Hi Detlev,
thank you for coming back on this topic.
As stated at that time, this implementation throws more question marks than everything else (even the doc realizes the contradiction between "single space = empty string" and "no blank-padded strings allowed"). As mentioned 2004, I think this could have been implemented otherwise and better. Anyhow, as this has been the way for over ten years now, the basic logic won't be changed.
Let me say it in a less diplomatic way: they have simply done a quite sloppy job on this. But from my point of view, this is not the actual problem, if you have done somenthing wrong you are anytime free to correct your mistakes. Its rather, and Im noticing similar circumstances on other SAP products also, SAP has a concern wide problem to admit a fail. I will never understand this: for my self Im more disappointed about faltering tryes to sell it as "its a feature" compared to statements like "we dont have the resources or motivation to fix this", like pretty demonstrated in this thread. Dont get me wrong, I like SAP and worked for a time in Walldorf HQ, its a great company and Im keepin my colleagues and company in good memories. But we obviuosly still need to learn to call a fail as a fail.
However, Im familiar with the docs you pointed me to, nevertheless thank you for that. At the end I will need to program helper classes and adapt my table structures to avoid mentioned problems.
cheers