I am aware but untested with needing to handle commits and have better error handling when on iPad but how is it handled when creating parent and child records on web direct?  Can someone point me to transactional issues like Mr. Giest has videod but on web direct?  I do not even know if it is safe to do.  Thank you very much for ideas even pointing me elsewhere.

A big issue with transactions in WebD is that when the user closes the connection, FM does an implicit commit so if that happens in the middle of your transaction then may still end up with bad data.

I was afraid of this.  Are there things I can do to stop them from leaving until finished?  

also, how do I identify if someone signs in if they are web direct?  I tried searching all the functions but do not see what I need.  Thank you, Wim.

Get(systemPlatform) http://www.filemaker.com/help/14/fmp/en/html/func_ref1.32.150.html will tell you whether the user is WebDirect

And no; unfortunately you can not prevent them from doing in the browser what they want to do.  It is one of the major concessions you have to live with from not using the native client (FMP / FM Go)

20 hours ago, David Nelson said:

I was afraid of this.  Are there things I can do to stop them from leaving until finished?  

No people can always close the browser...

What you can do is to flag records as incomplete until they finish your process or have them input into global fields only to create a record when you have all the data.

I'm dealing with this too.   What I've found is that:

 

I ended up using the OnRecordCommit event which triggers a script which normally returns False which blocks the commit --- unless my "Save" button has been pressed.    This is a common technique.

 

What I added was this:  The Save button calls a script which timestamp field indicating that the user actually pressed the Save button, and this is the flag that I use to indicate "was actually committed by the user on purpose".  Then it Commits the record.

 

So this way I can distinguish between two cases:   New records that were committed on purpose, and new records that were committed on accident.  You then need some additional logic to deal with records that exist but are missing the flag, but this is not that difficult.  For example, you could have a server-side-script that runs every X minutes and simply deletes records in that state.    Or instead of deleting the data, perhaps it emails the user to say "Hey, you left an incomplete entry, please fix it".