Where did "instance" go?

[Rhys Lewis 2]

I've started using "Child (Generalization)" instead of "Instance", and the sky hasn't fallen.  I'm going to miss being able to navigate easily between the thing and the parent/classifier, but I have to accept that this is bigger than I am.

While the child inherits stuff from its parent.  We often (colloquially) use the concept of parent-child when talking about meronymies.  So for me, that's not precise enough.  For inheritance I use "inheritor".

Paolo

Yes, I am looking for a hyponym, not a meronym.

[Modesto Vega]

[SNIP]and the sky hasn't fallen.

The sky rarely falls and the day it does we can all stop worrying 

[SNIP]

The deconstructionists would love that.  Meaning and language were literally dissociated, and yet everyone was left with the feeling of communication as if it had occurred.

There are 2 fundamental differences between natural languages and artificial languages, such as modelling languages and those used in mathematics and logic, namely:

• the former allows metaphors and other figures of speech to gauge how a message is received and what parts are lost in translation. I used both on my heliocentric/geocentrism reply.
• the later should be unambiguous

Now let's jump into the very interesting aspects of this thread.

[SNIP]
I'm going to miss being able to navigate easily between the thing and the parent/classifier, but I have to accept that this is bigger than I am.

What are you trying to achieve with having easy navigation? Traceability? Navigation? Both? Please note this a genuine question.

I've started using "Child (Generalization)" instead of "Instance", and the sky hasn't fallen.  I'm going to miss being able to navigate easily between the thing and the parent/classifier, but I have to accept that this is bigger than I am.

While the child inherits stuff from its parent.  We often (colloquially) use the concept of parent-child when talking about meronymies.  So for me, that's not precise enough.  For inheritance I use "inheritor".

Paolo

Yes, I am looking for a hyponym, not a meronym.

Red, Blue and Green are hyponyms of Colour, with the hyponym specialising the hypernym Colour. A hyponym is a type of the hypernym, in UML the hypernym generalises the hyponym or the hyponym specialises the hypernym.

A wing is a meronym of a plane, likewise a port is meronym of a switch or router. This is in UML is depicted by composition (or aggregation).

Typically inheritance is something associated with hyponyms and hypernyms and not with meronym.

The difficulty I always had with Sparx "Child (Generalization)" is what it does with the elements: adds "_child" at the end, and ignores the stereotypes (at least in V13). If it did not do that I will be a bit more helpful.

[qwerty]

I'm pretty sure you can express metaphors also in artificial languages.

q.

[Rhys Lewis 2]

Now let's jump into the very interesting aspects of this thread.

[SNIP]
I'm going to miss being able to navigate easily between the thing and the parent/classifier, but I have to accept that this is bigger than I am.

What are you trying to achieve with having easy navigation? Traceability? Navigation? Both? Please note this a genuine question.

Navigation for the author of the model and diagrams, rather than a function of the published artifacts.  As the model grows, and as other people add to it, I am often wanting to make sure that it remains coherent.  One of the problems I try to avoid is multiple copies of the same element - either because I have forgotten that I created one earlier or because some other person has created it.  I try to have as many ways to audit the integrity of the model as possible, including:

• Searching the model for objects of the same name/ip address
• Looking at other objects that should have a relationship to the object in question to see if a relationship leads me to the object with a name I wasn't expecting
• Going to the parent/classifier, then searching for that in all diagrams.  For example go from this switch to the platonic switch, and then from there I can find all diagrams that have "a" switch.

The thing that makes EA useful beyond the frustrations of its quirks and transient features is the model, which allows this kind of navigation.  If it were not for that I might as well use Visio or crayons.

[Paolo F Cantoni]

[SNIP]

Yes, I am looking for a hyponym, not a meronym.

Red, Blue and Green are hyponyms of Colour, with the hyponym specialising the hypernym Colour. A hyponym is a type of the hypernym, in UML the hypernym generalises the hyponym or the hyponym specialises the hypernym.

A wing is a meronym of a plane, likewise, a port is meronym of a switch or router. This is in UML is depicted by composition (or aggregation).

Typically inheritance is something associated with hyponyms and hypernyms and not with meronym.

