View unanswered posts | View active topics It is currently Mon Apr 23, 2018 9:00 pm



Reply to topic  [ 20 posts ]  Go to page Previous  1, 2
 Generic Patterns for IIP TIPs 
Author Message

Joined: Mon Feb 27, 2012 11:01 pm
Posts: 282
Location: Moscow, Russia
The graph is OK, except that there should be ClassifiedContainmentOfIndividual in it.

Quote:
Does your pattern have roles pointing at the yellow and the green instances of PossibleIndividual?


This question has no sense yet. The theory of patterns is only appearing. So far it is only a compact way to store template combinations for mapping purposes.

There is no RDF convention to store patterns, you can see an initial discussion for that at viewtopic.php?f=3&t=160

iRING TIP Manager stores patterns in a relational database, my Editor is using Python dictionaries for this purpose.

Direct instantiation of patterns is theoretically envisioned by JORD pattern spec. There are no software to instantiate patterns today. iRING Tools users are probable thinking about it, but they do not need it yet. I can do it any moment RDF representation is agreed upon.

So far the only way to instantiate a pattern is to instantiate all templates included. My Editor does this, green and yellow roles filled by autogenerated URIs (like you see it for temporal roles in your templates).

Quote:
If so, what do you gain with this denormalization? Can you fathom the combinatory explosion?


Do not understand the question. It is semantically correct way to model engineering situation.

Proliferation of Scolem entities is no worse (and no better) then proliferation of temporal parts in exports of life-cycle data for your templates. There should be a business logic in the adaptor to ensure that only one Boiler individual is created for all patterns describing several different properties of this BoilerSystem.

Can not say any good thins about SPARQL. I just hope that at some point people will pay attention to my SearchLanguage (which already has pattern search) and understand that SPARQL should be encapsulated to obtain a higher-level language. From this moment technical aspects of SPARQL speed will be addressed independently of syntactic complexity. Or probable people will follow the road of http://www.optique-project.eu/ .


Wed Jan 08, 2014 7:48 pm
Profile WWW

Joined: Sun Jan 22, 2012 10:02 pm
Posts: 283
Quote:
The graph is OK, except that there should be ClassifiedContainmentOfIndividual in it.
We'll talk about that when we discuss templates.

Quote:
This question has no sense yet. The theory of patterns is only appearing. So far it is only a compact way to store template combinations for mapping purposes.
There is no RDF convention to store patterns
This means that they are not in the scope of ISO 15926, or perhaps in some future new Part of it (probably competing with Part 11).

Quote:
It is semantically correct way to model engineering situation.
But it is not compliant with Parts 7 & 8. It will not allow for lifecycle integration, and it will predictably lead to chaos. I rest my case.


Thu Jan 09, 2014 12:30 pm
Profile

Joined: Mon Feb 27, 2012 11:01 pm
Posts: 282
Location: Moscow, Russia
Quote:
This means that they are not in the scope of ISO 15926, or perhaps in some future new Part of it (probably competing with Part 11).


This means it is still on the implementation side, not standardised. The fact that implementation is synchronised between at least 3 teams gives a good chance that future standardisation will be more mature.

Quote:
But it is not compliant with Parts 7 & 8.


It is not so. Instantiation, as I've already mentioned, goes as a series of P8 compliant instances of P7 templates. If patterns are mapped to your template set - it will deliver perfect lifecycle integration data.


Thu Jan 09, 2014 2:07 pm
Profile WWW

Joined: Sun Jan 22, 2012 10:02 pm
Posts: 283
Yes, but then incorporate temporal parts for individuals and class of temporal parts for classes.


Thu Jan 09, 2014 9:05 pm
Profile

Joined: Thu Jul 05, 2012 2:33 pm
Posts: 39
So patterns are a means of realising consistent Part 8 exchange between systems since the complexity of mapping to multiple templates for each property of a system has proved too complex. The start of the IIP work proved beyond doubt that template and class selection is not something individuals can do consistently in different environments unless they have libraries like TIPS where they can see mappings agreed by consensus.

