Forum Posts Feed
search for single repository
< Next Topic | Back to topic list | Previous Topic >
Posted by jerryk
Nov 8, 2008 at 06:55 PM
Let me then push the linking idea more aggressively. We can try to “link” items in separate repositories into some db program such as UR, WR, 1N, PB (Personal Brain). But has anyone done it successfully with simply the OS folder?
The Planner Beta (http://kftf.ischool.washington.edu/index.htm) is one gesture at this. Generally folks need to do the following:
* store files, emails, web pages, snippets
* organize the data by folders and/or tags (preferably in some outline/mmap like fashion)
* recall by navigating through the hierarchy or through search
An OS stores files just fine. One can also drag *.msg into a file folder. XYplorer has a nice feature where it will rename outlook emails upon drop to include sender, date, etc. Web pages are a pain, which is why so many of us use WR or surfulater or something else. MHTs or PDFs might be a solution (why should this be such rocket science?) Snippets of text are also a pain (although txt files are a possibility); this is also why so many of us use UR or Zoot.
WDS could index all of this.
The real weakness is in the outlining/organizing. Most file browsers are terrible. What I’d like is a view that allows one to see subfolders and their contents indented. Dopus has a view like this, I think, but I find that program difficult to use. Also, if we could somehow have an outline order (based on some *.xml file in the folder directory) so that we can move items up and down and indent, that would be great. And the xml file could also be used as a tagging system for files.
In any event, by now you get the gist. Does anyone use a file browser in this way? Which one?
Many thanks.
Posted by Chris Thompson
Nov 8, 2008 at 09:24 PM
I do work this way, for the most part.
A “view that allows one to see subfolders and their contents indented” is one of the standard OS X file browser views. There isn’t any way to control the order within a level except by alphabetization (or modification date, etc.), but you could rename files with prefixes.
I store my bookmarks as either PDFs or webarchives (basically the same as .mht), use the filesystem as an outline of topics, and then tag them with Spotlight comments. Images work the same way. The OS can create custom search views (only collapsed lists, not in outline context). I used to use a program called Leap that would create alternative graphical search views, but ended up switching to a different program (Together). But the program I use doesn’t have a proprietary database… the filesystem is its database… it’s just a convenient way of getting things in the right places and tagging them.
I generally don’t like the “old” notion of custom databases that you squirrel your information into and then can’t get it out. The filesystem exists and stores searchable metadata for a reason. But on the same token, data that already has a logical “home” (email, RSS items, music, etc.) makes more sense to keep in its original application, and you just store links to it in your project’s working hierarchy.
—Chris
jerryk wrote:
>Let me then push the linking idea more aggressively. We can try to “link” items in
>separate repositories into some db program such as UR, WR, 1N, PB (Personal Brain).
>But has anyone done it successfully with simply the OS folder?
>
>An OS stores files just fine. One can also drag *.msg into a file folder.
>XYplorer has a nice feature where it will rename outlook emails upon drop to include
>sender, date, etc. Web pages are a pain, which is why so many of us use WR or surfulater or
>something else. MHTs or PDFs might be a solution (why should this be such rocket
>science?) Snippets of text are also a pain (although txt files are a possibility);
>this is also why so many of us use UR or Zoot.
>
>WDS could index all of this.
>
>The real
>weakness is in the outlining/organizing. Most file browsers are terrible. What I’d
>like is a view that allows one to see subfolders and their contents indented. Dopus has
>a view like this, I think, but I find that program difficult to use. Also, if we could
>somehow have an outline order (based on some *.xml file in the folder directory) so
>that we can move items up and down and indent, that would be great. And the xml file could
>also be used as a tagging system for files.
>
>In any event, by now you get the gist. Does
>anyone use a file browser in this way? Which one?
>
>Many thanks.
>
Posted by jerryk
Nov 8, 2008 at 09:59 PM
Chris Thompson wrote:
>I do work this way, for the most part.
>
>A “view that allows one to see subfolders and
>their contents indented” is one of the standard OS X file browser views. There isn’t
>any way to control the order within a level except by alphabetization (or
>modification date, etc.), but you could rename files with prefixes.
Unfortunately, I am a Windows user and not likely to be able to switch anytime soon. It’s nice to hear that that’s a standard OSX view.
>
>I generally don’t like the “old” notion of custom databases that you
>squirrel your information into and then can’t get it out. The filesystem exists and
>stores searchable metadata for a reason. But on the same token, data that already has a
>logical “home” (email, RSS items, music, etc.) makes more sense to keep in its
>original application, and you just store links to it in your project’s working
>hierarchy.
I mostly agree… but this is where the file v. pile distinction can become relevant. I’m happy to pile most email becuase it’s fairly easy to serach (I keep coming back to NEO as a front end to Outlook email). But I probably want to “file” 5 % of my email, in which case, I might as well drag it into an OS folder if I can drag it into an Outlook folder.
Thanks for your thoughts.