Bug 329 - ESA - Definition of Persistent and Unique Identifiers
Summary: ESA - Definition of Persistent and Unique Identifiers
Status: RESOLVED CHANGE AGREED
Alias: None
Product: OAIS June 2012
Classification: Unclassified
Component: Section 1: Introduction (show other bugs)
Version: June 2012
Hardware: No All
: Needs some discussion Editorial
Assignee: David Giaretta (david@giaretta.org)
URL:
Whiteboard:
Depends on:
Blocks: 332 382
  Show dependency tree
 
Reported: 2021-02-15 13:01 UTC by David Giaretta (david@giaretta.org)
Modified: 2024-02-27 18:57 UTC (History)
4 users (show)

See Also:
Organisation of the submitter: ESA (Iolanda Maggio, submitted by David Giaretta)
Disposition of the suggested change:
Category of the suggested change: ---
Due date:
Explanation of the reason for the suggested change:
Missing definitions


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description David Giaretta (david@giaretta.org) 2021-02-15 13:01:44 UTC
Persistent Identifier and Unique Identifier missing definitions
Comment 1 David Giaretta (david@giaretta.org) 2021-02-23 12:15:48 UTC
OAIS uses the terms (once each)
- Persistent Identifiers  (as an example of Reference Information in Table 4-1)
- Unique Persistent Identifiers ("Unique Persistent Identifiers within some specified naming system" in section 6.2.4)

These are capitalised.

However "unique identifier" is used several times, but never capitalised.

How about:
- Identifier: An identifier is a name that identifies (that is, labels the identity of) either a unique object or a unique class of objects, where the "object" or class may be an idea, physical countable object (or class thereof), or physical noncountable substance (or class thereof). 
- Persistent Identifier: A long-lasting Identifier.
- Unique Persistent Identifier: A Persistent Identifier which is unique within a specific naming convention.
- Unique Identifier: An Identifier which unique within a specific naming convention (but may not be persistent)
Comment 2 Eld Zierau (elzi@kb.dk) 2021-02-24 06:41:27 UTC
I agree that we should use these terms with care, and we should be careful with the way we use persistent, since persistent identifier can be misunderstood as an identifier managed by a persistent identifier system like DOI.

Maybe the places should be revisited and re-evaluate what is meant for each of the terms, - then choose which to define and which to explain with examples.
Comment 3 David Giaretta (david@giaretta.org) 2021-03-30 10:52:19 UTC
(In reply to David Giaretta from comment #1)
> OAIS uses the terms (once each)
> - Persistent Identifiers  (as an example of Reference Information in Table
> 4-1)
> - Unique Persistent Identifiers ("Unique Persistent Identifiers within some
> specified naming system" in section 6.2.4)
> 
> These are capitalised.
> 
> However "unique identifier" is used several times, but never capitalised.
> 
> How about:
> - Identifier: An identifier is a name that identifies (that is, labels the
> identity of) either a unique object or a unique class of objects, where the
> "object" or class may be an idea, physical countable object (or class
> thereof), or physical noncountable substance (or class thereof). 
> - Persistent Identifier: A long-lasting Identifier.
> - Unique Persistent Identifier: A Persistent Identifier which is unique
> within a specific naming convention.
> - Unique Identifier: An Identifier which unique within a specific naming
> convention (but may not be persistent)

COrrection - add "is" in
- Unique Identifier: An Identifier which is unique within a specific naming convention (but may not be persistent)
Comment 4 David Giaretta (david@giaretta.org) 2022-09-21 13:59:11 UTC
Changes made:

- Identifier: An identifier is a name that identifies (that is, labels the identity of) either a unique object or a unique class of objects, where the "object" or class may be an idea, physical countable object (or class thereof), or physical noncountable substance (or class thereof). 
- Persistent Identifier: A long-lasting Identifier.
- Unique Persistent Identifier: A Persistent Identifier which is unique within a specific naming convention.
- Unique Identifier: An Identifier which is unique within a specific naming convention (but may not be persistent)
Comment 5 David Giaretta (david@giaretta.org) 2023-09-28 11:59:03 UTC
REOPENED for ISO review - had been accepted after CESG changes made so not in ISO submission.
Comment 6 Gerald Banon (gerald.banon@gmail.com) 2023-12-11 14:55:12 UTC
(In reply to David Giaretta from comment #4)
> Changes made:
> 
> - Identifier: An identifier is a name that identifies (that is, labels the
> identity of) either a unique object or a unique class of objects, where the
> "object" or class may be an idea, physical countable object (or class
> thereof), or physical noncountable substance (or class thereof). 
> - Persistent Identifier: A long-lasting Identifier.
> - Unique Persistent Identifier: A Persistent Identifier which is unique
> within a specific naming convention.
> - Unique Identifier: An Identifier which is unique within a specific naming
> convention (but may not be persistent)

Regarding some of the above definitions, I would like to suggest, for the future, the following more precise statements:

- Namespace: A Namespace is a set of names.

- Identifier:  An Identifier (of an object) is a name within a specific namespace that is assigned to that object.

- Persistent Identifier: A Persistent Identifier (of an object) is an Identifier (of an object) whose assignment to that object is done forever.

- Unique Persistent Identifier: The Unique Persistent Identifier (of an object) is a Persistent Identifier (of an object) whose assignment is made only to that object.
Comment 7 David Giaretta (david@giaretta.org) 2024-02-13 14:49:32 UTC
Combining Gerald's suggestions with what we had agreed previously:

- Namespace: A Namespace is a set of names.
- Identifier: An identifier is a name within a specific namespace that identifies (that is, labels the identity of) either a unique object or a unique class of objects, where the "object" or class may be an idea, physical countable object (or class thereof), or physical noncountable substance (or class thereof), and ideally allows the object(s) to be located. 
- Persistent Identifier: A long-lasting Identifier, ideally lasting forever.
- Unique Identifier: An Identifier which is unique within a specific naming convention (but may not be persistent)
- Unique Persistent Identifier: A Persistent Identifier which is unique within a specific naming convention.
Comment 8 Gerald Banon (gerald.banon@gmail.com) 2024-02-14 22:33:25 UTC
(In reply to David Giaretta from comment #7)
> Combining Gerald's suggestions with what we had agreed previously:
> 
> - Namespace: A Namespace is a set of names.
> - Identifier: An identifier is a name within a specific namespace that
> identifies (that is, labels the identity of) either a unique object or a
> unique class of objects, where the "object" or class may be an idea,
> physical countable object (or class thereof), or physical noncountable
> substance (or class thereof), and ideally allows the object(s) to be
> located. 
> - Persistent Identifier: A long-lasting Identifier, ideally lasting forever.
> - Unique Identifier: An Identifier which is unique within a specific naming
> convention (but may not be persistent)
> - Unique Persistent Identifier: A Persistent Identifier which is unique
> within a specific naming convention.

I would like to make some comments about the expression 'Unique Identifier'
There are two things that can be said:
1. The expression 'Unique Identifier' is redundant.
An identifier is a name within a specific namespace that identifies (that is, labels the identity of) either a unique object...
If the identifier of something identifies it, the identifier is necessarily unique.

2. The expression 'Unique Identifier' is ambiguous.
The definition:
'Unique Identifier: An Identifier which is unique within a specific naming convention'
might receive two interpretations:
2.1 A unique object is identified by only one name within a specific namespace;
2.2 The identifier within a specific namespace identifies only one object.

I think that what was intended by 'Unique Identifier' is 2.2. Am I right?
If so, the definition of Unique Identifier should be: An Identifier within a specific namespace that identifies only one object. Nevertheless, the expression 'identifies only one object' sounds redundant (see Item 1. above).