I am concerned about the injection of identifiers in Victors approach which needs some very careful consideration/system knowledge to ensure that the correct engineering information is exchanged and is repeatable. I don't think these supplimentary identifiers are throw away, non-persistent but key to semantic interpretation. I have no way of identifying made up identifiers as opposed to persistent maintained identities at the moment such as the RDF blank node concepts.


Tue Jan 21, 2014 11:13 am
Profile

Joined: Mon Feb 27, 2012 11:01 pm
Posts: 282
Location: Moscow, Russia
Andrew, do you mean my identifiers for generic patterns, like "ClassificationOfContentOfPartInAssembly"?

I had an intention to keep them informative from a datamodelling point of view, not engineering.

Generic patterns are used in the process of TIP design and approval, not in the mapping. To ensure correct mapping and subsequent exchange of engineering information TIPs should have names like Fluid Category or Fluid Code


Tue Jan 21, 2014 11:54 am
Profile WWW

Joined: Sun Jan 22, 2012 10:02 pm
Posts: 283
Patterns are, as Andy stated, "a means of realising consistent Part 8 exchange between systems".

To me that means that a predefined mapping is provided between a (well-defined) prompt and one or more interrelated template specializations.

If we take "Test Pressure" that translates into:

Attachment:
test-pressure.jpg
test-pressure.jpg [ 7.19 KiB | Viewed 1017 times ]

that translates into three templates:
- PRESSURE RATED ARTEFACT participates in HYDROSTATIC TESTING in the role of TESTED
- LIQUID COMPOUND participates in HYDROSTATIC TESTING in the role of TEST FLUID
- LIQUID COMPOUND has HYDROSTATIC TEST PRESSURE
(I used the designations from the RDL).

It is possible to give this pattern its own identity by collecting these three specialized templates in an instance of the template EnumeratedSetOf3Classes (http://www.15926.org/templatespecs/CL-SETOP-04.xml) with the Name "Test Pressure", and store it in the RDL. In that way each of the three collected templates keep their own identity and can be used individually as well, if so required.


Perhaps it is better to start a new thread on identifiers.


Tue Jan 21, 2014 12:30 pm
Profile

Joined: Mon Feb 27, 2012 11:01 pm
Posts: 282
Location: Moscow, Russia
Quote:
collecting these three specialized templates in an instance of the template EnumeratedSetOf3Classes (templatespecs/CL-SETOP-04.xml) with the Name "Test Pressure"


Yes, except for two other aspects (not yet used in iRING User Group TIP mappings, but envisioned in JORD TSPs):

1. Role optionality. Currently patterns do not have a strict FOL predicate semantics, as opposed to P7 templates. Therefore in some cases it will be possible to define pattern structures and rules for their expansion which will allow to use the pattern for partially complete data.

2. Mapping variety. Single pattern can be mapped to several template sets (one for one-time exchange, another - for life cycle integration), or even to a specialized non-template ontology (alternative realization of ISO 15926 like in TopQuadrant projects). This will enhance integration options and allow to reuse semantic efforts put into engineering modelling).


Tue Jan 21, 2014 12:49 pm
Profile WWW

Joined: Mon Feb 27, 2012 11:01 pm
Posts: 282
Location: Moscow, Russia
Another important example envisioned by Jan Glendinning and still never really implemented:

To make a mapping as simple for domain expert as possible, we can define something like ParentChildPattern. And have a rule engine in the adaptor which can decide which complex data modeller's concept it is -- Specialization, Classification or probable even PartWhole!


Tue Jan 21, 2014 12:53 pm
Profile WWW

Joined: Sun Jan 22, 2012 10:02 pm
Posts: 283
Don't try to boil the ocean!


Tue Jan 21, 2014 1:41 pm
Profile
Display posts from previous:  Sort by  
Reply to topic   [ 20 posts ]  Go to page Previous  1, 2

Who is online

Users browsing this forum: No registered users and 1 guest


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
cron
eXTReMe Tracker
Powered by phpBB® Forum Software © phpBB Group
Designed by ST Software.