Re: [Gnumed-devel] state of test results handling

[Karsten Hilbert]

On Sat, Mar 29, 2008 at 12:04:20PM -0700, James Busser wrote:

> - how did a semi-discontinuous selection of cells get highlighted blue 
> (RBC in the left column, together with all 4 cells in the right column)

The grid class allows very complex selections of non-contig
cells. This particular selection came about by clicking the
column label field (the date) thereby selecting the whole
column. Then holding down CONTROL and clicking the row label
(the "1") thereby adding the entire row to the selection.

Note that column "test" is technically a standard column but
is treated as a special column holding test result names.
Putting test names on the row label instead of the numbers
didn't work well enough for what we plan to do with it.

> - if I correctly interpret that the action of the "Review" widget will be 
> to "set or reset" the values of all items in the table
No. All items within the table that are *selected* when the
review is invoked.

> (based on the 
> checkbox selections within the widget)
yes

>, then this may be an unwise 
> implementation, since almost never will *all* displayed items warrant the 
> same attribute.
indeed :-)

> Can the setting / resetting of the item attributes be 
> based on a multiple selection rather than based on the total contents of 
> the table?
it is

> Under this alternative design, if nothing had gotten selected, 
> nothing would get coded as having been reviewed,
yes, it evens sends a notice to the statusbar
so people know what to do

> suggesting maybe all 
> items in the display should be pre-selected by default.
That looks rather ugly. Selecting the visible cells is
rather a snap via click-dragging.

> - would the checkbox "Technically abnormal" within the Review widget  
> serve to let the user overwrite (alter) whatever value had been provided 
> by test_org?
No way ! :-)  What the test_org thinks is what they think.
GNUmed wouldn't dare alter that even if a misguided user
thinks that'd be a good idea. A user review attaches to a
certain result (or document) what that user thinks about it.
One limitation (which seems logical to me): There can only
be one review per result per user (which can be altered, of
course).

Karsten
-- 
GPG key ID E4071346 @ wwwkeys.pgp.net
E167 67FD A291 2BEA 73BD  4537 78B9 A9F9 E407 1346