::scr FreeForm Filesystem

[Arvid Gidhagen]

> The other thing, that came up in a conversation last night, was - how
> does this looked to Usability types and Cog Scis </shameless attempt to
> smoke those creatures out>.
>
> By 'this' I mean, have data organised by query. Apple had big qualms
> about introducing aliases/shortcuts/symlinks because they didn't think
> that they mapped to a metaphor

It looks fine to me. I believe that the stubborn insistence on mapping
everything to metaphors that has been with us for quite a while
is actually more of a hindrance than anything. In some cases it
does work, of course, but in many cases metaphors only confuse
matters. Usually this occurs either because

1) the source domain would otherwise limit the target domain unduly,
   so there are lots of breaks in the metaphor (some breaks are OK,
   but if there are too many, the metaphor doesn't actually increase
   learnability)

2) the metaphor chosen is basically inappropriate (like dragging
   disks to the trash can)

or

3) in order to avoid 1), the UI designers limited the potential
   usage so as to comply with the source domain for the metaphor.


and also, as mentioned in 1), metaphors are usually only ever useful
for increasing learnability. Once you know the system, it doesn't
really matter whether it builds on a metaphor or not.

Queries are apparently tricky to master for "normal users" though.
Proper education would, I guess, be impossible. So the system
would probaly have to be *very* smart, almost as good at
understanding natural language as a human, to be accepted by the
mainstream.

> I believe there's been a lot of talk on various IA lists recently about
> taxonomys. How does this segue into those?

I don't know what the IA people put into the word "taxonomy",
but taxonomies usually have problems with multiple inheritance.
In computer science/AI taxonomies might be allowed to have
multiple inheritance, but it is troublesome (because you need
extra information about which path to pick given the context).
In other disciplines (like various flavours of linguistics or
cognitive anthropology) multiple inheritance is not allowed in
taxonomies. There has been a lot of debate about whether there
are any natural taxonomies at all except for in the ethnobiological
domain, and at the moment the answer seems to be "no".

- Arvid