Definitions and Comparisons

[Glenn9999]

I'm looking for a good comparison between the three things listed in the Topic Description. I found definitions for all three, but am finding it hard to answer the questions of what are they good for, why would I want to use one over the other (they seem like very similar concepts), and so forth.

[uid0]

Symbolic links are only available in Vista and later, but allow links across network shares.

Hard links (for files) and junctions (for folders) have been available on ntfs since at least win2k, junctions can cross local volumes but hard links only work within a local volume.

Hard links can be used to save space if you need the same file stored in more than one folder - e.g. if you were building a "x-in-1" windows install disk.

Junctions can be used to redirect folders - this can be useful if you have limited space available on one volume but need a folder to appear there.

http://en.wikipedia.org/wiki/NTFS_symbolic_link

http://en.wikipedia.org/wiki/Hard_link

dfhl duplicate file hard linker

http://en.wikipedia.org/wiki/NTFS_junction_point

sysinternals junction

[Glenn9999]

Okay, so a symbolic link and a hard link is essentially the same thing (same idea as a shortcut?) - just that the symbolic link can work for a network share as opposed to a hard link being only for the same drive? And a junction point is the same as a reparse point? Essentially to compare these concepts, the reparse point is for a directory, and a hard link is for a file, but they both essentially do the same job (make some disk resource appear to be there, but is redirected to another source?).

To add something else to the mix, how about object identifiers? I would think that would be a good enough idea (tracking files that are moved to their real locations), but the thing I'm wondering is that there has to be drawbacks to it that cause it to not be in wider use? That makes me wonder since OS/2 essentially had that feature built-in (and it was nice, too, no "resource is not found, searching drive" messages).

(If you're wondering, I've been fishing through the Windows API file reference and been trying to find stuff that isn't generally implemented and trying to figure out when/what/why to use them and get them implemented in a good way - that's how the file compression post in the programming forum came about)

Edit: I just noticed Sysinternals has a forum now, so that might be good for some more technical questions I've got

Edited April 25, 2009 by Glenn9999

[Former Big Iron Guy]

Glenn,

The *best* explanation and the best tool for these very esoteric "shortcuts" is found on Hermann Schinagl's website at Link Shell Extension (I.M.H.O. ! !)

His shell extension and file link name utility are pretty awesome.

I plan on using them to eliminate redundant data files in several games I have where the files in the "expansion" pack duplicate (at least hash-wise) the files in the original application. In my game program directories (Age of Kings/Conquerors, Age of Mythology/Titans, Cossacks: Back to War mod pack 1) I have right now 9196 files, with 832 originals with 1153 duplicates, said dupes taking up 10.90MiB of space. That directory is going on a space reduction diet real soon now.

And gaming isn't the only place this happens. If you use something like Chuck Baker's FEBE to back up Firefox extensions, same sort of thing can happen. In my 5 deep FEBE backup, I have 52 originals with 78 hash-duplicates taking up 15.13MiB of very precious network flash drive storage.

Try it and see. Works for me.

/s/ fbig

Edited June 15, 2010 by Former Big Iron Guy