https://bugs.documentfoundation.org/show_bug.cgi?id=114789
Elmar <[email protected]> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEEDINFO |UNCONFIRMED
Ever confirmed|1 |0
--- Comment #7 from Elmar <[email protected]> ---
I notice that if I stop what I am writing something here and then interrupt
that to save an attachment, this tool loses what I have typed. I spent about
1/2 hour typing notes and have had to retype 2x because of being distracted and
then forgetting about that and not saving the changes first. Is there no easy
way to avoid that?
I will try to reconstruct what I tried to say:
One of the the formative periods in my career was when I worked for Arthur
Andersen Consulting Division in the days before AA&Co was destroyed by the
Enron corruption (the division survived because it had already started
distancing itself and is Accenture today).
We had a motto: Analyse, Simplify, Automate.
When developing software, the basic principle was, you first design the optimal
solution. Then you de-optimise it, to cater for the weaknesses of the
technology and the users.
You probably know that early computers suffered from two basic problems:
limited data storage and computing power. So, for key functions (like adding
two numbers) you typically had three options: one instruction would use minimal
storage, another maximise speed, they were mutually exclusive. Then there was a
third which you could use if you did not care.
As the type of project I managed changed, I found that these principles applied
elsewhere as well.
For example, if we are building a plant, you have to consider the the file of
the plant. If it will be used for a very long time, you want to build the
highest quality plant will give you the best return in the long run even it it
take syears to pass the break-even point, because of the minimisation of
maintenance and stoppages during the life of the plant.
If you are going to use it for only a year or so, then you ahev to use
different rules.
Now to get to my point:
In a spreadsheet or any document the same applies.
I spent a about two hours sorting out the formatting of the Styles1-2 final.
Only a few minutes on doing the formulae and less in capturing the data.
I wanted something that would be as readable whether viewing on the screen, or
printing it in colour and monochrome (like a laser).
Some of this, I was able to do with styles, but I could only go so far (compare
Styles1-0, 1-1, 1-2).
I hate sheets where a column as 3 digits in and 5cm wide column because of the
heading. My early experience also taught me that people can interpret very
quickly what the standard stuff is ... if your are consistent ... so, if you do
headings and other standard stuff even in 7 or 6 point, that is quite OK.
The thing which changes every time is the content, and that is also why they
are reading the document. This needs to eb the clearest, least cluttered and
most emphasised.
So, what are the format issues that count?
1. Being able to format the contents according to the type of data. I you look
at my styles, you will see the ones that I tend to use the most.
2. Varying the borders - having thick borders around unshaded cells tends to
overpower the contents. But having thin borders around darker backgrounds
(shading) tends to make the lines almost invisible.
3. Being able to distinguish levels when totalling - so I colour code them
suing graded colouring (lighter to darker)
1. is quite acceptable is it is currently implemented.
2. is a problem because if I want to have the same content format but different
borders, then I have to create sub-styles for each variant. However, there
could be a simple solution here: if I apply a style, then I can override it at
the cell level. If I leave a border greyed, then it will only change the ones
that are not greyed.
However, if I first draw borders then apply a style, even if the style Borders
tab has the borders greyed then it will inherit going back in steps to the
Default style. Which is consistent, but it means there is no easy way to vary
the borders.
3. In the case of background, there is not even the greyed option. No Fill
basically means fill with white or what ever the background colour is in the
template or application.
The sequence should be (as it seems to be):
1. Application defaults, overridden by
2. Application Tools default overrides (typically stored in template),
overridden by
3. Default style, overridden by
4. Applied style, overridden by
5. Direct cell formatting
(there seems to be a level between 2 and 3 which happens for the current
session if I change the settings using the toolbar.)
There seems to be an easy solution:
Consistently utilise the Y/N/empty binary values, where empty is equivalent to
greyed.
If Y - then apply the format
if N - do not do anything
if grey - inherit
Then if one added a tickbox into any tab, you could determine whether to apply
the format or not.
You could make it easy to see if a tab applies or not without having to go into
it if the tab name text were to be white.
Then the tab contents could be determined by the greyed feature.
MS Excel does a partial implementation of this. They present you a list when
you modify a style where you can tick boxes to indicate if you want to apply
the different format sets. But in typical MS fashion, this seems to be more an
afterthought than a design issue.
--
You are receiving this mail because:
You are the assignee for the bug._______________________________________________
Libreoffice-bugs mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/libreoffice-bugs
