The ORM Foundation

I'm trying to populate with sample data the Person entity sub-type of Party(.Handle) by clicking on the solid arrow joining both entities. I can enter handles in the Party column (not in the Person column, it makes sense since I'm using the same key), but what about when I want to enter a last name (value of Person)? I click on the corresponding role but the Person colum doesn't allow me to select a Party handle. I hope I'm explaining the problem right...

Me either. When I have a fact type that has one role that is a subtype, I cannot populate that side of the fact type. I can't find a way to cause the drop-down list of role instances to appear, no matter where I type.

I am unable to use the ORM Sample Popualtion Editor to edit the MalePatient side of the population for MalePatient has ProstateStatus or of Malepatient is a subtype of Patient. Clicking on the MalePatient entity shows a blank and disabled population editor.

Hi,

Not sure if it would help (haven't tested it), but try giving explicit names to each role in the model.  I found that doing this helped cure some ills with ring constraints.  If that doesn't help, the fault is likely in sample population handling of the current nORMa tool CTP.  Matt C. mentioned a while back that the team trys to focus on one or two areas for improvement at a time - leaving some areas laging a bit.  I don't think the Sample Population features have been high on the list, compared to other priorities (but one of the team memebers would be the source for that sort of inside info).

I also recall some problems with removing sample population data in earlier CTPs.  If you start seeing lines in the error list that seem odd, check to make sure sample population data you thought you changed/removed was.

The hint about explictly naming roles (and other elements), has worked as a sort of cure-all for a number of woes.  You'll likely notice it helps when it comes to generating logical schemas and physical implementations.

Hope it help.

BRN..

Hi,

I forgot to mention, the image you used looks good.  What procedure did you use to include it, and what file format did you use? 

BRN..

The role names didn't help.

 I then tried something that "helped" in the sense that I think it shows where the bug is. I gave a reference mode to the subtype. This allowed data entry in the Sample Population Editor.

 Unfortunately, it adds a bogus extra column to my database table, so has to go. However, it suggests that the bug is that the Editor doesn't notice that the subtype has a parent type and shares the reference mode through the subtype link.

Hi John,

You might try keyword searching the old forum posts (from the SourceForge venue).  I belive they were all transfered here.  My recollection is a thread with the issue of reference modes and subtyping was addressed before. 

Check the libaray here, or search for Dr. Halpin's published articles on Subtyping - I'm sure I've seen a couple.  They won't be tool specific, but should help with determining what's actually required by the theory. 

As for the nORMa tool sample data editor, I think everyone would like that made more user friendly, and comprehensive, at some point.  Sample data is important for validation, but if it's a pain to workwith, most tend to avoid the step.  My guess is that general UI considerations are going to be lower priority than stablity and other issues.  If this were a commercial product, the marketing folks would have that switched around.

Glad you found the probable bug.  On the SourceForge site, they have (or did), a bug reporting tracker.  Check to see if it's still active for the ORM project there.

BRN..

It just isn't done. This was next on my list, after some name generation issues and fixing the unary binarization pattern. This is the easiest of the blatantly undone sample population issues--as has been noted in this thread, the identifier from the supertype is not recognized.

The other major undones for entering sample population are:

• Automatically populating implied facttypes (available as expansions below objectifications in the Model Browser). Please do not try to populate these in the meantime, you will eventually crash.
• Migrating sample population settings when a prefered identifier is changed.
• Allowing sample population with no preferred identifier, and migrating the settings when a pid is added

My apologies for not having this in a readme. At least it doesn't crash, which it did a couple of months ago.

I will also take a look at some of the other usability issues discussed in this thread. The trick is to get a balance where you are in 'instant edit' mode to allow easy typing but are still able to use the arrow keys to navigate around the grid. 

-Matt