(a) ‘links’ in system and head
(b) building things
In a way organisation/maintenance is about creating explicit links
thinking – what different ways are ‘links’ explicitly represented in
* email-email shared folders
* email-email shared recipient/sender
* file-file – shared directory
* url-url – shared category/tag in bookmarks
* email-file – attachment
* email-url – url in the email body
* file-url – url in the document text
(above ’email’ means email *message*)
Sem desktop people pick up other stuff from content or metadata
also implicit time connections "write email whilst editing file"
Some link data is available to the the system ‘for free’
(automatically generated) – e.g. attachment to email, sender, time
connections (but may be random)
some the user needs to generate (manual) – categories, tags
(b) BUILDING THINGS
Although the web tags stuff would be sexy app, I was thinking very
much what I know I could build easily myself – desktop stuff always
seems harder! In the long term the distinction is probably breaking
down, but not for a long time yet.
Idea – you focus on the desktop, but we work out what protocols,
APIs, interaction methods are needed to link the two.
For example, I’ve written before about the need for ‘desktop URIs’
more easily enable desktop apps to cross-link to each other an to the
If you do desktop things that are ‘web open’ then (given I think it
is so easy!) I can always add some sort of web link at soem stage.