In my opinion, the definition of Identifier:
'An identifier is a name within a specific namespace that IDENTIFIES (that is, labels the identity of) either a unique object...' already contains the idea of unicity. For this reason, the definition of 'Unique Identifier' is probably not necessary.
Comment 9 David Giaretta (david@giaretta.org) 2024-02-19 20:03:11 UTC
(In reply to Gerald Banon from comment #8)
> (In reply to David Giaretta from comment #7)
> > Combining Gerald's suggestions with what we had agreed previously:
> > 
> > - Namespace: A Namespace is a set of names.
> > - Identifier: An identifier is a name within a specific namespace that
> > identifies (that is, labels the identity of) either a unique object or a
> > unique class of objects, where the "object" or class may be an idea,
> > physical countable object (or class thereof), or physical noncountable
> > substance (or class thereof), and ideally allows the object(s) to be
> > located. 
> > - Persistent Identifier: A long-lasting Identifier, ideally lasting
> forever.
> > - Unique Identifier: An Identifier which is unique within a specific naming
> > convention (but may not be persistent)
> > - Unique Persistent Identifier: A Persistent Identifier which is unique
> > within a specific naming convention.
> 
> I would like to make some comments about the expression 'Unique Identifier'
> There are two things that can be said:
> 1. The expression 'Unique Identifier' is redundant.
> An identifier is a name within a specific namespace that identifies (that
> is, labels the identity of) either a unique object...
> If the identifier of something identifies it, the identifier is necessarily
> unique.
> 
> 2. The expression 'Unique Identifier' is ambiguous.
> The definition:
> 'Unique Identifier: An Identifier which is unique within a specific naming
> convention'
> might receive two interpretations:
> 2.1 A unique object is identified by only one name within a specific
> namespace;
> 2.2 The identifier within a specific namespace identifies only one object.
> 
> I think that what was intended by 'Unique Identifier' is 2.2. Am I right?
> If so, the definition of Unique Identifier should be: An Identifier within a
> specific namespace that identifies only one object. Nevertheless, the
> expression 'identifies only one object' sounds redundant (see Item 1. above).
> 
> In my opinion, the definition of Identifier:
> 'An identifier is a name within a specific namespace that IDENTIFIES (that
> is, labels the identity of) either a unique object...' already contains the
> idea of unicity. For this reason, the definition of 'Unique Identifier' is
> probably not necessary.

This raises the question in my mind as to whether there are identifiers which are not unique.
For example we use the example of an ISBN, but an ISBN does not identify a specific book, but rather a set of probably identical books. This is essentially what we mean by "unique class of objects". 

So does a "unique id" refer to a unique object i.e.a unique book.

For example in a repository we may have multiple copies of a digital object and we don't care which one we get when we ask for a copy. On the other hand we do need to locate specific copies at various times.

That argument seems to lead to the conclusion that we do need the term "unique identifier"
Comment 10 David Giaretta (david@giaretta.org) 2024-02-19 20:21:57 UTC
Another thought is that we talk about "namespaces" but what we actually need is that an identifier has to always be resolvable through a resolution system. Otherwise the identifier is just a string of characters.

SO perhaps we need something like
From:
- Identifier: An identifier is a name that identifies (that is, labels the identity of) either a unique object or a unique class of objects, where the "object" or class may be an idea, physical countable object (or class thereof), or physical noncountable substance (or class thereof). 

To:
- Identifier: An identifier is a name and an associated resolution system that identifies (that is, labels the identity of) either a unique object or a unique class of objects, where the "object" or class may be an idea, physical countable object (or class thereof), or physical noncountable substance (or class thereof).

