I am finding several places in the MM .NET framework where it appears that unnecessary and sometimes painfully slow testing is occuring. Maybe this testing is important, but with my limited experience I cannot explain why. I am hoping that this thread will elicit an explanation of why the testing is important or if it is not important, then I hope we can eliminate unnecessary testing and speed up the framework.

Let me describe a few specific cases:

When using the new EntityCentric business objects, each individual entity is really just a wrapper around each individual row in the ADO.NET DataTable, so to make this discussion easier I will refer to the rows rather than the entities even though sometimes the testing is done against the actual entities and sometimes against the rows. When a user edits rows in the DataTable and then clicks Save, all rows in the DataTable are tested for the required rules generated by the Business Layer Generator. This includes testing deleted rows as well as unchanged rows. I do not see any justifiable reason to test rows marked as deleted or rows that are unchanged and therefore will not be saved back to the database.

Also when using EntityCentric business objects, and the user clicks Cancel, every entity (DataTable row) is canceled individually. Even unchanged rows are canceled and the canceling process is horribly slow. I do not see any justifiable reason to cancel entities (DataTable rows) that have not been changed. During testing I selected all 5000 records from an SQL table and edited one record and then clicked the Cancel button and the canceling process took about 5 minutes.

Please don't tell me that I shouldn't select so many records unless I intend to use all of them. I understand that, but my question is why does the framework even test deleted and/or unchanged rows and why does it cancel changes on rows that have not been changed? Maybe there are legitimate reasons and if so, I hope somebody can explain them to me.

Sam Tenney