Calc style management

users
#1

Dear all,

we're working with pretty complex spreadsheets in a project and I've
been pushing for the strict adherence to styles (rather than direct
formatting). But I am approaching levels of complexity that make it
really hard for everyone, and I would love to hear some advice on
what I am doing wrong or how our approach could be improved.

Say you have a simple table with row 1 being a "header" row, row 20
a "results" row, and "plain" rows between, column A holding "dates"
and column B "percentages".

Currently, we have 6 styles to represent all combinations, and after
formatting the entire column A as "dates-plain" and column B as
"percentages-plain", we need to format A1 as "dates-header", B1 as
"percentages-header", A20 as "dates-total", and B20 as
"percentages-total". If ever column A or B needed reformatting (e.g.
because rows were pasted wrongly), then the manual overwriting of
the cells in rows 1, 20 and between would need to be redone.

Is there a better way to approach this? Can styles be somehow
combined, such that cell A1 would simultaneously be a "header" and
a "date" cell, given that the two styles are perpendicular and
don't really affect each other?

Thanks for any insights!

#2

Dear all,

we're working with pretty complex spreadsheets in a project and I've
been pushing for the strict adherence to styles (rather than direct
formatting). But I am approaching levels of complexity that make it
really hard for everyone, and I would love to hear some advice on
what I am doing wrong or how our approach could be improved.

Don't be too picky. Concentrate on the cells that need constant editing.
There is nothing wrong when you apply some hard attributes to fixed
(protected) content.

Say you have a simple table with row 1 being a "header" row, row 20
a "results" row, and "plain" rows between, column A holding "dates"
and column B "percentages".

Blending data, appearance and calculation is the major weak point of all
spreadsheet tools. Spreadsheets are quick and dirty tools. Quick and
dirty tools are expert tools and never "fool proof" by any means. All
the meticulousness invested by millions of "Excel experts" is a waste of
time. A single copy and paste or the unforeseen error of an untrained
user may override everything.

Currently, we have 6 styles to represent all combinations, and after
formatting the entire column A as "dates-plain" and column B as
"percentages-plain", we need to format A1 as "dates-header", B1 as
"percentages-header", A20 as "dates-total", and B20 as
"percentages-total". If ever column A or B needed reformatting (e.g.
because rows were pasted wrongly), then the manual overwriting of
the cells in rows 1, 20 and between would need to be redone.

Is there a better way to approach this? Can styles be somehow
combined, such that cell A1 would simultaneously be a "header" and
a "date" cell, given that the two styles are perpendicular and
don't really affect each other?

... and if you include borders, you have percentages-plain-left,
percentages-plain-upper. percentages-plain-lower, percentages-plain-right

#3

Hi :slight_smile:
Have you tried a dedicated spreadsheet program such as Gnumeric? It might
be worth a try, just for yourself at first.

It probably wont solve this specific problem but a dedicated tool sometimes
trumps something that is combined with many other tools. A "swiss army
knife" is brilliant but a carving knife is better for some things sometimes
and a proper screwdriver is better at others sometimes.

http://www.gnumeric.org/

Regards from
Tom :slight_smile:

#4

Yes, I did. Did you too?

#5

Currently, we have 6 styles to represent all combinations, and after
formatting the entire column A as "dates-plain" and column B as
"percentages-plain", we need to format A1 as "dates-header", B1 as
"percentages-header", A20 as "dates-total", and B20 as
"percentages-total". If ever column A or B needed reformatting (e.g.
because rows were pasted wrongly), then the manual overwriting of
the cells in rows 1, 20 and between would need to be redone.

Is there a better way to approach this?

Maybe.

Is this semantic markup, or presentation markup, or just sommer markup?

Can styles be somehow combined, such that cell A1 would simultaneously be a "header" and

a "date" cell, given that the two styles are perpendicular and don't really affect each other?

What is the difference in the presentation markup between header and date?

jonathon

#6

also sprach Andreas Säger <villeroy@t-online.de> [2015-04-15 17:56 +0200]:

Don't be too picky. Concentrate on the cells that need constant
editing. There is nothing wrong when you apply some hard
attributes to fixed (protected) content.

My experience varies. In one case, for instance, we are preparing
a funding application for a children's theatre to the state and the
document should look consistently formatted (it's been returned to
us before due to formatting inconsistencies… don't ask…). We are
maintaining the document in Git, and even though we all use
Libreoffice 4 on Linux, it's not uncommon that some formatting gets
lost between saving the document on two different computers. I found
that styles mitigate this problem a little bit.

> Say you have a simple table with row 1 being a "header" row, row
> 20 a "results" row, and "plain" rows between, column A holding
> "dates" and column B "percentages".

Blending data, appearance and calculation is the major weak point
of all spreadsheet tools. Spreadsheets are quick and dirty tools.
Quick and dirty tools are expert tools and never "fool proof" by
any means. All the meticulousness invested by millions of "Excel
experts" is a waste of time. A single copy and paste or the
unforeseen error of an untrained user may override everything.

I fully agree. And yet, fundamentally, I don't see why this couldn't
be addressed. Data and calculation are perpendicular aspects and you
can think of them as "content" and completely ignore everything
regarding formatting about them. I.e. at the content level, it's
just cell references, data, and formulae.

The problems seems to be on the formatting ("appearance") layer.
From a user's perspective, I would currently conclude that each
cell's appearance depends on (a) the formatting attributes defined
in the (single) style applied to the cell, and (b) the formatting
attributes applied directly to the cell.

The solution IMHO would be to "simply" allow multiple styles to be
applied to each cell, ideally even from different sources. With
this, the appearance would depend on the attributes derived from

  1. styles applied to the column
  2. styles applied to the row
  3. direct formatting applied to the column
  4. direct formatting applied to the row
  5. styles applied to the cell
  6. direct formatting applied to the cell

The order of column-before-row could probably even be made
configurable per-sheet. Other than that, the merging would be
trivial: attributes obtained later in the above series would
override attributes set earlier, and each attribute could either be
set (forced) to on or off, or unset (in which case it's transparent
and gets ignored).

also sprach toki kantoor <toki.kantoor@gmail.com> [2015-04-16 02:19 +0200]:

Is this semantic markup, or presentation markup, or just sommer
markup?

I am not entirely sure what you are differentiating between. Some is
semantic (e.g. "dates"), some is presentation (e.g. "header").
I don't know what "sommer markup" is.

What is the difference in the presentation markup between header
and date?

Nothing. A date field means the data should be rendered as a date.
A header field means that the font should be e.g. bold on a gray
background. And there can be header-date fields, meaning that data
should be rendered as a date, in bold, and on gray background.

Makes sense?