The APARSEN project defined:
 Identifier_Management_Scheme

Definition: 	A system that deals with identifying entities in a system by using identifiers. In the system IDs are used only as a way to make unambiguous reference to an entity and not as tokens to access to the system (this allows to distinguish ID management systems from authentication services).
Comment 11 David Giaretta (david@giaretta.org) 2024-02-19 20:57:54 UTC
The page http://www.alliancepermanentaccess.org/index.php/consultancy/dpglossary/#Identifier might be useful.
Comment 12 Gerald Banon (gerald.banon@gmail.com) 2024-02-20 12:20:47 UTC
(In reply to David Giaretta from comment #9)
> (In reply to Gerald Banon from comment #8)
> > (In reply to David Giaretta from comment #7)
> > > Combining Gerald's suggestions with what we had agreed previously:
> > > 
> > > - Namespace: A Namespace is a set of names.
> > > - Identifier: An identifier is a name within a specific namespace that
> > > identifies (that is, labels the identity of) either a unique object or a
> > > unique class of objects, where the "object" or class may be an idea,
> > > physical countable object (or class thereof), or physical noncountable
> > > substance (or class thereof), and ideally allows the object(s) to be
> > > located. 
> > > - Persistent Identifier: A long-lasting Identifier, ideally lasting
> > forever.
> > > - Unique Identifier: An Identifier which is unique within a specific
> naming
> > > convention (but may not be persistent)
> > > - Unique Persistent Identifier: A Persistent Identifier which is unique
> > > within a specific naming convention.
> > 
> > I would like to make some comments about the expression 'Unique Identifier'
> > There are two things that can be said:
> > 1. The expression 'Unique Identifier' is redundant.
> > An identifier is a name within a specific namespace that identifies (that
> > is, labels the identity of) either a unique object...
> > If the identifier of something identifies it, the identifier is necessarily
> > unique.
> > 
> > 2. The expression 'Unique Identifier' is ambiguous.
> > The definition:
> > 'Unique Identifier: An Identifier which is unique within a specific naming
> > convention'
> > might receive two interpretations:
> > 2.1 A unique object is identified by only one name within a specific
> > namespace;
> > 2.2 The identifier within a specific namespace identifies only one object.
> > 
> > I think that what was intended by 'Unique Identifier' is 2.2. Am I right?
> > If so, the definition of Unique Identifier should be: An Identifier within
> a
> > specific namespace that identifies only one object. Nevertheless, the
> > expression 'identifies only one object' sounds redundant (see Item 1.
> above).
> > 
> > In my opinion, the definition of Identifier:
> > 'An identifier is a name within a specific namespace that IDENTIFIES (that
> > is, labels the identity of) either a unique object...' already contains the
> > idea of unicity. For this reason, the definition of 'Unique Identifier' is
> > probably not necessary.
> 
> This raises the question in my mind as to whether there are identifiers
> which are not unique.
> For example we use the example of an ISBN, but an ISBN does not identify a
> specific book, but rather a set of probably identical books. This is
> essentially what we mean by "unique class of objects". 
> 
> So does a "unique id" refer to a unique object i.e.a unique book.
> 
> For example in a repository we may have multiple copies of a digital object
> and we don't care which one we get when we ask for a copy. On the other hand
> we do need to locate specific copies at various times.
> 
> That argument seems to lead to the conclusion that we do need the term
> "unique identifier"

Do we need the term "unique identifier"?

The concept of copy implies the existence of a master.

Recalling the ISBN example: the publisher has the master with its own ISBN and all the available copies of that master have also the same ISBN. We don't need to distinguish between copies, there are all the same. To locate the book master we look for the publisher, to locate a copy we look for a special kind of library that has that book.

In the digital world, the same identification scheme is valid.

A Digital Repository has an identified master and other Digital Repositories may have identified copies of that identified master (master and copies using the same identifier). It seems natural that each Digital Repository owns just one of these identified item (a master or a copy). Therefore, to locate such an item we look for a special Digital Repository which has such identified item.

These other arguments seem to lead to the conclusion that we don't need the term "unique identifier"
Comment 13 Gerald Banon (gerald.banon@gmail.com) 2024-02-20 12:32:01 UTC
(In reply to David Giaretta from comment #10)
> Another thought is that we talk about "namespaces" but what we actually need
> is that an identifier has to always be resolvable through a resolution
> system. Otherwise the identifier is just a string of characters.
> 
> SO perhaps we need something like
> From:
> - Identifier: An identifier is a name that identifies (that is, labels the
> identity of) either a unique object or a unique class of objects, where the
> "object" or class may be an idea, physical countable object (or class
> thereof), or physical noncountable substance (or class thereof). 
> 
> To:
> - Identifier: An identifier is a name and an associated resolution system
> that identifies (that is, labels the identity of) either a unique object or
> a unique class of objects, where the "object" or class may be an idea,
> physical countable object (or class thereof), or physical noncountable
> substance (or class thereof).
> 
> The APARSEN project defined:
>  Identifier_Management_Scheme
> 
> Definition:   A system that deals with identifying entities in a system by
> using identifiers. In the system IDs are used only as a way to make
> unambiguous reference to an entity and not as tokens to access to the system
> (this allows to distinguish ID management systems from authentication
> services).

Another question:
Do we need the term "persistent identifier"?

