Namespaces

[HansTeijgeler]

This only adds to the confusion.

A namespace per 15926 part is fine, but the data model was defined in Part 2 and not in Part 8.
So when an item is defined (again) in Part x we use the URI of that Part x, and when in Part y we use the URI of Part y?

In Part 8 the split in ontologies has been defined, which is fine. But the Part 8 URI had to be used for ISO reasons.
Most stuff we have and still will add has not been defined in Part 8. We already have three or four URIs for the data model, and by far most RDL items have not been defined in Part 4. If you keep using that URI for that version of Part 8 for all new items it would be misleading.

So let's grab the bull by the horns and come with a compact URI set with a format that does take care of versioning.
Very long URIs like

Code: Select all

http://standards.iso.org/iso/ts/15926/-8/ed-1/tech/reference-data/data-model#

simply take too much storage space, because they are present three times in every triple.

[Pavel Selchukov]

Hans, Onno, sorry for my interfere in discussion, but I have two questions

1. If part Part 8 states namespaces in format "http://standards.iso.org/iso/ts/15926/- ... data-model#" it means we have to use exactly this format, isn't it?

2. Slide power-set.pdf - it is more interesting. If possible_individual have the same set of members as class_of_individual (both class extensions contain exactly the same set of individuals) - so why you consider them as different entities?

The basic definition of set equivalence is "Two sets A and B are equivalent if they consist of exactly same members".

And why individual is a member of equipment item?

[OnnoPaap]

Hans, Onno, sorry for my interfere in discussion, but I have two questions
1. If part Part 8 states namespaces in format "http://standards.iso.org/iso/ts/15926/- ... data-model#" it means we have to use exactly this format, isn't it?

I will answer question 1.

Part 8 is a Technical Spec (the TS in ISO/TS 15926-8). That means that it is regarded something to be tested in industrial projects, where-after comments may be incorporated before it levels up to International Standard (IS). The 15926.org is a gathering place for these comments (see forums called "ISO 15926 standard, the parts")

Hans made the remark that very long namespace addresses will cost extra storage space in triple stores. This is true if the (long) namespace addresses are stored in every triple. I have seen triple stores having this.

So he proposed as a comment to the TS: use shorter addresses.

[HansTeijgeler]

Hi Pavel,

You wrote:

If possible_individual have the same set of members as class_of_individual (both class extensions contain exactly the same set of individuals) - so why you consider them as different entities?
The basic definition of set equivalence is "Two sets A and B are equivalent if they consist of exactly same members".

[HT] They are equivalent, but not equal.
I quote W3C:

owl:equivalentClass is a built-in property that links a class description to another class description. The meaning of such a class axiom is that the two class descriptions involved have the same class extension (i.e., both class extensions contain exactly the same set of individuals).
The use of owl:equivalentClass does not imply class equality. Class equality means that the classes have the same intensional meaning (denote the same concept). In the example below, the concept of "President of the US" is related to, but not equal to the concept of the principal resident of a certain estate. Real class equality can only be expressed with the owl:sameAs construct. As this requires treating classes as individuals, class equality can only be expressed in OWL Full.

<owl:Class rdf:about="#US_President">
<owl:equivalentClass rdf:resource="#PrincipalResidentOfWhiteHouse"/>
</owl:Class>

The entity type <possible_individual> is not equal to the instance 'Individual' of the entity type <class_of_individual>.
Let me explain:
1) A <possible_individual> is a <thing> that exists in space and time.
2) A <class> is a <thing> that is an understanding of the nature of things.
So an instance of the first is something that exists in space and time, and an instance of the second is an "understanding", so quite different.

In case you instantiate <possible_individual> you should further qualify that instance by stating that it is also an instance of <actual_individual> and/or <arranged_individual> and/or <whole_life_individual>.
In case you instantiate the instance 'Individual' of <class_of_individual> what do you get?

An instance of <possible_individual> is a member of the instance of <class_of_individual> called 'Individual' (which is not really adding any information).
All instances of <possible_individual> are member of that class 'Individual' and all members of the class 'Individual' are instances of <possible_individual>.
So the class extension (i.e. the collection of members alias instances) of both is the same, and owl:equivalentClass is legitimate.
But they are not equal and hence not interchangeable.

I suggest to use that instance of <class_of_individual> called 'Individual' only as the top of its class hierarchy.

Regards,
Hans

[Pavel Selchukov]

Hello Hans!

I think it is time for understanding of a nature of Class.

Could you explain to us what exactly conception ISO 15926 standard keeps in mind about a Class.

Because, now we see clear that this conception is out of classic set theory - this is not equivalent for set (Cantor theory), and this is not a set which couldn't be subset of another (Neuman).

