[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: Re: [office-collab] Compromise on change-tracking proposals
Robin, You are right, it is best to ask the TC for permission. There is no need to for discussion if it is a compromise or a new proposal, still I would like to explain why I think it would be a good compromise. From my understanding the ECT was submitted as the elegance of the simplicity of the GCT is only granted for XML model based ODF application. Applications like OpenOffice or MS Office 'need a good K.I. to find the mapping from the XML to their model' to quote one of their developers. ECT takes therefore the document level approach, but neglects the beauty of a generic approach. Allow me a little magic trick. We start with chapter 5.2 of your latest updated draft: 5.2 Edit-operationIt may be difficult for an editing application to determine the editing operation that was applied in order to achieve a particular set of changes represented by a CT, or CT group or a CT stack. Therefore editing operations can be defined in the standard and identified using the delta:edit-operation attribute. For each definition of an edit operation, there will be rules to say what type of change is allowed. For example, an edit operation defined as a 'text-replacement' would not be allowed to include the deletion of a paragraph, or the change of a format attribute. This mechanism could easily be extended so that a given editing application which has an operation that is not defined in the standard, would be able to create a new definition perhaps using its own namespace as a prefix. You say there might be a label for the operation, as it might be difficult for an editing application to identify the editing operation.If the label is an addition that sometimes help, why not always providing it to enhance usability? Making it mandatory to be able to trust upon it? And we should be also aware that those labels have to be standardized, otherwise Apache OOo might label "deleteColumn", while LibreOffice uses "columnDelete". A potential nightmare for interoperability. So if we agree to standardize all labels of operations, our two proposals are close to equal. The difference is: I neglect the redundant verbose XML change, only keeping the CT labels in the document, move them out of the content into their own file "undo.xml" and specify the XML change of each label on TC level in a specification. Saving a lot of space in the document and allowing ODF applications with no XML model to deal with CT on a higher abstraction level. In the end I only refactored your GCT approach. The generic design came from GCT, the high level requirement from ECT, therefore I was mentioned it to be a good compromise. Best regards, Svante Am 12.09.2011 17:37, schrieb Robin LaFontaine: Svante, |
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]