In the real world the term "persistent identifier" does not exist. In the digital world the term should not exist either, but rather the term "persistent hyperlink".  

As in the real world, in the digital world, an identifier is just a label (a string of characters). However, in the digital world, once a digital destination object has a location on the Web, it is possible to embed within a digital source object a computational procedure that, when activated, brings the digital destination object onto the user's screen. This procedure is known as hyperlink.

It is advantageous to distinguish between the terms "identifier" (which is a name) and "hyperlink" (which is a procedure).

The word "persistent" should refer to hyperlinks, not to identifiers.

Actually, an hyperlink becomes "persistent" when its attribute value involves the identifier of the digital destination object and a resolver domain name.

Actually, it is the attribute value of a persistent hyperlink that consists of "a name and an associated resolution system". This value is used to persistently locate the digital destination object. 

These arguments seem to lead to the conclusion that we don't need the term "persistent identifier".
Comment 14 David Giaretta (david@giaretta.org) 2024-02-21 12:45:32 UTC
(In reply to Gerald Banon from comment #13)
> (In reply to David Giaretta from comment #10)
> > Another thought is that we talk about "namespaces" but what we actually
> need
> > is that an identifier has to always be resolvable through a resolution
> > system. Otherwise the identifier is just a string of characters.
> > 
> > SO perhaps we need something like
> > From:
> > - Identifier: An identifier is a name that identifies (that is, labels the
> > identity of) either a unique object or a unique class of objects, where the
> > "object" or class may be an idea, physical countable object (or class
> > thereof), or physical noncountable substance (or class thereof). 
> > 
> > To:
> > - Identifier: An identifier is a name and an associated resolution system
> > that identifies (that is, labels the identity of) either a unique object or
> > a unique class of objects, where the "object" or class may be an idea,
> > physical countable object (or class thereof), or physical noncountable
> > substance (or class thereof).
> > 
> > The APARSEN project defined:
> >  Identifier_Management_Scheme
> > 
> > Definition:   A system that deals with identifying entities in a system by
> > using identifiers. In the system IDs are used only as a way to make
> > unambiguous reference to an entity and not as tokens to access to the
> system
> > (this allows to distinguish ID management systems from authentication
> > services).
> 
> Another question:
> Do we need the term "persistent identifier"?
> 
> In the real world the term "persistent identifier" does not exist. In the
> digital world the term should not exist either, but rather the term
> "persistent hyperlink".  
> 
> As in the real world, in the digital world, an identifier is just a label (a
> string of characters). However, in the digital world, once a digital
> destination object has a location on the Web, it is possible to embed within
> a digital source object a computational procedure that, when activated,
> brings the digital destination object onto the user's screen. This procedure
> is known as hyperlink.
> 
> It is advantageous to distinguish between the terms "identifier" (which is a
> name) and "hyperlink" (which is a procedure).
> 
> The word "persistent" should refer to hyperlinks, not to identifiers.
> 
> Actually, an hyperlink becomes "persistent" when its attribute value
> involves the identifier of the digital destination object and a resolver
> domain name.
> 
> Actually, it is the attribute value of a persistent hyperlink that consists
> of "a name and an associated resolution system". This value is used to
> persistently locate the digital destination object. 
> 
> These arguments seem to lead to the conclusion that we don't need the term
> "persistent identifier".

There are several comments I should make:
1) In the real world I like to think of my full name as a persistent identifier - although it may not always be unique.
2) Not all identifiers for digital objects are what would normally be called hyperlinks. For example the DOI is used as an identifier and the doi.org site says "A DOI is a unique number made up of a prefix and a suffix separated by a forward slash. This is an example of one: 10.1000/182". It then goes on to say how it can currently be resolved - I guess the implication is that the resolution mechanism may change.
3) My main point is that some identifiers are only meant to be used in the short term e.g. my physical address at the moment. Other identifiers are meant to be valid i.e. to get me to the object of interest over the long term (perhaps forever). This basis difference is what justifies the separate definitions.
I personally have doubts that what are claimed to be persistent identifiers are actually going to work over the long term (but they may).  

My conclusion is that we _do_ need a term "persistent identifier", and its use is commonly understood.

