton.ac.uk>> wrote:
>> 
>>> Hi Wendy,
>>> 
>>> I'm definitely with your initial reaction. We define a machine in
>>> general as something that does work to achieve a goal. A Turing machine
>>> has a set of final states. I don't see email as having anything like
>>> that kind of teleological goal-directedness, except at a very abstract
>>> level like it allows people to communicate with people.
>>> 
>>> To me it's much more like a component tool or platform. The point of
>>> social machines, it seems to me, is that whatever else they are, they
>>> are characterised by people getting together to do something. Email is
>>> something you would use for that in many circumstances, but it is not
>>> itself characterised by that. It's on the level of a mobile phone
>>> network, a railway system, or a social networking site, in my view. Once
>>> it starts to be used to achieve some kind of goal, then it becomes a
>>> component of a social machine.
>>> 
>>> Is my view.
>>> 
>>> Kieron
>>> 
>>> 
>>> On 29/10/2013 10:10, Wendy Hall wrote:
>>>> Hi
>>>> 
>>>> I'm at dinner in Melbourne and the question has arisen - is email a social machine?
>>>> 
>>>> My initial reaction was no, because email is not of the Web. But there are people here who are persuading me otherwise
>>>> 
>>>> Thoughts?
>>>> 
>>>> Wendy
>>>> 
>>>> PS Apologies if this question is already answered in the SM classification work but I don't have all the information to hand
>>>> 
>>>> Sent from my iPhone
>>>> _______________________________________________
>>>> Sociam-soton mailing list
>>>> Sociam-soton at ecs.soton.ac.uk<mailto:Sociam-soton at ecs.soton.ac.uk>
>>>> http://mailman.ecs.soton.ac.uk/mailman/listinfo/sociam-soton
>>> _______________________________________________
>>> Sociam-soton mailing list
>>> Sociam-soton at ecs.soton.ac.uk<mailto:Sociam-soton at ecs.soton.ac.uk>
>>> http://mailman.ecs.soton.ac.uk/mailman/listinfo/sociam-soton
>> _______________________________________________
>> Sociam-soton mailing list
>> Sociam-soton at ecs.soton.ac.uk<mailto:Sociam-soton at ecs.soton.ac.uk>
>> http://mailman.ecs.soton.ac.uk/mailman/listinfo/sociam-soton
>> _______________________________________________
>> Sociam-soton mailing list
>> Sociam-soton at ecs.soton.ac.uk<mailto:Sociam-soton at ecs.soton.ac.uk>
>> http://mailman.ecs.soton.ac.uk/mailman/listinfo/sociam-soton
> 

--
Hugh
023 8061 5652