todays call notes: note new term proposal "identifier"

[Chris Stoeckert]

Attending: CS, JF, MC

1. Jennifers use case:
Jennifer has posted her use case as a google docs and will make a
first pass at annotating to generate specific questions for the branch.

2. Went through tracker items
a. software: this was discussed and summarized by Bjoern under the
"programming" thread. Action item is for all to review discussion and
adjust ontology as needed.
b. image/ graph terms: CS will check with Liju regarding status of
is_rendered_by. Also there was further discussion on the distinction
between timage and report graph. Action item was to apply use cases to
clarify whether the distinction was necessary and if so help improve
definitions. CS will do microarray, MC will do flow cytometry. JF will
do another.
c. information entities model number, serial number: Agreed that a
term to capture these is needed and currently missing. Propose a term
"identifier" which the DENRIE branch should agree on and include in OBI.

Term: identifier
Definition: An identifier is an information content entity which can
take the form of label, symbol or token that names or identifies an
entity or a collection of data.
Definition source: http://www.pcmag.com/encyclopedia_term/0,2542,t=identifier&i=44716,00.asp
Examples: version, lot, name, ID, serial number

d. genome version number: would be covered under "identifier" proposal
e. objective (several). Still working out upper ontology regarding
realizable information entities (whether objectives are realizable).
In the mean time can continue to add to children of objective.

Thanks,
Chris

[Alan Ruttenberg]

On May 12, 2008, at 12:22 PM, Chris Stoeckert wrote:

> c. information entities model number, serial number: Agreed that a
> term to capture these is needed and currently missing. Propose a
> term "identifier" which the DENRIE branch should agree on and
> include in OBI.
>
> Term: identifier
> Definition: An identifier is an information content entity which
> can take the form of label, symbol or token that names or
> identifies an entity or a collection of data.
> Definition source: http://www.pcmag.com/encyclopedia_term/
> 0,2542,t=identifier&i=44716,00.asp
> Examples: version, lot, name, ID, serial number

Issues - not all symbols are identifiers - multiple inheritance
issue? Don't know what token is. Wondering whether we should avoid
general "identifier" category and instead focus on the specifics. We
may or may not want to distinguish between the information content
and the encoding. For instance a serial number can be encoded in a
bar code or as digits.

Serial number: An information content entity that is a sequence of
characters borne by part of manufactured product or its packaging
that is intended to be unique to each individual in some class of
products, and so can serve as a way to identify an individual product
within the class. Serial numbers can be encoded in a variety of other
information objects, such as bar codes, numerals, or patterns of dots.

Model number: An information content entity that is an identical
sequence of characters borne by part of each of a class of of
manufactured products or its packaging that is intended to be
unique to a class of products, and can so serve as a way to identify
a specific product class within the set that the manufacture makes.
Model numbers can be encoded in a variety of other information
objects, such as bar codes, numerals, or patterns of dots.

Lot number: An information content entity that is an identical
sequence of characters borne by part of manufactured product or its
packaging for each instances of a product class are manufactured in a
lot. Manufacturing as a lot might be due to a variety of reasons, for
example, a single process during which many individuals are made
from the same portion of source material. Lot numbers can be encoded
in a variety of other information objects, such as bar codes,
numerals, or patterns of dots.

Version number - running out of steam at the moment, but note that it
doesn't need to be a number, and that we probably want to distinguish
version number of software (something like revision number) version
number of database (something that marks points in the accumulation
of sets of information) version number of an information entity which
is computed (e.g. genome build number), which refers to the process
as well as the result, and version number applied to a product (which
either applies to the design, or is really a model number).

> d. genome version number: would be covered under "identifier" proposal

Properly, genome representation build number, identifying both the
iteration of the process used to compute the representation, and the
information content entity that results.

-Alan

[Melanie Courtot]

Hi,
The idea was to not define the specifics yet, and instead have a general "identifier" category , put the specifics as example of usage, and once we would be happy with identifier add children as needed.

I agree that all symbols are not identifiers, but I think all identifiers are made of symbols (symbol being any of letters, digits, icons...) In that case we should be ok vs multiple inheritance?
Would the definition "an identifier is an information content entity which names or identifies an entity or collection of data" be better?

I don't think we were trying to address encoding (bar code or digit) but we were rather trying to express the information content entity.

For version number I take it as representing the current state of something, this something being a database, software... (the first option you describe, which covers the iteration of the process as well as the result) I didn't get the difference with version number of a product? I would say if new version number something has been changed, so a process has been accomplished.

Thanks for the definitions, I'm wondering about the why of the packaging precision? (I understand packaging as the box you discard once opened) Is it to cover cases where the model number for example is written on the box and not the instrument? Even in this case, would it be right to consider that the model number applies to the instrument?

