In a specific project where we wanted to enter one field extra following extract and import of hospital records the following two problems emerged. Both aspects will be fixed in August. Just now there is a pause in development until August.

Be reassured that we do all we can to solve such issues since they are in essense what we call "show stoppers" of using the software - unless we manage to find and fix the cause. Which we - expressed with modesty - have done until now. Apart from a special issue in estimation of certain lifetables in EpiData Analysis. But for that one have decided to fix it as part of the rewriting of the software to the "epx based" standa

The current two issues relates to: EpiData Entry Client Program Version: 2.0.3.15 r481, Core version: 2.0.0.0 r1161, FPC Version: 2.6.4, Platform: i386-Win32

1. Lack of "value changed" upon entry in existing field with currently no information in that field Normally when you use the value label box (F9) to select values and change a value, then the current record is marked internally as edited. Upon change to the next record the system should ask "save record".

But apparently this does not happen, the EntryClient just moves to the next record without saving the value

2. The standard behaviour of EpiData software when a user moves a given window to a new "non-default" position is that next time that window is opened it will be placed in that same position.

This is currently not the done with EntryClient for the value label box and the "confirm change to records" box.

The remedy until fixed was the following: I added an integer field of length 1. Then I changed the range to 1-1 and added mustenter requirement. In other words the only possible entry is a 1 (one).

This shows that sometimes combinations of attributes can give problems, which we - unfortunately - have not identified yet. When we see such situations they are then added to the list of structured testing, which we apply before public release. The list is increasing with increased number of handled special situations.

The solution could only be identified by a structured "trial and error" process, where I managed to find the cause, which can then be remedied.

Unfortunately in some situations it is very difficult to create such a process. In particular when an issue is only present in some versions. An example of this is the 32/64 bit windows v8.1 issue which have struck some users and another issue where we have seen that a given epx related file works very well in Linux and Mac, but in Windows 7 creates an access violation on attempt to edit an existing record. .

Any structured input to solving such problems is highly welcomed.

Regards Jens Lauritsen EpiData Association Denmark.