For Review: A Framework for Composition - De-composing structure

30 views
Skip to first unread message

matthe...@informationjunction.co.uk

unread,
Mar 11, 2021, 3:23:40 AM3/11/21
to UK-ND...@googlegroups.com, Yeomans, Steven

Dear Colleagues,

Attached is a copy of the report: “A Framework for Composition - De-composing structure”. This is our first piece of the Foundation Data Model. We would appreciate your reading it and your comments on it both in terms of content and readability.

However, I have to apologise, because as is often the case at this time in the financial year, we are up against deadlines, which together with producing this taking longer and proving more challenging than we had anticipated means that for your comments to be addressed in this edition, we need them by COB Monday 15th March. If you can’t make that deadline, please still provide your comments. There is more to do on this subject and in the new financial year we will be addressing challenges which we identified, but did not have time to address in this report. We will therefore either be updating this document or producing an Part 2, and which of those we choose to do would be influenced by the comments we receive.

Just to wet your appetite, here is an excerpt from the introduction.

 

“At the core of the notion of a component breakdown is the component as an integral (dependent) part of the composite whole. This leads to a rich formal structure, one that requires careful consideration to capture well enough. If one is not sufficiently aware of this structure, it is difficult to determine what is required to produce a reasonably accurate representation – in particular, one that is sufficiently accurate to support interoperability.

 

“In this report, we describe this rich formal structure and develop a framework for assessing how well a data model (or ontology) has captured the main elements of the structure. This will enable people to both assess existing models as well as design new models. As a separate exercise, as an illustration, we develop a data model that captures these elements.”

 

We look forward to receiving your comments which can be posted here or sent to me privately as you wish.

Regards

Matthew West

 

 

Dr Matthew West OBE
Technical Lead – National Digital Twin programme
https://www.cdbb.cam.ac.uk/what-we-do/national-digital-twin-programme

 

image001.png
image002.jpg
FDM Seed - A Framework for Composition - De-composing structure v 0.31_CPa - branch for review.docx

Graham Meaden

unread,
Mar 11, 2021, 8:17:47 AM3/11/21
to UK NDT FDM
Matthew, I am struggling to open the document in Chrome.  Is it missing a file type extension?
thx, Graham

Chris Partridge

unread,
Mar 11, 2021, 8:22:21 AM3/11/21
to fdm ndt
Hi Graham,

Isn't is a docx file - so shouldn't you be using Word?

Regards,
Chris Partridge


Chris Partridge | Chief Ontologist | BORO Solutions Limited | www.BOROSolutions.co.uk
M: +44 790 5167263 | e: partr...@borogroup.co.uk

BORO Solutions Limited | Registered Office: 2 West Street, Henley on Thames, Oxfordshire RG9 2DU
Registered in England & Wales | Company No: 06025010 | VAT No. GB 905 6100 58



--
You received this message because you are subscribed to the Google Groups "UK NDT FDM" group.
To unsubscribe from this group and stop receiving emails from it, send an email to uk-ndt-fdm+...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/uk-ndt-fdm/a42787fa-5791-4bc4-807a-b3192627cc35n%40googlegroups.com.

Graham Meaden

unread,
Mar 11, 2021, 9:18:24 AM3/11/21
to uk-nd...@googlegroups.com
Hi
Just clicking to download is the problem. I shall try another browser.

Graham Meaden
Business Architect | Founder

E-mail: graham...@kipstor.com
Mobile: +44 7841 204 874
Office: +44 203 126 4962 (UK)


Kipstor Limited
BASE Bordon Innovation Centre
Broxhead House
60 Barbados Road
Bordon
Hampshire
GU35 0FX
United Kingdom

http://www.kipstor.com

Kipstor Limited is registered in England and Wales, registered number: 08825651.  The registered address is: 43 Berkeley Square, W1J 5AP, London, United Kingdom.

The information contained in this e-mail message (and any attachments) is intended only for the use of the individual to whom it is addressed and is private and confidential. If you are not the intended recipient, or the employee or agent responsible for delivering this message to the intended recipient, any dissemination, distribution or copying of this communication is strictly prohibited. If you have received this message in error, please destroy it (and any attachments) and notify the sender immediately by reply e-mail. Thank you for your cooperation.



