From mboxrd@z Thu Jan 1 00:00:00 1970 From: Andrew Clausen Subject: Re: aal_assert() Date: Sat, 11 Jan 2003 00:26:28 +1100 Message-ID: <20030110132628.GA1056@gnu.org> References: <20021231005440.GA1573@gnu.org> <15901.23872.24060.68353@laputa.namesys.com> Mime-Version: 1.0 Return-path: list-help: list-unsubscribe: list-post: Errors-To: flx@namesys.com Content-Disposition: inline In-Reply-To: <15901.23872.24060.68353@laputa.namesys.com> List-Id: Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: Nikita Danilov Cc: reiserfs-list@namesys.com On Thu, Jan 09, 2003 at 02:30:08PM +0300, Nikita Danilov wrote: > These "message ids" are traditional way reiserfs code uses to identify > errors (probably because early during development preprocessor features > were unknown) and Hans feels himself attached to them emotionally. :-) Well, I'll give some arguments against this scheme: (1) it's more code, more work, more space, etc. Programmers are minimalists, and adding extra stuff feels dirty. (2) "ids" are the Wrong Way to identify things... things should be identified by their essence, not by tags. This is the whole OO vs relational database flamewar. Relational databases (the "by essense/being, not tag") won, game over. You don't query things by tags, but by what they are. I know this is all quite subjective... I just wanted to provide some emotional arguments as well as rational ones *grin* Another thing: I'd like to see libaal reused in all file system tool implementations. That's my main motivation for cleaning it up. My first target is ntfs... Cheers, Andrew