OASIS Open Document Format for Office Applications (OpenDocument) TC

Hi,

so for the records. My problems with the conversion algorithm is:
a) It is non-deterministic
b) It does not addess all attributes related to numbering (e.g. continue, restart, list-style-name, ...)
c) No sample; so I couldn't very my understanding  :-)

At the end of the day I am insisting on a normative algorithm, because I want to know whether ODF will contain two
different ways of doing numbering or one conceptual way of doing numbering with two different textual representations.
I want this conversion algorithm to be easily implementable. So e.g. not requirering global knowledge of the document,
which would e.g. a SAX-based transformation impossible.

I personally have not figured out a way to write a deterministic conversion algorithm. I guess the choicce of what list
style will be related to the list-id is always a non-deterministic choice.

E.g. 

On Monday 12 March 2007 17:32, Florian Reuter wrote:
> I personally have not figured out a way to write a deterministic conversion
> algorithm. I guess the choicce of what list style will be related to the
> list-id is always a non-deterministic choice.

[snip examples]

You end up with 3 different ways of doing it mostly because you are mixing 
proposals.

The proposal that Oliver sent out was to have the list-override on the 
text:list-item. Not on the text:list.

Try doing it like that and you'll see its a lot easier already.
-- 
Thomas Zander

Florian Reuter wrote:
> Hi,
> 
> so for the records. My problems with the conversion algorithm is:
> a) It is non-deterministic
> b) It does not addess all attributes related to numbering (e.g. continue, restart, list-style-name, ...)
> c) No sample; so I couldn't very my understanding  :-)
> 
> At the end of the day I am insisting on a normative algorithm, because I want to know whether ODF will contain two
> different ways of doing numbering or one conceptual way of doing numbering with two different textual representations.
> I want this conversion algorithm to be easily implementable. So e.g. not requirering global knowledge of the document,
> which would e.g. a SAX-based transformation impossible.
> 
> I personally have not figured out a way to write a deterministic conversion algorithm. I guess the choicce of what list
> style will be related to the list-id is always a non-deterministic choice.
> 

The main purpose of the proposed (semi-)algorithm was, that the 
introduction of the *new* attributes will not break the possibility to 
convert between both structures under the assumption that both 
structures can be convert into each other in ODF 1.1. Thus, it's not 
complete.
The other purpose of this algorithm was, that you requested it.

Regards, Oliver.

> ...