Quoting I l <[EMAIL PROTECTED]>: > Lets say you want to store someones picture. > > In the database, you would insert the picture, who owns that picture, maybe > the ip address and request headers of where that picture came from, the > category, sub-category, sub-sub-category in which the picture belongs to, > etc. You can gather and store as much information about that picture as you > want with ease. > > When you want to access that file and its attributes, you only have to do > one database query.
...which decomposes to multiple file system accesses. > Wouldn't you agree that this structure is much easier to manage than storing > the file in a directory, then storing that extra information in a database? > Then to retrieve, you must do a database query and find the file in the > FILESYSTEM (hoping it is still there). The code is much more complex First, I'm probably doing more system work to access the file contents as part of a database record, since the attributes of that superblobby field are different from those of the short metadata describing it. Second, I'm forcing EVERY access to that picture, even the ones by standalone non-webbish utilities that don't care about the metadata at all, to use a heavyweight database engine instead of the filesystem code embedded in the operating system... which gets used by the DBMS anyway, but less efficiently. -- PHP General Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php