Thanks,
Melanie

[Alan Ruttenberg]

On May 13, 2008, at 12:27 AM, Melanie Courtot wrote:

Hi,
The idea was to not define the specifics yet, and instead have a general "identifier" category , put the specifics as example of usage, and once we would be happy with identifier add children as needed.

Except I'm a bit worried about identifier as a class, hence the emphasis on the more specific classes. 

I agree that all symbols are not identifiers, but I think all identifiers are made of symbols (symbol being any of letters, digits, icons...) In that case we should be ok vs multiple inheritance?

What else could be used as an identifier? Gene sequence?

Would the definition "an identifier is an information content entity which names or identifies an entity or collection of data" be better?

Then a URI is an identifier, as is any term in the ontology?

I don't think we were trying to address encoding (bar code or digit) but we were rather trying to express the information content entity.

I wasn't sure, but thought it worth clarifying.

For version number I take it as representing the current state of something, this something being a database, software... (the first option you describe, which covers the iteration of the process as well as the result) I didn't get the difference with version number of a product?

Can a version number be a model number?

I would say if new version number something has been changed, so a process has been accomplished.

That corresponds to my understanding. But a but more. There is an implication that it is either the same or similar to the previous thing in some way.

Thanks for the definitions, I'm wondering about the why of the packaging precision? (I understand packaging as the box you discard once opened) Is it to cover cases where the model number for example is written on the box and not the instrument?

Yes. As dependent continuant it needs to be borne by something. So the definition should say what.

Even in this case, would it be right to consider that the model number applies to the instrument?

We've been using "is about". So it would be about the instrument, but it might be borne by something else.

-Alan

[Melanie Courtot]

Hi,

I most probably won't be able to make the denrie call on monday, but
here are my thoughts on identifier.

- I understand that it is a very general term, and yes I would say
that anything could be an identifier, a URI, a gene sequence etc.
Everything that you use to name a thing, uniquely or not, is for me an
identifier.
- I'm ok if you decide not to have identifier at the end, but I still
believe that all our specific terms share something common, and that
it would be nice to have a common parent to cover general cases
- regarding version number: I also agree that you keep same version
number only if similarity in some way to a previous thing. Even if you
modify completely the code behind your software, you would call it new
version if for example it is still doing the same thing.
- "Can a version number be a model number?"
From what you said the version number of a product could be a model
number, but I can't think of an example (maybe because I'm thinking
software when I hear version)
- thanks for the packaging explanation - I thought it was ok to omit
the packaging and directly say that the instrument has number x.

What about moving forward on the thing we all agree upon, i.e. the
specifics you described, and try to have a clean definition for those
(and I think we are close with the ones you proposed)? We know that we
will need to have them being children of a general identifier or not.
Then we can see later on (probably with use cases/requests from users)
if we need a general identifier or if those are enough.

Thanks,
Melanie

[Chris Stoeckert]

I too will not be able to make the call on Monday as I will be away at
a retreat.

I also agree with Melanie regarding trying to move ahead with creating
a set of terms and definitions that specify the different labels we
use to distinguish between different instances of genome builds,
different instances of the same instrument (serial number). If
"identifier" has too loose a connotation we can try to come up a
different label but feel we should use it for now and work to improve
it and the definition. The definition was quickly grabbed off the web
and sounded right but but the term and definition were meant as
starting points for a general type of requested terms.

Cheers,
Chris

[Barry Smith]

>Serial number: An information content entity that is a sequence of
>characters

[we need to define 'sequence of characters' elsewhere
the sequence of characters should be part of series of sequences of
characters (which serve in some ways like a clock)]

>borne by part of manufactured product or its packaging
>that is intended to be unique to each individual in some class of
>products

and to serve as a way to identify an individual product
and which reliably does so

>within the class. Serial numbers can be encoded in a variety of other
>information objects, such as bar codes, numerals, or patterns of dots.

BS

[Melanie Courtot]

Hi all,

Re-reading this, I am wondering if we shouldn't have "sequence of
characters" which has objective "identification" instead of the class
"identifier".
We could have Information Content Entity, with child sequence of
characters, with children like serial number, model number etc...and
these children would have the objective "identification".

I am not sure having sequence of character would be better than
identifier, because it is pretty general too, but I think we should
consider the objective (and also that we would anyway need an
identification objective, e.g. for biomaterial identification)

What do you think?
Melanie


On May 19, 4:15 pm, Barry Smith <phism...@buffalo.edu> wrote:
> >Serial number: Aninformationcontententitythat is a sequence of

> >characters
>
> [we need to define 'sequence of characters' elsewhere
> the sequence of characters should be part of series of sequences of
> characters (which serve in some ways like a clock)]
>
> >borne by part of manufactured product or its packaging
> >that is intended to be unique to each individual in some class of
> >products
>
> and to serve as a way to identify an individual product
> and which reliably does so
>
> >within the class. Serial numbers can be encoded in a variety of other

> >informationobjects, such as bar codes, numerals, or patterns of dots.
>
> BS