Re: [office] OpenDocument lists - my view included are some proposals

[Oliver-Rainer Wittmann - Software engineer - Sun Microsystems Inc<Oliver-Rainer.Wittmann@Sun.COM>]

Thomas Zander wrote:
> On Friday 09 February 2007 11:31, Oliver-Rainer Wittmann - Software Engineer - 
> Sun Microsystems wrote:
>> ODF lists defined by using list blocks - <text:list> elements:
>> - Each top-most list block denotes a certain list and defines the
>> counter domain for its including list items and sub lists. 
> ok

>> The list 
>> style, which will be applied to this list, can define start values for
>> each list level.
> I don't agree to that.
> A child list has its own style which can specify the start value. The child 
> list-style may be inherited from the parent.
> This means that your statement is only true for some cases (when the child 
> list does not have a list-style) but not for all.
Can you please give definition for the following terms:
- "child list"
- "child list-style"
- "inherited from the parent"
I can imagine, what you mean. But, if you want me to be precise, I 
except that you also try to be precise.
My interpretation for these terms are the following:
- "child list" == sub list
- "child list-style" == list style specified for a sub list
- "inherited from the parent" == Inheritance of attributes from the 
list style of list level x to the list style of list level x+1. The 
same that I tried to express in my view about lists.
Does my interpretation meets your intentation?

Ok, I can also support your view on the start values for a list.
Repeating: We would define, that each list block can define its own 
start values via its list style. If a list block doesn't define a 
start value, the start value of the surrounding list block is used.

> 
>> - Each list block restarts the counter of the list level, the list
>> block belongs to.
> Could you explain this? This is not a full English sentence.
Imagine the following (meta) structure for a certain list:
<text:list>
	<text:list>
	</text:list>
	<text:list>
	</text:list>
</text:list>
We have two list blocks on list level 2. Each of these list blocks 
restarts the counter for the list level 2. These list blocks belongs 
to list level 2. Restarting the counter for a certain list level means 
to set its value to the defined start value.
> 
>> - Each list block can have its own list style and the complete list
>> level definitions of this list style are applied to its list items.
>> E.g., a list block on list level 3 defines an own list style containing
>> list level definitions for list level 2 and 3. Thus, these list level
>> definitions are used, if needed, to build the list label of its list
>> items and to layout its list items.
> Hmm, you sure?
Yes.
> 
> So you actually think we should have
>  1. head
>  a.a head 2
> if the list-level 2 has a different definition for the level 1?
Yes.

> Who would want that?
I don't know. But, I'm not in the position to imagine, what each user 
of the OpenDocument file format wants to do.
Thus, I decided to give the user the freedom to decide on its own, if 
she/he wants to override the list level definition of a superior list 
level or not.
> 
> No, I think that if the level 3 uses a list-style thats different from the 
> other levels then that is used for the formatting (character style) and the 
> counter, but if it shows counters from parent list-styles then they should be 
> inherited from the parent style without change.
> So;
>   1. Head
>   1. a  Head2
> where the head2 is build up from:
> *  "1."  the counter inherited from the parent item. The counter-style, start 
> value and prefix/suffix are inherited as well.
> * "a"  the local style determines the counter style and the start value, 
> prefix suffix.
> * "Head 2" the parag text.
I think the user should decide - as I already stated above.

> 
>> - Each list item can restart the counter of its list level with a
>> certain value.
> ok
> 
>> - Each list item can override the defined list style of its list block
>> by the new text:style-override attribute. The complete list level
>> definitions of the overriding list style is used to override the
>> definitions of the overriden list style. If the overriding list style
>> only contains the definition for one list level, only the list level
>> definition of this list level is overriden. Thus, the list style, which
>> is applied to the list item is a somehow a list style, which is formed
>> by taking the list style of the list block overriden by the attributes
>> and elements of the list style given in the text:style-override attribute.
> 
> Please write shorter sentences, your writing gives me a headace. ;)
> If I understand correctly this is the same as 2 points back. Which I disagree 
> with.  I do think that we brought up that topic a couple of times on this 
> list in the last month as well.
> Reusing the text from the parent list was the concensus.
Then I have missed this consensus. Where can I find the statement 
about this consensus.

> 
>> - A list block can also continue the numbering of its preceding list
>> block without restart by using attribute text:continue-numbering - see
>> chapter 4.3.1 of ODF specification 1.1 about list blocks. 
> 
> Yeah, I read that but still am quite at a loss as how to interpret it.
> Can you perhaps give a (useful) example?
>
Imagine the following part of a document:
...
1. P1
2. P2
3. P3
4. P4
...
- enumeration item 1
- enumeration item 2
- enumeration item 3
...
Two lists are present in this part of the document - a numbered one 
and a bullet one. Now imagine what happens, if the user cut-and-paste 
the bullet list between "2. P2" and "3. P3".
I expect the following result:
...
1. P1
2. P2
- enumeration item 1
- enumeration item 2
- enumeration item 3
3. P3
4. P4
...

The numbered list is "separated" by the bullet list.
Yes, you can use <text:numbered-paragraph> and the proposed attribute 
list-id to represent it in the OpenDocument file format. But, I want 
to keep, that each list represented with <text:numbered-paragraph> can 
be expressed with <text:list> and vice versa. Thus, the above 
"separated" numbered list should be also represented by using 
<text:list>. That, why I proposing the extension.

The simple case, which is currently covered by the specification is, 
that a list is separated by normal paragraphs.

> When I read it I thought that this was an indication that the list was not 
> properly created. If you want to continue numbering, then merge the lists. 
> That's why we have both the text:list hierarchy (which makes up a whole list) 
> and the numbered-paragraph with a list-id. Both make it clear that a set of 
> paragraphs together make a list, and thus should have consecutive numbering.
> I.e. I don't see a reason to have a feature that allows you to combine 
> separate lists into one when you could just do exactly that; by combining 
> different paragraphs into a list.
> 
>> For sub lists 
>> everything is clear. But for top-most list blocks I want to extend the
>> ODF specification: The current specification states, that the numbering
>> is only continued, if the preceding list has the same list style. I want
>> to apply this statement not only to the direct preceding list, but any
>> preceding list. Thus, if text:continue-numbering is true, the numbering
>> of a preceding list, which has the same list style applied, is continued.
>> Use case: Think of a list with 3 list items, each list item containing a
>> couple of paragraphs. In the paragraphs of the second list item you now
>> insert a small enumeration using a bullet list. Now, your list is broken
>> into two parts by the bullet list. To "join" these parts together to one
>> list, you can now use the text:continue-numbering attribute.
> 
> I think this is false.
> This just indicates that you should not have used a text:list structure for 
> the non-continues list, but you should have used numbered-paragraphs for the 
> 3 list items.
> On top of that; your example doesn't actually need the proposed extention. It 
> would work fine with the current continue. Wouldn't it?
>
I don't think that I will work this the current specification. The 
current specification only talks about the direct preceding list and 
that its numbering can be continued. I want to extend this to all 
preceding lists.

Regards, Oliver.