Fwd: Re: [Clipart] metadata: aren't the keywords actually categories (and can keywords be added)?

[Nathan Eady]

Glen Turner wrote:

> Hard links are seen as evil by system administrators and I would
> strongly encourage you not to use them.

We are currently not planning to use them in our packages.

> They don't cross filesystems (since they are an inode reference)
> so sysadmins are constrained as to the partitioning of the system's
> disks (and this factor alone has seen them disappear from modern
> Unix systems).
> 
> Also, consider finding the hardlink reference.  ls -l suffices
> to determine where a symlink goes. For a hardlink you need
> "ls -i -R | sort -n" for the filesystem.  Sysadmins hate that
> sort of crap.

In my more theoretical moments, when contemplating filesystem design, I 
have considered a sort of hybrid "firm link" that is similar to a hard 
link but retains a list of all the links.  Ultimately I concluded that 
it's not worth it.  (Among other things, if the list is to say accurate 
when directories further up the hierarchy are renamed, the act of doing 
so either has to recurse down looking for links that have to be updated 
(bad) or else the list has to contain an inode pointer to the containing 
directory, which means that a directory cannot be relocated to a 
different inode without checking every directly-contained file.)

And of course it wouldn't help OCAL unless every filesystem in 
significant widespread use supports it, which seems unlikely.

> The "best practice" way of doing a link is a "relative symlink"

Symlinks are not supported on enough systems for our purposes.
(Neither are hardlinks.)  For the present, I think we will
continue using copies.