[Pipet Devel] Re: window sill

[Brad Chapman]

> You may have seen the 'window sill' mentioned on the Loci list.  
This is
> pretty much a status bar for the windowlets that appear in the 
workspace. 
> Attached is a screenshot of one prototype window sill.

I dig the window sill idea! Although I have no buisness saying 
anything about GUI design, I'll make a few comments below :-)

> INFO
> * Visible only when window is shaded
> * Button contains a text message for each info item (not 
context-sensitive
> help)
>      * Time elapsed while unprocessed
>      * Unprocessed/processing/processed
>      * Location
>      * Help message for window's node (truncated)

Will the info button be replacing the "help" window thingy that I 
started using, or in addition to that?

> OCCUPY/UNOCCUPY
> * Click button1 to make self window pane occupy parent window
> * Click button3 to unoccupy self window pane from parent window
> * Window sill also occupies parent's
> 
> (When I say 'parent', I mean a node that contains a network of 
nodes, one of
> which being the node in question.)

This is a really cool idea, but will the old windowlet with the 
network GUI get recycled by python memory collection if it gets popped 
out of a parent window? Just throwing that out, I really have no idea 
how memory collection works in pyGtk. Maybe it will still be fine as 
long as you maintain a reference to it.
    Also, I was having a lot of problems getting windows to refresh 
properly after resetting the contents of a windowlet...

> VIEW
> * Click button1 or button3 to cycle through views:
>    * GUI view: Gtk (composite) interface

Can you describe this view? What does it look like?

>    * Network normal: Network with icons, links and windows

Is this the view we currently have?

>    * Network world: Only links shown but for entire network (no 
scrolling
>      needed).  Single-click in workspace switches view back to 
network
>      normal, focused in area of workspace where clicked (like a zoom
> feature)

Jeff, I was trying to play around with the canvas zoom feature as soon 
as I saw it was a feature, because I thought it would be a really cool 
thing to use. The GUI ended up really ugly for me because not 
everything scaled in the same way, especially the text box, if I 
remember correctly.

>    * Info view (all messages in single view):
>          * Time elapsed while unprocessed
>          * Unprocessed/processing/processed
>          * Location
>          * Help message for window's node
>          * (others)
>    * Context-sensitive help message: for ONE selected node in 
window's
>      workspace, not for the window's node

How are the old views "stored"? For instance, if you have the view we 
have now, with links and windows and all that, will the view be 
exactly the same (ie. all the loci in the same places, all of the same 
windows open, etc) when you scroll back around to it? 
    And again, I'm not positive if you'll see the refresh problem I 
couldn't fix.

Just playing devil's advocate on all this stuff. The ideas are really 
cool, and I look forward to seeing them in action :-)

Brad