However I am still not clear whether OAIS is the place for a definition. 

Remember in the first comment http://review.oais.info/show_bug.cgi?id=329#c0 I wrote:
-----------------------------------
OAIS uses the terms (once each)
- Persistent Identifiers  (as an example of Reference Information in Table 4-1)
- Unique Persistent Identifiers ("Unique Persistent Identifiers within some specified naming system" in section 6.2.4)

These are capitalised.

However "unique identifier" is used several times, but never capitalised.
------------------------------------

I don't think that using a term once in a table or in one sentence is a string reason for OAIS to put it in the glossary.

It would be easier to remove the capitalisation for Persistent Identifiers and Unique Persistent Identifiers.
Comment 15 Gerald Banon (gerald.banon@gmail.com) 2024-02-23 04:44:24 UTC
(In reply to David Giaretta from comment #14)
> (In reply to Gerald Banon from comment #13)
> > (In reply to David Giaretta from comment #10)
> > > Another thought is that we talk about "namespaces" but what we actually
> > need
> > > is that an identifier has to always be resolvable through a resolution
> > > system. Otherwise the identifier is just a string of characters.
> > > 
> > > SO perhaps we need something like
> > > From:
> > > - Identifier: An identifier is a name that identifies (that is, labels
> the
> > > identity of) either a unique object or a unique class of objects, where
> the
> > > "object" or class may be an idea, physical countable object (or class
> > > thereof), or physical noncountable substance (or class thereof). 
> > > 
> > > To:
> > > - Identifier: An identifier is a name and an associated resolution system
> > > that identifies (that is, labels the identity of) either a unique object
> or
> > > a unique class of objects, where the "object" or class may be an idea,
> > > physical countable object (or class thereof), or physical noncountable
> > > substance (or class thereof).
> > > 
> > > The APARSEN project defined:
> > >  Identifier_Management_Scheme
> > > 
> > > Definition:   A system that deals with identifying entities in a system
> by
> > > using identifiers. In the system IDs are used only as a way to make
> > > unambiguous reference to an entity and not as tokens to access to the
> > system
> > > (this allows to distinguish ID management systems from authentication
> > > services).
> > 
> > Another question:
> > Do we need the term "persistent identifier"?
> > 
> > In the real world the term "persistent identifier" does not exist. In the
> > digital world the term should not exist either, but rather the term
> > "persistent hyperlink".  
> > 
> > As in the real world, in the digital world, an identifier is just a label
> (a
> > string of characters). However, in the digital world, once a digital
> > destination object has a location on the Web, it is possible to embed
> within
> > a digital source object a computational procedure that, when activated,
> > brings the digital destination object onto the user's screen. This
> procedure
> > is known as hyperlink.
> > 
> > It is advantageous to distinguish between the terms "identifier" (which is
> a
> > name) and "hyperlink" (which is a procedure).
> > 
> > The word "persistent" should refer to hyperlinks, not to identifiers.
> > 
> > Actually, an hyperlink becomes "persistent" when its attribute value
> > involves the identifier of the digital destination object and a resolver
> > domain name.
> > 
> > Actually, it is the attribute value of a persistent hyperlink that consists
> > of "a name and an associated resolution system". This value is used to
> > persistently locate the digital destination object. 
> > 
> > These arguments seem to lead to the conclusion that we don't need the term
> > "persistent identifier".
> 
> There are several comments I should make:
> 1) In the real world I like to think of my full name as a persistent
> identifier - although it may not always be unique.
> 2) Not all identifiers for digital objects are what would normally be called
> hyperlinks. For example the DOI is used as an identifier and the doi.org
> site says "A DOI is a unique number made up of a prefix and a suffix
> separated by a forward slash. This is an example of one: 10.1000/182". It
> then goes on to say how it can currently be resolved - I guess the
> implication is that the resolution mechanism may change.
> 3) My main point is that some identifiers are only meant to be used in the
> short term e.g. my physical address at the moment. Other identifiers are
> meant to be valid i.e. to get me to the object of interest over the long
> term (perhaps forever). This basis difference is what justifies the separate
> definitions.
> I personally have doubts that what are claimed to be persistent identifiers
> are actually going to work over the long term (but they may).  
> 
> My conclusion is that we _do_ need a term "persistent identifier", and its
> use is commonly understood.
> 
> However I am still not clear whether OAIS is the place for a definition. 
> 
> Remember in the first comment http://review.oais.info/show_bug.cgi?id=329#c0
> I wrote:
> -----------------------------------
> OAIS uses the terms (once each)
> - Persistent Identifiers  (as an example of Reference Information in Table
> 4-1)
> - Unique Persistent Identifiers ("Unique Persistent Identifiers within some
> specified naming system" in section 6.2.4)
> 
> These are capitalised.
> 
> However "unique identifier" is used several times, but never capitalised.
> ------------------------------------
> 
> I don't think that using a term once in a table or in one sentence is a
> string reason for OAIS to put it in the glossary.
> 
> It would be easier to remove the capitalisation for Persistent Identifiers
> and Unique Persistent Identifiers.

