OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

office message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]


Subject: Re: [office] list-override proposal


On Wednesday 07 March 2007 09:59, Florian Reuter wrote:
> Hi Thomas,
>
> I guess we're getting closer to a common understanding.

Great!
Do you understand my explanation of the idea to add a property to the 
numbered-paragraph and another property to the text:list to make them more 
equal ?

> So --- sample 5 seems to express a main difference in understanding.
> However my example is not very good. Unfortionately it can be read in two
> different ways :-(
>
> I changed it in the attachements.
>
> So what we have is
> 1. A [L1, continue=false]
> 2 .B [L1, continue=true]
> 3. C [L1, continue=true]
> I.  D [L2, continue=false]
> II. E [L2, continue=true]
> 4. F [L1, continue=*true*]
> 5. G [L1, continue=true]
> I.  H [L2, continue=false]
> II. I [L2, continue=true]
>
> So the continue=true statement on paragraph "F" actually continues the list
> identified with the style "L1" (and not "L2").
>
> So it this our main difference in understanding?

The main difference is probably the auto-continue idea in numbered-paragraphs.
I'll try to show why I said you almost never need continue with 
numbered-paragraphs.

The rule;
A numbered paragraph automatically continues the last numbered-paragraph on 
that same level using the same style.  If the previous has a different style, 
then a new list is started (i.e. it does not continue)

Which means that paragraph "F" would try to continue paragraph E but fails 
because of the different style. And will start a new list.
Oliver made a proposal to change this in 1.2. The change, when accepted, would 
continue from the last that has the same style. Which means your logic 
_would_ work in 1.2 (again, if that proposal is accepted)

Do note that your text:list sample5 has the same problem. It will also only 
work when that proposal is accepted.

That means that apart from my example using list-id, there would be a second 
way to encode sample5. As you did in the updated pdf.

There is a big BUT here.
Every time you use continue=false, you are effectively starting a new list. 
Not adding to the previous list.

See the attachment which shows 3 of the lists, and all of them have a 
different left-indent.  The B & C share indent, the C&D share indent, but 
they don't share the indent of each other. That is because they are different 
lists. (I know; you may need to magnify to see the difference in this 
example)

This means that using the list-id proposal changes one thing pretty 
fundamentally.
list-id:  all paragraphs make up one list.
continue: each markup is a different list.

So, this little detail is important on two areas;
1) data in the odt is not limited to viewing purposes, its also useful for 
datamining and conversion. There is a real difference in using 3 lists instead 
of 1.
2) Each list can have a different implied maximum-width.  See;
http://www.koffice.org/kword/pics/200701-Counters.png
Notice how the two lists at the bottom of the left column (starting 
with 'roman numbering') have the same left-alignment point for the start of 
each paragraph. But the two lists don't share that indent.
They have this because of the widest list-label for each of the two lists 
being different.
The effect of cutting one list into 3 lists means the alignment of the 
paragraph-content goes out the window.

Bottom line;
a) continue numbering is something I don't like. It does not encode the actual 
data, just a way to display it.
b) list-id is a required extention that can replace almost all usages of the 
continue-numbering for numbered-paragraphs.

Cheers!
-- 
Thomas Zander

aligned.png

PGP signature



[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]