I tested some things, and it seems the problem isn't SHBrowseForFolder itself. But it's more strange: A simple BrowseForFolder-application works well on 2k, but my real application doesn't! I did try and error and I found out, that I get a good path, when I run SHBroswForFolder before I open a database via ODBC (CDatabase) ?!?! Don't have a good idea now, but it has definitly to do with the DB-Open. Maybe it's a bug? This works great on XP or better on any machine with IE>5. Ingmar

Thanks for the help. I tried it out, but if gives me the same GUID back again I put in?! For your question, I use itemIDList = SHBrowseForFolder(&browseInfo); ...do some checks and then SHGetPathFromIDList(itemIDList, szPathBuffer); And at this point I've got that GUID in szPathBuffer. I found so many samples, everyone does it that way, but I already tested on 2 win2k machines and it doesn't work. (Except I update to IE6, but this is currently not an option). Maybe something is wrong with my browseinfo structure? I did nothing special, maybe some flags are needed I didn't set? Ingmar

Hello, I've got a class to browse folders similiar to CSXBrowseFolder here on code-project. It worked fine until I needed to run my software on a fresh installed Windows 2000 with IE5. Selecting e.g. "My Documents" gives me not the real path but an Id like "::{xxxxxxxxx}". I'm not sure if I'm right, but I think I need to resolve this "link" using the shell functions. But whatever I tried it didn't work. If I select a usual folder (e.g. c:\test) it works like expected. I found a message on MSDN, that win2k/ie5 doesn't follow these links. But newer versions do so. Seems I get the Id of the not real existing folder, while IE6 resolves these links automatically. Maybe someone can give me hint how to solve this problem? Thanks in advance! Ingmar