The difficulty I always had with Sparx "Child (Generalization)" is what it does with the elements: adds "_child" at the end and ignores the stereotypes (at least in V13). If it did not do that I will be a bit more helpful.

Hey Modesto,

reading your post triggered a thought in my head (dangerous, I know)

Looking at your colour example, I would have said that "Red, Blue and Green are instances of Colour".  I thought (and you agreed) hyponyms were a TYPE_OF.

One "trick" I use is to use the indefinite article to determine if I have an inheritance or restriction (IS_WHERE).  Thus I can say a spoon is a TYPE_OF cutlery, but it is more difficult to say a green is a TYPE_OF colour.

That's because Green IS a colour WHERE the frequency is x.

In our modelling technology (as I've said before) we have Inheritance, Restriction and unspecified Specialization (not sure yet whether it's an inheritance or restriction).

As I said, just a thought.

Paolo

[Rhys Lewis 2]

One "trick" I use is to use the indefinite article to determine if I have an inheritance or restriction (IS_WHERE).  Thus I can say a spoon is a TYPE_OF cutlery, but it is more difficult to say a green is a TYPE_OF colour.

That's because Green IS a colour WHERE the frequency is x.

A PoE switch is a switch that is a type of switch.  The IE-4000-4T4P4G-E is a switch that is a type of PoE switch.  The switch in the guard house is the IE-4000-4T4P4G-E that used to be in the storeroom.

In this diagram we have the inventory of PoE switches, showing the location that each is allocated to.  The switch with serial number 12345 is now assigned to the guard house location instead of the storeroom location.  This increments the version of the diagram to 1.2.

Version 1.1 was an instance of the switch diagram that has been superseded by version 1.2.

The logical diagram shows a camera connected to a PoE switch.
The PTZ technology diagram shows a PTZ camera connected to a PoE switch, and the bullet technology diagram shows a bullet camera connected to a PoE switch.
The guard house physical diagram shows the PTZ camera with serial number 12346 connected to the IE-4000-4T4P4G-E with serial number 12345, and includes their IP addresses.

IE-4000-4T4P4G-E is a hyponym of PoE switch, which is a hyponym of switch.  Switch is a hypernym of PoE switch, which is a hypernym of IE-4000-4T4P4G-E.

So does the "Realized by" arrow point from the logical switch to the technology one, or the other way around?

[Glassboy]

So does the "Realized by" arrow point from the logical switch to the technology one, or the other way around?

ArchiMate relationships should form a coherent English sentence when read from source to target.  They're also generally the opposite direction to UML.

[Rhys Lewis 2]

So does the "Realized by" arrow point from the logical switch to the technology one, or the other way around?

ArchiMate relationships should form a coherent English sentence when read from source to target.  They're also generally the opposite direction to UML.

I always thought that too.  So the "Realized by" arrow should point from blue (logical) to green (technology).  Ie. The switch in the guard room instance of the camera plus switch logical pattern is realised by a IE-4000-4T4P4G-E.

However I have seen a lot of green to blue "Realized by" arrows, including in Gerben Wierda's Mastering Archimate, so I am beginning to doubt my understanding of it.

[Glassboy]

I always thought that too.  So the "Realized by" arrow should point from blue (logical) to green (technology).  Ie. The switch in the guard room instance of the camera plus switch logical pattern is realised by a IE-4000-4T4P4G-E.

However I have seen a lot of green to blue "Realized by" arrows, including in Gerben Wierda's Mastering Archimate, so I am beginning to doubt my understanding of it.

Are you looking at the shapescript direction or the underlying tracability view?  They're the opposite of each other.

[KP]

The more tangible entity (source) realizes (makes real) the more abstract entity (target).

[Rhys Lewis 2]

The more tangible entity (source) realizes (makes real) the more abstract entity (target).

So in the passive voice convention of ArchiMate, the Realized by arrow goes from abstract to tangible to show the relationship you have described?

[KP]

As long as the arrowhead is at the abstract end 

[Glassboy]

As long as the arrowhead is at the abstract end 

As long as the ArchiMate shapescript arrowhead is at the abstract end.  If you lose the stereotype it will be drawn the other way.

[Modesto Vega]

Now let's jump into the very interesting aspects of this thread.