You received this message because you are subscribed to a topic in the Google Groups "UK NDT FDM" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/uk-ndt-fdm/s9w6KS0xn7c/unsubscribe.
To unsubscribe from this group and all its topics, send an email to uk-ndt-fdm+...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/uk-ndt-fdm/CA%2B8EkRoc9BaCt6iTHxFnR1XCE8XWjJuYVrL-a9BH74Ji-QX78Q%40mail.gmail.com.

Graham Meaden

unread,
Mar 11, 2021, 9:20:58 AM3/11/21
to UK NDT FDM
Safari works fine. Chrome (plug-ins) trying to be too clever as usual!
thx

Chris Partridge

unread,
Mar 11, 2021, 9:21:22 AM3/11/21
to fdm ndt
Hi Graham,

Just checked, I can click to download in Chrome. It works fine for me.
So this suggests it may be a problem your end?

Regards,
Chris Partridge


Chris Partridge | Chief Ontologist | BORO Solutions Limited | www.BOROSolutions.co.uk
M: +44 790 5167263 | e: partr...@borogroup.co.uk

BORO Solutions Limited | Registered Office: 2 West Street, Henley on Thames, Oxfordshire RG9 2DU
Registered in England & Wales | Company No: 06025010 | VAT No. GB 905 6100 58


To view this discussion on the web, visit https://groups.google.com/d/msgid/uk-ndt-fdm/CACoDOzSn-FmgEVJ%3DHsa8So70OE%3D5SSw-Cbrjq_JYd%2BiQfz3HGQ%40mail.gmail.com.

Andrew Davis

unread,
Mar 11, 2021, 10:22:43 AM3/11/21
to UK NDT FDM
Hi Matthew. 

A great read and quite though provoking. The diagrams were excellent and easier to read for those less skilled in formal logic.

I particularly liked the section on atomic vs. molecular composition (very useful mental model to consider when defining Business Intelligence data models) and Cover as a concept to ensuring completeness of analysis. 

It also raised a few thoughts on the subject matter, which I have posted here just for stimulation rather than as formal comments. I also apologise if these points are a bit "A Level" as opposed to to the "Post Doc" standard you chaps are working at.

Null Component and Exceptions Only
On the null component concept, in life sciences, the absence of something can be critical - for example someone who has an organ or limb missing either at birth, or removed at a later stage. 
This "missing" thing is important and needs to be explicitly stated rather than implied by comparison with a standard "template" of component organs that should be present, or simply not mentioned because having something is "normal".
The fact that I may have 2 eyes won't be mentioned on my medical record, but the fact I have one missing or ne that is damaged certainly would be.
Hence body components will only be mentioned on an exceptions basis - only if the component is of specific interest. 
  • If the component was removed, then it is simply a temporal part, but its absence needs to be easily accessible rather than delving back through temporal parts to discover when it was removed. 
  • If it was never present, that is a slightly different scenario but still needs capturing.
If I was producing a reference data model to cover this scenario, by default, a person would have no component parts.
  • A component may be assumed to be present because a standard template says it would normally be there
  • I would only create data for a component if it is in some way interesting
  • I would create the persons missing organ as a component, but then indicate it as "not present".  
This component could then be linked to either Event which was either its "removal" (e.g. accident/surgery) or an "investigation" which determined it was not present (i.e. a diagnosis).

Mixing
It also stimulated some thoughts regarding "mixing" of fluids and the resultant identity of the material formed by it.
A good example would blood and transfusions. Blood is a composite of multiple materials which can be separated to various levels out if required (water, enzymes, even down to individual cells) , or considered as a whole.
During a transfusion, foreign blood becomes mixed with the patients so afterwards, the identity of the persons blood becomes a mix of their the donors. 
After approx. 4 months, the foreign matter is largely replaced by the patients so it becomes purely theirs and no longer has any traces from the donor - unless they have been unfortunate enough to contract something which has multiplied.
I don't know if this impacts on your analysis but it presents an interesting use-case to test your hypothesis to see how it would be accommodated because it is a little different from the usual car analogies.

When is something a Molecule or an Atom?
The level of detail that is useful varies depending on the business context. Something that is atomic to me (e.g. as a consumer) may be a molecule to another person (e.g. as a supplier). 
Sometimes an item can be both to the same person in different contexts. 
A good example would be food and allergens - I don't normally care what the ingredients of a cake I procure are, but when it is on sale, I need to consider the ingredients so I can warn people with allergies.
The Construction History section covers this, but the ingredients as a general concept are more of an upwards classification of the item as opposed to a constructional component i.e. "Contains Dairy" (classification?) vs. than "10ml Milk" (decomposition).

