[sugar] Ideas for Journal: How epiphany browser manages bookmarks just with tags
C. Scott Ananian
cscott at laptop.org
Fri Sep 19 17:20:32 EDT 2008
On Fri, Sep 19, 2008 at 4:57 PM, Eben Eliason <eben.eliason at gmail.com> wrote:
> "Two") to get at the thing you're looking for. So, again, I'm not
> sure that order really matters.
> Of course, if it DID really matter for a reason I'm not presently
> considering, we could allow tags of the form:
> To match on "A", "B", "A B" "B A", and also "A/B" (but not "B/A"). In
> other words, the addition of the slash to the tag format is used
> similarly to the way quotes are used to group two tags into one.
> Instead of grouping, however, it orders instead. It's interesting,
> but I'm not sure I see a good utility there yet.
I think there is compelling utility in terms of mapping tag space to a
filesystem and back. I hope that (like quotes) there is not all that
often when you need to use them -- but they are a useful power-user
feature. The following are all related searches:
A B -- matches these tag sets "A B C", "B A C", "A/B C", "B/A C", etc.
A/B -- matches "A/B", "C/A/B", "A/B/C", etc.
B/A -- matches "B/A", "C/B/A", "B/A/C", etc.
"A B" -- matches "'A B' C", etc
/A/B -- matches "A/B/C" but not "C/A/B"
They express slightly different meanings, but in many cases will be
indistinguishable from A B. I think they add a lot of power to the
tag language, although naive users won't need to use it often.
Completion on a search for "A/B" would suggest "A/B/C" if there is a
'subfolder' C; it would suggest "A/B C" if there was an item labelled
with A/B and tag C. If a young user has never made subfolders, then
the slash-separated options will never be suggested and all this power
An interesting point Eduardo brought up was the relationship between
"folders" and "saved searches". Do "tag completions" (ie "sub
folders" or "related tags") show up in the journal itself, or only in
a pane during a search? If they show up as first class objects, then
it might be nice to have searches in general as first class objects.
I think I'm arguing that "tag completions" are not the same thing as
journal items, and only show up during a search. I could be convinced
Another interesting point: gmail's UI never lets you see the results
of the 'empty' search (that is, all objects). By default the search
is restricted to 'in:Inbox' and the easiest UI mechanisms always
restrict the search to a 'folder'. You can click 'Search Mail' with
an empty search query, though, and what you get is very similar to
what one might imagine the Journal to be: a chronological list of all
your email (activity instances), grouped by thread (version), with a
list of clickable tags on each which you can use to find other similar
emails (activity instances). Gmail does have a flexible 'rules'
system to help automatically tag/categorize/file documents, though; it
may be worthwhile thinking what the Journal could do in that regard.
( http://cscott.net/ )
More information about the Devel