[SNIP]
I'm going to miss being able to navigate easily between the thing and the parent/classifier, but I have to accept that this is bigger than I am.

What are you trying to achieve with having easy navigation? Traceability? Navigation? Both? Please note this a genuine question.

Navigation for the author of the model and diagrams, rather than a function of the published artifacts.  As the model grows, and as other people add to it, I am often wanting to make sure that it remains coherent.  One of the problems I try to avoid is multiple copies of the same element - either because I have forgotten that I created one earlier or because some other person has created it.  I try to have as many ways to audit the integrity of the model as possible, including:

• Searching the model for objects of the same name/ip address
• Looking at other objects that should have a relationship to the object in question to see if a relationship leads me to the object with a name I wasn't expecting
• Going to the parent/classifier, then searching for that in all diagrams.  For example go from this switch to the platonic switch, and then from there I can find all diagrams that have "a" switch.

The thing that makes EA useful beyond the frustrations of its quirks and transient features is the model, which allows this kind of navigation.  If it were not for that I might as well use Visio or crayons.

I’ll tackle the other post replies when I have a better screen and keyboard. To tackle the issue of duplication we have created a dashboard with a custom SQL statement to list just those elements in the repository with the same name, base class, stereotype and belonging to the same root branch.

By checking regularly, we got duplicates more or less under control.

[Modesto Vega]

[SNIP]

Yes, I am looking for a hyponym, not a meronym.

Red, Blue and Green are hyponyms of Colour, with the hyponym specialising the hypernym Colour. A hyponym is a type of the hypernym, in UML the hypernym generalises the hyponym or the hyponym specialises the hypernym.

A wing is a meronym of a plane, likewise, a port is meronym of a switch or router. This is in UML is depicted by composition (or aggregation).

Typically inheritance is something associated with hyponyms and hypernyms and not with meronym.

The difficulty I always had with Sparx "Child (Generalization)" is what it does with the elements: adds "_child" at the end and ignores the stereotypes (at least in V13). If it did not do that I will be a bit more helpful.

Hey Modesto,

reading your post triggered a thought in my head (dangerous, I know)

Looking at your colour example, I would have said that "Red, Blue and Green are instances of Colour".  I thought (and you agreed) hyponyms were a TYPE_OF.

One "trick" I use is to use the indefinite article to determine if I have an inheritance or restriction (IS_WHERE).  Thus I can say a spoon is a TYPE_OF cutlery, but it is more difficult to say a green is a TYPE_OF colour.

That's because Green IS a colour WHERE the frequency is x.

In our modelling technology (as I've said before) we have Inheritance, Restriction and unspecified Specialization (not sure yet whether it's an inheritance or restriction).

As I said, just a thought.

Paolo

The use of the indefinite article may be of assistance to establish if something is a type of something else but it is not bullet proof. "Spoon" is a type of cutlery irrespective of the article preceding, likewise "POE Switch" is a type of Switch irrespective of the article preceding it. I'll get into colours shortly.

According to Merriam-Webster, the definite article is used in English to refer to a person or thing that is identified or specified. While the OED, states "The main function of the definite article (in English, the) is to specify the noun given, while the indefinite article (in English, a or an) marks a noun as being generic."

To take Rhys' example, the use of a definite article before "Switch" or "POE Switch" denotes a switch (of a particular make and model) with a specific IP address. It is not the actual physical switch but a representation of a physical switch. Whether a Realization is the right or wrong way to depict the relationship between the representation of a physical switch and its hypernyms, I don't know and need to ponder about it.

Colours are one of those examples where the "article" rule to decide if something is a type does not work, but there many others: water, methane, and Sodium chloride are chemical compounds (there types of of chemical compounds), and a AFIK, in English, preceding them by an article without further lexicalisation does not work. The fact that red has a dominant wavelength of approximately 625–740 nanometres and that reds range from the brilliant yellow-tinged scarlet and vermillion to bluish-red crimson, and vary in shade from the pale red pink to the dark red burgundy, does not mean red is not a type of colour and that vermillion, crimson, and so on are not types of red.

In other words, because you can define something based on physical parameters does not mean there are not a type of.

Incidentally, POE Switch is also an example of lexicalisation.

To sum up, lexicalisation plays a role when trying to decide if something is a type of.