Thanks for your help!
It surely is a big difference with Http/HTML scripting. Moving around with transactions starts/ and ends does not seem like an option to me since it does not help to keep structure in the script. When I do this and look at the script in a few months time again I will wonder why certain actions are outside of the transaction start and end labels. Of course I can add comments, but I think it's far better if the script speaks for itself (and I thought that was one of the promises of TruClient scripting anyway)
I can agree that in reality the user is causing the typing interval, however in most cases one would be interested in processing time of the server once the full text is received. That information is now lost in the "noise" of the accumulated typing time.
So question: when I would have a fixed amount of characters for the text entered, would it then be fair to subtract the typing time afterwards from the total transaction time: transaction time - ((typing time)*(typing interval)) ? Would this then be comparable with the transaction time that would otherwise be measured when an Http/HTML scripting technique would have been used?
Peet