[replying to someone who wanted a good way to store many tiny configuration files, a la Windows' registry] By the way, Hans Reiser's original rationale for writing ReiserFS was not to write a journaling filesystem that will take over Linux, per se - it was *exectly* to write a filesystem that will alleviate the need for database files, configuration files, registries, tar files and things like that. He believed that all these were specialized namespaces that were invented only because the typical filesystem namespace was either not efficient enough (e.g., it is a waste to store each 10 byte registry line as a separate file) or did not have strong enogh ways to specify relationships between filesystem objects (directory-child and symlinks were the only ones) or to do sophisticated searching on these namespaces - such as the ones databases can do. See the "future vision" in the reiserFS site (http://www.namesys.com/) for a full description of his vision. Note that the current ReiserFS implementation (version 3) isn't very much like his complete vision. It does support some things needed for this vision: efficient support for tiny files and huge directories, and journalling (making the filesystem more reliable, like you'd expect from databases and configuration files his filesystem was aimed to replace). But Reiser admits that "Ok, it's time to fess up. The interesting stuff is still in the future. Because they are nifty, we are going to add database and hypertext like features into the filesystem. Only by using balanced trees, with their effective handling of small files (database small fields, hypertext keywords), as our technical foundation can we hope to do this. That was our real motivation..." ReiserFS 4, which will be released in a couple of months, will still be a long way from Resier's original vision. -- Nadav Har'El | Thursday, Oct 17 2002, 11 Heshvan 5763