Re: FW: Major change to a SIOC term: User is now called UserAccount

I am not motivated by philosophical issues.

It depends on how you want to use something.  Conflating people and roles
and accounts may not matter in a localized situation.  It seems more likely
to sting you in a worldwide context of using linked data.  It is limiting in
the way you suggest at the end of the message.  Making distinctions cleanly
(which by accident may please a philosopher), makes it easier and cheaper
for a system to evolve and grow.

I know second hand of cases where silly mistakes like conflating role with
account cost $millions. Why?  The system could not evolve to the future
needs.

The fact that SIOC:User was always a subclass of foaf:OnlineAccount is
probably a good thing.

I have not carefully studied SIOC, so will not otherwise comment on it.

Michael

On Tue, Mar 30, 2010 at 1:40 PM, Simon Reinhardt
<simon.reinhardt@koeln.de>wrote:

> Hello Michael
>
>
> Michael F Uschold wrote:
>
>> This has a knockon effect, which could create new confusion.
>>
>> Now the creator of a message is a UserAccount.  But this is not sensible.
>> Accounts do not create messages, User's create messages. A User has a
>> relationship with an account, e.g. AccountHolder.
>>
>> There may be confusion between DC-Creator whose domain presumably does not
>> include Accounts.
>>
>
> I don't see a problem here. An account is just an entity of managing
> things, a sort of proxy through which actions like creations are seen. This
> works perfectly fine in the computer world, maybe not so much if you get
> more conceptual/philosophical. But I don't see a need to do that. You don't
> always know the person behind an account and it might even be shared by
> multiple persons.
>
> Also note that this was a change in naming to clarify the intended meaning
> of the term. sioc:User has always been a sub-class of foaf:OnlineAccount and
> thus the meaning wasn't changed. It is sort of one of the basic assumptions
> of the SIOC model that it's perfectly ok for accounts to work as
> placeholders for actions and properties within community platforms. This is
> how those platforms are modelled anyway so they can be nicely described with
> SIOC rather than having to introduce extra indirection just for some
> philosophical reasons.
>
> Another way to look at this is to see accounts as roles you can take on
> which are detached from your person. Or a bit more practical: imagine you
> have multiple accounts on a platform (one for acting as a user, one for
> administrating). Isn't it rather important under which account you did
> something?
> In online games people sometimes create multiple accounts to gain
> advantages - how would you describe that you transferred game money from one
> of your accounts to another?
> Or how would you model that various people use one administrator account?
> Who "created" that new user account? Surely whoever was logged in as the
> administrator at that point in time. But even if you knew who it was - if
> you attached that information to the person you loose information about
> which account was responsible for it.
>
> I wouldn't necessarily say that dcterms:creator excludes that scenario. Its
> range is defined as dcterms:Agent which has the following definition:
>  Definition:    A resource that acts or has the power to act.
>  Comment:       Examples of Agent include person, organization, and
> software agent.
>
> Regards,
>  Simon
>

Received on Wednesday, 31 March 2010 17:06:43 UTC