I also found your comment "if a component is missing from a BOM or a WMS, then the composite product or project is not whole, it is incomplete. It probably won’t work or won’t work properly." quite amusing in a medical context - looking at para-Olympic athletes suggests humans are great at adapting :-) 

Regards

Andrew Davis
Enterprise Data Architect
Nuffield Health

matthe...@informationjunction.co.uk

unread,
Mar 11, 2021, 10:42:35 AM3/11/21
to uk-nd...@googlegroups.com

Dear Andrew,

Very pleased to hear you found it a good read.

Thank you for these reflections and examples. Very helpful. Our whole approach is to work from data or thought experiments to test what we are doing, so these examples from a different discipline are particularly useful. We will work them through the sausage machine.

Regards

Matthew West

Chris Partridge

unread,
Mar 15, 2021, 10:21:02 AM3/15/21
to fdm ndt
Hi Andrew,

Many thanks for taking the time to comment.
I've quickly added some clarifying comments inline - to make sure everyone is in the same ballpark.

inline ....

Reading this, I think we need to make the distinction between integral and replaceable part clearer in the paper. And also that we are focusing on the individual breakdown, not the class level. Hopefully the comment make this clear.

On Thu, 11 Mar 2021 at 15:22, Andrew Davis <andrewjd...@gmail.com> wrote:
Hi Matthew. 

A great read and quite though provoking. The diagrams were excellent and easier to read for those less skilled in formal logic.

I particularly liked the section on atomic vs. molecular composition (very useful mental model to consider when defining Business Intelligence data models) and Cover as a concept to ensuring completeness of analysis. 

It also raised a few thoughts on the subject matter, which I have posted here just for stimulation rather than as formal comments. I also apologise if these points are a bit "A Level" as opposed to to the "Post Doc" standard you chaps are working at.

Null Component and Exceptions Only
On the null component concept, in life sciences, the absence of something can be critical - for example someone who has an organ or limb missing either at birth, or removed at a later stage. 
CP> Agreed. But these are both catered for - though in different ways.
CP> The For the first case, the key distinction is integral and replaceable parts - see e.g. Appendix A. In the case where body parts are removed, there is a my body part x - it just doesn't exist at all time (froma 4D perspective). For an integral parts, we are interested the 'dependent' part - rather than the replaceable one - so in the case of, https://en.wikipedia.org/wiki/Oscar_Pistorius, Oscar's right leg, not the natural leg that was amputated nor the prosthetic legs he wore from time to time (which we presumably playing the role of his legs).
CP> In the case, where it is missing from birth and never replace, then it order to be able (semantically) to talk about missing, we need to have a counterpart (in a possible world semantics) with which we compare - and relative to this the part is missing.      
This "missing" thing is important and needs to be explicitly stated rather than implied by comparison with a standard "template" of component organs that should be present, or simply not mentioned because having something is "normal".
CP> I understand and sympathise with the 'language' view - but from an ontological view one would have to explain how a something missing and so non-existent also existed. 
The fact that I may have 2 eyes won't be mentioned on my medical record, but the fact I have one missing or ne that is damaged certainly would be.
Hence body components will only be mentioned on an exceptions basis - only if the component is of specific interest. 
CP> I wonder in this case whether one is mentioning a part rather than taking on board a whole decomposition into components. This would then be handled by the 'flat' mereology. 
  • If the component was removed, then it is simply a temporal part, but its absence needs to be easily accessible rather than delving back through temporal parts to discover when it was removed. 
CP> I think here the distinction between integral and temporal parts (see above) is being conflated. 
  • If it was never present, that is a slightly different scenario but still needs capturing. 
CP> See above. 
If I was producing a reference data model to cover this scenario, by default, a person would have no component parts.
  • A component may be assumed to be present because a standard template says it would normally be there
CP> The relation between the class-level template and the instances is more subtle - as you note above. 
  • I would only create data for a component if it is in some way interesting
CP> One of the purposes of components is to suggest a breakdown. So presumably you would have legs and eyes in your breakdown, not the right 55% of the leg or eye or the combination/fusion of the leg or eye. Componentisation play many roles. However, once you have the breakdown, you can use it in the way you see fit. 
  • I would create the persons missing organ as a component, but then indicate it as "not present".  
CP> See above - I think you might have difficulties explaining the ontology of 'not present' - though of course there is no harm in presenting the underlying ontology in this way.  
This component could then be linked to either Event which was either its "removal" (e.g. accident/surgery) or an "investigation" which determined it was not present (i.e. a diagnosis).
CP> If you linked something to a removal event, then it could not be in principle, an integral part. 

