Friday, September 09, 2005
Appealing file managers.
After blathering on at some length about file managers, I ran across some of the ideas that the Appeal project is thinking about. Here is a link to the Content Manager proposals.
This is interesting, and gets the juices running. How to address the challenge of presenting the complexities of data in a way that looks good and is not confusing. The many ways that one document could be presented, or better, the inherent metadata contained within the document (and hopefully collected and collated by something like Tenor) could overwhelm a user, all the while granting the user enormous power. Using the smooth pop-up panes similar to what we see when we hover over an icon in Kicker suggests possibilities. A transient and 'melting' overlay, with possibilities of further drill-down and instantaneous back-out permits deepening levels of complexity. The more you drilled into an object, the more detailed the overlay. This would be very similar in action to taking a pile of papers, flipping through them, glancing at the ones that seem pertinent until you find the one you want. Easy in, easy out. If you find what you want, or want to bookmark or keep your thumb somewhere, allow a level of overlay to be opened as some kind of container.
This overlay view could apply to folders, either virtual or physical. Overlays would show different paths of inquiry or action, leading to another overlay or action screen. Some kind of visual or textual representation would show the user where they are.
When the same data is presented in many different ways and contexts, it is very easy to get confused. This has been the downfall of virtual file systems. Actually confused isn't the right word. It is difficult to remember how you got somewhere. It doesn't take very many levels to get confusing. This week I was working in a complex that has a new control system for their ice plant. The graphical display shows the equipment status and allows changing setpoints and viewing logs. There are 6 major screens. Everyone including me had to click on a few different views before finding what we saw before and would like to see again. But, and this is important, to do what I needed to do required access to much more. Similar to KDE users opening Konsole to do routine tasks that are cumbersome or impossible to do graphically.
A minor nit, and perhaps an important philosophical point. Please please please don't forget something so basic that we can miss it; computers are tools for managing complexity. If I have two documents in my life that I need to handle, edit, send, or whatever, two only, I don't need a computer. The internet, birds, flowers, cultures, scientific pursuits, any endeavor is appealing to us due to the infinite variety that is presented to us. We love abundance. That is why we have 250 gig hard drives. If you are doing a mockup, put 20, 30 items in it. It will change how you view the problem. Or at least force us to think about how we can easily get it down to two or three.
Totally agree! In the current VFS solution consisting of KIO slaves such as FISH:/ and SMB:/, every network location has its own hierarchy, separate from other network locations and from the local filesystem. FUSE and its dependent projects (SMB for FUSE, SSHFS, Fusedav etc.) solve this problem by enabling unpriviledged (non-root) users to mount network directories in the filesystem, thus giving every network file a directory address under /.
Even better, FUSE will be part of the official Linux kernel starting in 2.6.14!
Since FUSE is implemented at the system level, it gives transparent network access to all programs (shell utilities, QT, GTK, Motif etc.) without requiring that they be modified. Unfortunately, there is staunch opposition (see the long thread below):
from some KDE developers who don't want to give up the suboptimal VFS solution or to at least implement a GUI for mounting with FUSE.
Large selection of Websites for sale, starting as low as $19.99 come by and check it out.
Subscribe to Post Comments [Atom]
Links to this post:
Subscribe to Posts [Atom]