And we are not familliar with conception of "meaning" in set theory, because as I understand you talking about "sets are equal, meaning is different", so please explain it in "for dummies" format.

[MatthewWest]

In ISO 15926 a class is a set, but it uses a simpler set theory than standard ZF set theory, without the axiom of foundation, but with a construction rule. You can have any set you can construct by pointing to its members. This prevents Russel's Paradox (you cannot point to a paradoxical object because they can't exist), but allows sets to be members of themselves and other useful loops.

Going back to the initial question, possible_individual and class_of_individual do not have the same members. However, the RDL class Individual is a member of class_of_individual and is equivalent to (another name for) possible_individual and therefore does have the same members.

Class_of_individual can have any subtype of possible_individual (including possible_individual) as a member, i.e. the limit on membership of possible_individual is the powerset of possible_individual.

The reason for having two objects to represent the same thing is the limitations of the modelling tools. EXPRESS does not have any means of showing that one entity type is a member of another, and the only relationship you can create between an entity type and a data record is being a member of (so showing that a member of class_of_individual is a subtype of possible_individual is not possible). This requires the repetition of an object at both the data model and instance level where these kind of relationships are important. In OWL this is sometimes called punning.

So in summary, possible_individual and the RDL entry Individual both refer to the same class with the same members (i.e. if you have an object that is an instance of possible_individual it can be implied that there is a classification relationship between it and the RDL entry Individual). We need both because of the limitations of the modelling languages we are using.

Regards
Matthew West

[OnnoPaap]

A summary about this topic.

• (OnnoPaap) There are three types of namespaces being used. We have to choose a type. Examples are:

  Code: Select all

  [1] http://rds.posccaesar.org/2008/02/OWL/ISO-15926-2_2003#ClassOfActivity
  [2] http://dm.rdlfacade.org/data#ClassOfActivity
  [3] http://standards.iso.org/iso/ts/15926/-8/ed-1/tech/reference-data/data-model#ClassOfActivity
  

• (OnnoPaap) Namespaces are not endpoint addresses (refer topic on pURL servers for technical solutions)
• (vvagr) There should not be a distinction between Core RDL and Part 2 data model classes. [Further references to classes, power sets and equivalent classes are omitted]
• (HansTeijgeler) Namespaces type we choose should be much shorter, because they are stored in every triple. Refer document on power sets. Example:

  Code: Select all

  http://dm.15926.org# - data model
  http://rdl.15926.org# - reference data
  http://tm.15926.org# - template model
  http://tpl.15926.org# - templates
  http://meta.15926.org# - meta data
  

• (vvagr) we don't need sandbox namespaces for dm: and tm:, use the already defined part 8 namespaces.
• (HansTeijgeler) There are no standardized URIs yet.
• (HansTeijgeler) Most stuff we have and still will add has not been defined in Part 8. We already have three or four URIs for the data model, and by far most RDL items have not been defined in Part 4. If you keep using that URI for that version of Part 8 for all new items it would be misleading.
• (Pavel Selchukov) If part Part 8 states namespaces in format "http://standards.iso.org/iso/ts/15926/- ... data-model#" it means we have to use exactly this format, isn't it?
• (OnnoPaap) Part 8 is a Technical Spec (the TS in ISO/TS 15926-8). That means that it is regarded something to be tested in industrial projects, where-after comments may be incorporated before it levels up to International Standard (IS). The 15926.org is a gathering place for these comments (see forums called "ISO 15926 standard, the parts")

I propose to start a poll on namespaces. Agreed? (or reply with further discussions)

At such poll all members of the MMT are asked to vote (a no opinion vote will be available)

[vvagr]

Let's discuss the set of options to vote on and start a vote!

As I can see, the choice should be between Hans proposal of concise new system, Part 8 set (including dm) and my proposal, which is essentially Part 8 with dm replaced with old PCA namespace.

[OnnoPaap]

Hans said I should not add to the confusion, but there is another thing about namespaces we should decide upon in my opinion.

Attached a spreadsheet which is attached to the working copy of ISO 15926-6. It shows 3 namespaces for each part 6 metadata term.

The question is, should our preferred namespace end with slash (/) or hash (#) ?

[vvagr]

We have in Part 8 meta http://standards.iso.org/iso/ts/15926/- ... a/metadata#

What's wrong with it?

Why do we see http://standards.iso.org/iso/15926/tech/reference-data# ? Is this due to some new rules adopted by ISO?

Comparing:

http://standards.iso.org/iso/15926/tech/reference-data#
and
http://standards.iso.org/iso/15926/tech/reference-data/

I just wonder - how can they coexist? There will be some huge problems with resolvability.

In my opinion there should be just one URI, in the Part 8 meta namespace. Person-interpretable identifier should not be used to form an URI.