Re: Bookmark Patch Praise, Suggestion

[Peter Harvey <pah06 uow edu au>]

On Sun, 2004-08-29 at 16:24 -0400, Britt Selvitelle wrote:
> About a week ago I built w/ the Peter's latest bookmark patch and I must
> say, after using it a bit I find it very intuitive and natural. I guess
> I just wanted to throw one more vote in for its inclusion in the main
> CVS branch.

Thanks!

> In order to remedy this, I think a subtle title above each hierarchal
> topic would do the trick.

I see your point, and in many instances it would be nice to have. It was
actually the original idea proposed by Eric Newman (which my patch tries
to realise). His mockup is still online. I probably wouldn't make the
headings as large, but you get the idea.
  http://webpages.charter.net/hick_ninja/epiphany_menu_after.png

It does have some disadvantages though. Consider this example from my
own bookmarks. Should there be a title above the 'Constraints Journal'
bookmark? Is it necessary? How do we made that decision?
  http://www.dsl.uow.edu.au/~harvey/epiphany/Research.png


I'll also give the more formal explanation of why I've avoided creating
titles for subdivisions. First, this patch is based on my understanding
of Formal Concept Analysis, in which we define a 'concept' by:
      * a collection of attributes (called topics in Epiphany) and
      * a collection of objects (called bookmarks in Epiphany)

Formal Concept Analysis says that, given a collection of attributes
(four legs, bites, barks, hairy) you can identify the concept 'dog'.
Similarly, given a collection of objects (Fido, Lassie, Snoopy) you can
also identify the concept 'dog'.

In Epiphany I assume that each bookmark is sufficiently well named that
you can identify the concept that a collection of bookmarks belong to
*without* a heading. By that reasoning, my answer would be to rename
your bookmarks so that you can identify the concept (in this case FFTA).
Sorry!


There is one hope though. Can someone provide a nicely reasoned answer
to the problem of "when should we add a topic heading to subdivisions?".
I expect the answer lies in the answer to "when is it difficult to
derive the concept by looking at a collection of bookmarks?". It is
clearly not defined by the size of the subdivision (large subdivisions
don't need it, and neither do really small ones). If this can be
answered, I would be happy to implement it.

Regards,
Peter.