Mixing
It also stimulated some thoughts regarding "mixing" of fluids and the resultant identity of the material formed by it.
CP> Mixing is an interesting topic - going back to Aristotle's   On Generation and Corruption (De Generatione et Corruptione)
A good example would blood and transfusions. Blood is a composite of multiple materials which can be separated to various levels out if required (water, enzymes, even down to individual cells) , or considered as a whole.
During a transfusion, foreign blood becomes mixed with the patients so afterwards, the identity of the persons blood becomes a mix of their the donors. 
CP> Excellent example, but again the distinction between the integral - my blood and the replacement blood is not so clear (it seems to me).
CP> So there is no change in my blood - in the integral sense - but the 'foreign' blood changes state, becoming for a time a part of my blood - again see Appendix A. 
After approx. 4 months, the foreign matter is largely replaced by the patients so it becomes purely theirs and no longer has any traces from the donor - unless they have been unfortunate enough to contract something which has multiplied.
CP> I think there is a Ship of Theseus type question here - my naive view was that once the blood is in the person doing blood-like things, it is the person's blood. Is that not how things work. Of course, the alternative way is possible to model, just it seems odd to me (but I am no expert). If the alternative is the official view, then it is a really interesting case.
I don't know if this impacts on your analysis but it presents an interesting use-case to test your hypothesis to see how it would be accommodated because it is a little different from the usual car analogies.
CP> Yes, mixtures provide good use cases. many thanks. 

When is something a Molecule or an Atom?
The level of detail that is useful varies depending on the business context. Something that is atomic to me (e.g. as a consumer) may be a molecule to another person (e.g. as a supplier). 
CP> Exactly. And what the proposal does is give an exact formulation of this insight. 
Sometimes an item can be both to the same person in different contexts. 
CP> Exactly.
A good example would be food and allergens - I don't normally care what the ingredients of a cake I procure are, but when it is on sale, I need to consider the ingredients so I can warn people with allergies.
The Construction History section covers this, but the ingredients as a general concept are more of an upwards classification of the item as opposed to a constructional component i.e. "Contains Dairy" (classification?) vs. than "10ml Milk" (decomposition).
CP>  

I also found your comment "if a component is missing from a BOM or a WMS, then the composite product or project is not whole, it is incomplete. It probably won’t work or won’t work properly." quite amusing in a medical context - looking at para-Olympic athletes suggests humans are great at adapting :-) 
CP> I can see this, but I hope you can see why most of the textbooks on BOM and WNS stress the cover aspect.
CP> If IKEA started supplying you with 'Graham'-flat packs with key parts missing, I'm sure you would complain - ditto is your staff gave you WBSs with key deliverables missing.
CP> Also, you comment related to how we classify the construction - not the individual construction, which is the topic of the paper.  

Graham Meaden

unread,
Apr 15, 2021, 3:13:10 PM4/15/21
to UK NDT FDM
@andrew
To your point about Null Component and Exceptions Only there is a question of whether the TLO will proceed with a closed world assumption or an open world assumption.

matthe...@informationjunction.co.uk

unread,
Apr 16, 2021, 3:48:23 AM4/16/21
to uk-nd...@googlegroups.com

Dear Graham,

A good question, but not really one for Andrew.

 

@andrew

To your point about Null Component and Exceptions Only there is a question of whether the TLO will proceed with a closed world assumption or an open world assumption.

https://en.wikipedia.org/wiki/Closed-world_assumption

[MW] Open and closed world assumptions are largely a red herring for the NDT, since it is really about how reasoners are going to work, not what it is you are talking about. The NDT is not about reasoning it is about data sharing, so the open/closed world problem does not really arise (except in as much as we used reasoning for quality management of e.g. the reference data).

Regards

Matthew

image001.png
image002.jpg
Reply all
Reply to author
Forward
0 new messages