information artifact class

[Bjoern Peters]

All,

We are about to have OBI ready for the 1.0 release. In that spirit, the
IAO terms that are subclassed by OBI terms need to be in presentable
shape as well. In that spirit, I want to re-raise a number of older
issues that I think we should and can resolve quickly enough.

* information artifact - To clarify the scope and exclude music / bee
dances etc. from IAO, Barry / Alan had originally proposed to limit it
to 'information artifacts', which meet a number of criteria. To make
this clear to the ontology user, and avoid discussions on where music
etc. goes in the ontology, can we have information artifact as the top
level class of IAO under generically dependent continuant? We don't yet
have to have a perfect definition, maybe something like this (shortened
from Barry's original email):

An information artifact is a GDCs that bears information about
something. It is communicatable between sentient beings and certain
machines, can be preserved over time, can be replicated with no (or very
little) loss, and is formulated in a background language.

Currently there are three top level terms in IAO:
* _identifier
* digital quality (no definition)
* information content entity

It is unclear to me if / how the 'information content entity' is
different from the general definition of information artifacts.

- Bjoern

--
Bjoern Peters
Assistant Member
La Jolla Institute for Allergy and Immunology
9420 Athena Circle
La Jolla, CA 92037, USA
Tel: 858/752-6914
Fax: 858/752-6987
http://www.liai.org/pages/faculty-peters

[Alan Ruttenberg]

Currently I believe what you mean to be information artifact is
information content entity. The other siblings represent work that is
not completed and I expect them to not be there as we further edit the
ontology.

To others, note that I have done some recent edits. I can summarize if
someone wants - they are mostly in response to OBI needs. There is
still much to do.

A current issue concerns the question of what we need (or want, at
this stage of development) terms for hypothesis, proposition,
conclusion and similar. There is a narrow sense, namely as kinds of
sections of reports, that can be defined, and there is a broader sense
which is complicated because we enter the realm of self references -
axioms in the ontology itself become subject to description by the
ontology.

I'd rather avoid that if at all possible.

-Alan

[Bjoern Peters]

Alan Ruttenberg wrote:
> Currently I believe what you mean to be information artifact is
> information content entity. The other siblings represent work that is
> not completed and I expect them to not be there as we further edit the
> ontology.
>

Good. Can we label it information-artifact entity then? Can the
definition or at least editor notes state the conclusion of the email
discussions limiting it to not include bee dances / songs and DNA
molecule sequences (as occuring in nature not databases?). Can the
siblings be placed under information artifact?

- Bjoern

[Albert Goldfain]

On Fri, Jan 23, 2009 at 2:02 PM, Bjoern Peters <bpe...@liai.org> wrote:

An information artifact is a GDCs that bears information about
something. It is communicatable between sentient beings and certain
machines, can be preserved over time, can be replicated with no (or very
little) loss, and is formulated in a background language.

In general I think this is an acceptable definition.  I was just wondering if there has been any discussion of collapsing the somewhat awkward conjunction "sentient beings and certain machines" into something like "cognitive agents suitably trained for processing such information" (which would have sentient beings and computers as subtypes). 

Also, I am wondering whether the "preservation over time" is referring to physical preservation (endurance of the artifact) or what might be called semantic preservation (stability of the information-content/meaning over time).  I can preserve a binary on my Linux system for decades, but will it be interpreted the same way after those decades?  It should be made explicit which kind of preservation is meant.

-Albert

[Barry Smith]

At 12:25 PM 1/26/2009, Albert Goldfain wrote:


>On Fri, Jan 23, 2009 at 2:02 PM, Bjoern Peters
><<mailto:bpe...@liai.org>bpe...@liai.org> wrote:
>
>An information artifact is a GDCs that bears information about
>something. It is communicatable between sentient beings and certain
>machines, can be preserved over time, can be replicated with no (or very
>little) loss, and is formulated in a background language.

surely GDC rather than GDCs?

'communicable' rather than 'communicable'

do we really need to use 'information' (the mass noun); can't we just
say 'that is about something' making sure that we define 'about' and
'communicable' in such a way that it is clear that only information
entities can be about something and only their content can be communicated?

am not quite happy with the 'or very little'; can't we get at this by
stating that they can always be replicated in such a way that their
information content is perfectly preserved without loss? (thus may
not be quite true for the sake of an old scrap of paper with a diary
entry whose date we cannot determine)

[Bjoern Peters]

Incorporating Barry's feedback, a revised definition would read:

An information artifact is a GDC that is about something.
It is communicable between sentient beings and certain
machines, can be preserved over time, can be replicated without loss,

and is formulated in a background language.

I agree with Albert's first comment, and would add an editor note:

Editor note: "we may want to create a union of "sentient beings and
certain machines"
and name it something like "cognitive agents capable of processing
information"

Regarding the second comment, as GDCs are not material entities, it
seems clear to me that this is not about physical preservation, and I
don't think we need to capture that.

- Bjoern

Barry Smith wrote:
> At 12:25 PM 1/26/2009, Albert Goldfain wrote:
>
>
>
>> On Fri, Jan 23, 2009 at 2:02 PM, Bjoern Peters
>> <<mailto:bpe...@liai.org>bpe...@liai.org> wrote:
>>
>> An information artifact is a GDCs that bears information about
>> something. It is communicatable between sentient beings and certain
>> machines, can be preserved over time, can be replicated with no (or very
>> little) loss, and is formulated in a background language.
>>
>
> surely GDC rather than GDCs?
>
> 'communicable' rather than 'communicable'
>
> do we really need to use 'information' (the mass noun); can't we just
> say 'that is about something' making sure that we define 'about' and
> 'communicable' in such a way that it is clear that only information
> entities can be about something and only their content can be communicated?
>
> am not quite happy with the 'or very little'; can't we get at this by
> stating that they can always be replicated in such a way that their
> information content is perfectly preserved without loss? (thus may
> not be quite true for the sake of an old scrap of paper with a diary
> entry whose date we cannot determine)
>
>
>

-Albert






>
> >
>

[da...@georgetown.edu]

Why bother with "sentient beings..." at all? Just cut the phrase at "communicable" and the definition still works. Wouldn't it be understood that communication only occurs between those entities capable of it?

----