Now I understood your way of using the word identifier.

You use the word identifier in a broad sense. I was using it in a strict sense.

In the BROAD sense, an Identifier (of an object) is a name within a specific namespace that is assigned to the object. It may not always be unique and/or persistent.

In the STRICT sense, an Identifier (of an object) is a name within a specific namespace that is assigned EXCLUSIVELY and FOREVER to that object.

Commenting 1) - Uniqueness

In the broad sense, "someone's full name" is a person's identifier. It is NOT unique, as two people can have the same name. Therefore, in the strict sense, "someone's full name" is NOT an identifier of a person.

Commenting 2) - Identifier versus hyperlink

From what I said in my Comment 13, identifier and hyperlink are different in nature and, eventually a hyperlink may refer to an identifier (in the strict sense). For example, the HTML hyperlink <a href="https://doi.org/10.1000/182">xxx</a> refers to the DOI 10.1000/182. It is used to access a digital object (here, an html file) located at https://doi.org/the-identifier/resources/handbook.  When referring to identifiers like DOI (as in the above example) or IBI (Internet Based Identifier), hyperlinks are said persistent (i.e., location independent). However, hyperlinks may be much more, for example, resolution mechanism independent (when using relative URL - I can give you a pointer to a demonstration of such hyperlinks). In the latter case, the hyperlink shown in the above example becomes: <a href="doi/10.1000/182">xxx</a>.

Commenting 3) - Persistence

In the broad sense, "MY physical address" is an identifier of Myself. It is NOT persistent, as one day you may move to another home place. Therefore, in the strict sense, "YOUR home address", is NOT an identifier of YOU. A "home address" is an identifier of a home place, not, in the strict sense, for the person who lives there.

I personally know the doubts that what are claimed to be persistent identifiers are actually going to work over the long term. However, I could show you some interesting recent examples of what I call "Fully Persistent hyperlinks" that are independent of a global resolver.

Considering the atual stage of OAIS, I agree with you that OAIS may not be the ideal place for definitions of identifiers and hyperlinks.
Comment 16 Eld Zierau (elzi@kb.dk) 2024-02-23 14:14:38 UTC
I suggest the following definition:
"long-lasting reference to a document, file, web page, or other digital object, this could e.g. be a Uniform Resource Name or an identifier via a persistent indentifier system"
In this way we do not say anything about hyperlinks (which may not be the case) and it is also indirectly said that it is not necessarily identifiers via persistent identifier systems.
Comment 17 Mark Conrad (mark.conrad.iso@gmail.com) 2024-02-26 16:51:35 UTC
I would recommend rejecting the comment because OAIS is not a good place to resolve these definitions.
Comment 18 David Giaretta (david@giaretta.org) 2024-02-27 18:53:15 UTC
DAI20240227 after many weeks of discussion decided to reject this suggestion because the terms Persistent Identifiers  and Unique Persistent Identifiers are each used only once. The term "unique identifier" is used several times, but never capitalised, and is used in a rather informal way. Because of this OAIS is not a good place to define these terms.
However we should remove the capitalisation for Persistent Identifiers  and Unique Persistent Identifiers.
Comment 19 David Giaretta (david@giaretta.org) 2024-02-27 18:57:35 UTC
We are changing the capitalisation