Dear Operating System Vendors.

I no longer want to know where my files are stored. I no longer care. I have hordes of directories on my various computers called stuff ,downloads and documents , and the effort that it would take to organise them into a proper heirarchy is just not worth it. The heirarchical filesystem is a really wonderful thing for programmers and websites, but it just doesn't cut it for personal use.

I no longer care where my files are stored.

I'll be burned at the next stake over from Charles when the time comes, for this filesystem heresy. Just the other night, a co-worker was asking me about how diligent I was in organizing my email. I told her, "Not at all. I leave it all in one pile and then run the Find command on it later." She was shocked that I, alpha geek and info freako, didn't have some intricate taxonomy of folders into which mail was sorted by carefully crafted filters.

Years ago, when I first started using email, I did indeed do this with procmail and other arcane beasties. Then, I found myself cursing that I couldn't do cross-folder searches very easily. Also, the filters and folders started making less sense as their structure represented only one possible scheme for finding what I was looking for, and I was needing many possible kinds of schemes over time. So, eventually it all ended up in one pile, and searches became my way of finding things.

I abandoned bookmarks for Google by the same principle. Now, my bookmarks consist completely of bookmarklets and a few stray links to local on-disk pages like Python documentation. In fact, I'm wishing that I could create bookmark folders that are fed by Google API powered persistent searches.

So, now I'm looking balefully upon my filesystem. I haven't had much chance to play with BeOS, but I've read about the design of the BeOS file system and drooled. I hear about Microsoft's Longhorn and its WinFS and grind my teeth - I very much dislike Microsoft, but if they pulled this off, I'd have to sing their praises. Apple? Do they have any aces up their sleeves in this regard? Don't let a new fanboy down. :)

Anyway, that's what I want to see: Storage without explicit organization, but with super-rich metadata for super-fast searches. Allow me to create views made from persistent searches - my "project folder" is simply a collection of resources tied together by a common tag, one of many. And, if I want to form a project hierarchy, make my persistent searches into file objects too.

The main thing in all this, though, is that it be woven very deeply within the OS. I don't want a helper app. I want this to replace the standard metaphor completely.

RDF triples at the inode-level anyone? Heh, heh.

shortname=ooochg

Archived Comments

  • When BeOS went under, Apple hired their filesystem guy. Remember, Apple tend not to say _anything_ about future directions until they have something Steve Jobs can demo at MacWorld. (c.f. Joel Spolsky's latest essay)
  • I keep waiting for ReiserFS (for linux) to get this... http://www.namesys.com/whitepaper.html
  • Check out e4Graph, here --JYL
  • I share your annoyance with present file/filing systems! The main change needed as I see it is to move the emphasis from bookmarks as locators to bookmarks as identifiers (of course somewhere around there needs to be a locator). Given this then it makes things like including the same item in many folders (/categories) easy and doesn't tie you to hierarchical systems. One of the things I want my Ideagraph project (http://ideagraph.net) to be able to do is to maintain all this stuff in a friendlier (usable!) fashion. Ok, so it's layering on top of the OS but anything's got to be better than the organization Win (even Lin) gives you.
  • here is a kde/sql implementation of that idea from newdocms
  • Maybe this stuff shouldn't be in the OS, in case you want to change OS later!
  • I abandoned bookmarks for Google by the same principle. Now, my bookmarks consist completely of bookmarklets and a few stray links to local on-disk pages like Python documentation. In fact, I'm wishing that I could create bookmark folders that are fed by Google API powered persistent searches. Why don't you? There is huge potential in this idea, particularly if the search is distributed across many eclectic researchers/clients. This is actually what I wanted to do for the Google API contest- but, as usual, the idea didn't fit into the requested parameters. P2P bookmark cluster searches would eliminate the lag of publishing in the associative link loop. Making us, collaboratively, an order of magnitude smarter by virtue of enhanced proximity. If you don't I will.
  • I abandoned bookmarks for Google by the same principle. Now, my bookmarks consist completely of bookmarklets and a few stray links to local on-disk pages like Python documentation. In fact, I'm wishing that I could create bookmark folders that are fed by Google API powered persistent searches. Why don't you? There is huge potential in this idea, particularly if the search is distributed across many eclectic researchers/clients. This is actually what I wanted to do for the Google API contest- but, as usual, the idea didn't fit into the requested parameters. P2P bookmark cluster searches would eliminate the lag of publishing in the associative link loop. Making us, collaboratively, an order of magnitude smarter by virtue of enhanced proximity. If you don't I will.