The ORM Foundation

Hello Sjir,

First, I rather interjected a response into your thread, even though I hadn't read through the numerous proceding posts and replies.  I saw the post regarding using the ORM Foundation website as a test case, and Ken's reply to that.  Having talked to Ken a bit about the task of managing this site (which uses a powerful but complicated application called Community Server), I well understood his caution about opening up the structure of the site, its composition and management processes as a test case.  Ken knows more about Community Server than I do, but I think he'll admit to only a partial understanding of all that it entails.  My point was that a comprehensive model of the website would require a more complete knowledge than he has of C-S; and and that the burden of providing the expert knowledge he does possess in this regard, would be too much to ask.

I do like the idea of using the structure and workings of this website as a test case, but I think it will work only if the limitations to the UofD are well defined.  Your reply to me mentions some requirements, including "complete lists."  I take it that these should be comprehensive to the domain; but what perimeter would you set for this domain?  My suggestion is to limit the UofD to only those aspects of the website that are appearant to the user/members.  The advantage of this is that we all become equal as knowledge workers.  There would be no need for ken to provide details of the data structure, rules or procedures used by the C-S appliaction.  In effect, we would be creating a model that could be used to create (the high level logical structure for) an application to do what C-S does - though perhaps even better!  This would also have the advantage of being specific to ORM Foundation; so the generic aspects of the C-S capablities are not a concern.  The model would also be helpful should OF ever choose to port the website to another server application.  Creating a test case based on the OF website would then have a practical benefit to the organization, beyond the use as an example and learning excercise.

There is the question of the structure of the ORM Foundation that Ken has implemented using the facilities of the C-S application.  You might say that these need to be dicovered using Ken's unique knowledge (making him a kind of super knowledge worker), but I don't think that's required.  What we as user/members can discover of the structure, through what we can see as implemented in the website should sufice.  The model could always be modified later, if a more exact reflection of the actual structure of the OF be desired.

Would such a limited scope for a fact based conceptual model satisfy the objectives you had in mind for an OF website test case?  Is setting the limits for the UofD to only terms that might be collected from the experience of using the website reasonable?  I realize that it's not like having an established lexicon specific to the domain - but even that may be useful in that it allows a degree of freedom in selecting terms.

Looks like I have some reading to do, to begin to understand the context of your post and this sub-forum.

Thank you for the suggestion.  I hope we find a way to use it to advantage.

BRN..

Hello Sjir,

The color editing didn't come through, but from the bracketed insertions, It looks like we are on the same page.  I'll try to put togther a statement that we can use as a post here - to sort of kick things off - that will outline the purpose and general parameters for the excercise.  If you wish to edit that, or make additions - that will be fine.  After that, it will be just a matter of members submitting models, partial models, or adding comments and suggestions.  We'll see what we get, and go from there.  I'll have a go at an ORM2 model using NORMA myself.  I'm guessing there will be those wishing to collaberate on a model, but they can figure out the division of labor on their own.

Though I know he's busy, I'll ask Ken if we should keep the exercise as part of this forum, or in another part of the site.  The biggest issue I see is making diagrams, etc..., available for members to review.  I know that a better facility for these has been requested, but from Ken, I know that is still a work in progress.  Probably, the best method is to include links to URLs outside of this website. 

The more I've though about it, the more benefits I see of such an excercise.  Thanks again for the suggestion.

BRN..

Hello Sjir,

I've come back to your original question about modeling the OF website to post a proposal for a general fact-based modeling exercise.  I don't know to what degree what I propose will satisfy your objectives, but it is just a proposal, and certainly open to modification.  I'll just paste in the text here from a Word document I wrote (where it's easier to organize and edit):

A proposal for a fact-based conceptual modeling exercise

Working from a suggestion by Sjir Nijssen (in the SBVR forum), I’m proposing an exercise where OF members may submit conceptual data models for comment and review in this forum.  It is important to note that the proposed exercise is intended as a form of “sandbox” for experimentation and comparison of methods and ideas – this is not a production project.

Ken Evans has recently added the facility to attach images and files to a forum posts.  By doing so, any member can post a model for review – but the exercise I have in mind is a bit different. 

A problem with individual selection of modeling subject (the organization to be modeled), is that those that wish to review or comment on the submitted model must have as complete knowledge of the subject as the modeler – otherwise, the review is limited to syntax checking.  One solution to that problem is to use a modeling subject readily knowable to both the modeler and model reviewers.

Using an existing organization as a common modeling subject raises a number of issues, security and privacy concerns among them.  Also, providing all the information required to create a model, to both the modeler(s) and to those who will review the model, imposes a significant burden on someone in that organization.  In both cases, it’s access to “inside information” about the organization that is the source of the problem.  An alternative is to use a fictitious entity as a modeling subject.  In this case, there is no real proprietary information to protect, but the requirement to provide detailed information about the substance and workings of the entity remains.  This imposes a workload on the creator of the fictitious entity to generate, and publish, a great deal of simulated material.  Besides, these artificial entities are rarely convincing subjects.  A better alternative would have the following characteristics:

·       Is a real functioning entity

·       Exposes no proprietary information

·       Has sufficient published material to constitute the raw material for a model

I think that the public face of an internet website meets these criteria.  I suggest that the ORM Foundation website be used as the modeling subject.

