Sorry Andi, you missed my point which was/is that we shouldn't have to search all over to try and find the necessary parts for a program to run, they should all be there and if necessary have a note in the 'readme' file that states they should be put in a different dir.
We agree here, the readme should explain what's necessary. Most of the time this is done correctly that way. Sometimes some readme is not complete. In such case a ticket or in case no ticket system for the program is available, a mail to the programmer is what a user has to do when he finds such error. At least when he gets a lot of work for free. Remember, errors happen as porters are only humans. But -
and no rpm/yum is not the answer.
I've interpreted your post in a way that you think you can run different current programs (ports) reliable, stable, reproducible and manageable in a way without yum/rpm/ANPM. If this is what you think then there are only a few explanations for that comes to my mind. Either you -
1) have only very little experience with current ports
2) you use only a very few ported programs in a very limited scenario
3) you do not update or try out many different ports
4) you're a super power OS/2 user who knows and remembers exactly where you've put everything on your disk and your paths and you have a super memory and memorizes every detail of every dll you ever put into some program directory and remember the correct order how you have to start which application and you're knowing all the details and problems of beginlibpath/endpath and don't have the need to run big applications with much memory requirements and ...
Hope 4) is valid for you ;-).