The objectives of the exercise would be:

·       To allow OF members to create and submit fact-based conceptual models of the publicly exposed face of the subject website – using whatever methods and tools they wish

·       To allow OF members to review and comment on these submissions

·       To facilitate comparison of various approaches and styles of fact-based modeling

·       To stimulate discussion of the conceptual modeling process

A few notes about what the objectives would not be:

·       To create an “official” model of the OF (or any other), website

·       To model the assumed (from data not publicly available), internal structure of a website

·       To create a competition between members

·       To add to the burden of the administrator of the subject website

Your questions, comments and suggestions regarding this proposed modeling exercise are welcomed and appreciated.

BRN..

Hi John,

Just to comment on your last point - and to try to make my point clear.

You talk about a "concept" as though it can be separated and specified separately from the context in which it is used.
This is an important point that I want to explore by challenging your assertion that:

"The  concept is  "mourning armband", not "black armband"  or "white armband". "

It seems to me that the term "mourning armband" is itself only a small part of a much bigger story (and inseparable from it).
I think that your are missing the plot when you try to "correct me" by saying ""The concept is "mourning armband", not "black armband"  or "white armband".""

What about the little boy who says "Look daddy, that man is wearing a white armband!"  His concept is formed and expressed by what he sees.
Thus, it seems to me that your attempt to use a lexical approach to specify "concepts" is a (mostly) futile exercise. 

I have attached a small ORM diagram to illustrate my assertion that "concepts are a component of a context"

Ken

Hello Sjir,

I noted your bracketed comments to my modeling exercise proposal.  If other comments and suggestions are forthcoming, I'll create a new thread in this forum - to make it easier to track responses to the suggestion.

One question that comes to mind is how best to digest submitted models?  I think a method that does not require administrative action would be best.  I'd like to have members be able to submit models whenever and as often as they please - yet not have submissions overlooked, just because they happen to arrive in a bunch.  A possible option is to have each submission as not only a new post (with attached image, and/or file), but as a new thread.  The advantage would be that the software here keeps track of unreplied to threads.

Tracking unreplied to threads would clearly show any thread (and so, submission), that has yet to be reviewed.  I don't know if using this method would burden the system.

If anyone has other ideas on addressing an asyncronous pattern of digesting submissions, please offer them here.

Thanks,

BRN..

Hello Brian,

Please select what you think is the most convenient way of organising this; as far as I am concerned you are a much more experienced expert in Community Server than I am.

As said before I believe it is enough challenge to limit the submission to three approaches: CogNIAM, ORM and SBVR.  

I therefore like to propose as title for all threads:

A practical exercise in CogNIAM, ORM and SBVR: towards a common fact orientation approach.

I am happy to commit time to this effort. I hope enough other fact orientation experts will dedicate some of their valuable time to this effort.

Kind regards

Sjir 

Dear Ken,

My watch (Amsterdam time) says it is 22:40 when I wrote the previous post. Is it intentional that the website says 17:40?

Kind regards

Sjir

Hi Sjir,

The site time is set to GMT.

Each member can also set his or her preferred time in the personal profile.

Having said that, my forum times are also showing GMT-6.
I have asked technical support for their opinion on this and I will let you know the result.

By the way, if you have your personal profile time set to GMT+1 then your emails are correctly timestamped.

Ken 

Ken,

Thanks for the quick response.

I was a bit too early with my answer. After the message I checked my personal profile in order to adjust it and .... it was GMT+1 already. Mystery?

Kind regards

Sjir

Hello Sjir,

Thanks for the suggestions.  As soon as I see some other members express an interest in the exercise, I'll begin the new thread.

BTW, I am in no way an expert in the Community Server application that powers this website.  I actually noticed the feature of displaying unanswered threads because it is prominently displayed on the main Forums page, along with other customized information about forum posts and activity.  Significantly, this is just the type of discoverable fact that we hope members will include in the models they submit as part of the exercise.  The example also points out that I did not need to ask anything of the site administrator; and it was not a matter of having any special knowledge about the website, or the Community Server application that powers it.

More later,

BRN..

Hi Ken,

Thanks for the example of a posted image in support of a forum post.  I noticed you used a .png file - have you found that to be the best format for ORM models?  I was smart enough to click on the image to see it displayed clearly (and it that form was quite readable).  Did you use VEA to create the model fragment and image?

BRN..

Hi Brian,

Yes - I used VEA to create the model.
But I used Snagit to get the model from VEA into png.

I chose png because it creates a small efficient file and helps to keep the site running fast.. 

Ken

Ever since exposure to business rules in a three-day course on the topic while working at a local software company I had this intuition that ORM and SBVR had some deep underlying commonalities. This thread has reinforced that feeling, and gives hope to my idea that the two approaches (one emphasizing data, the other emphasizing business rules) might some day have a common language and set of communication techniques.

It is a bit hard to tell if this disussion terminated years ago, or merely shifted to a different location.