W3C home > Mailing lists > Public > public-change@w3.org > May 2013

RE: Deleted content in situ or separate? (was: Generic Change Tracking draft spec)

From: Dennis E. Hamilton <dennis.hamilton@acm.org>
Date: Fri, 31 May 2013 10:28:58 -0700
To: "'Robin LaFontaine'" <robin.lafontaine@deltaxml.com>, <public-change@w3.org>
Message-ID: <000201ce5e24$576a2210$063e6630$@acm.org>
I think there are three flavors of this:

1. Treating the main text as an “after-look” where any deleted material and other information about former states of the document is elsewhere.
2. Having the main text be a “before-look” and changes to be applied to it are kept on the side.
3. Having some composite where transformations are mingled with the text and it is necessary to be able to unravel that to get to the after-look or any preceding looks.

There are hybrids. For example, the original ODF approach provides an after-look with in-line markers so that there can be coordination with the subsequent changes, especially deletions, but all changes in fact, because the change entries also carry metadata and comments.  But it is easy to ignore the markers to present the after-look.  

One use-case has to do with whether or not the material is authenticated in some manner, including being digitally signed.  If one wants to be able to amend a document while preserving the authenticity of the starting point, (2) is appealing.  (1) is acceptable if it is possible to perfectly restore to and verify the signed previous state.  (New states can be signed as well, using a variety of approaches.)  My real-world version of this is contract offers and counter-offers made by revision and round-tripping of the contract document.


I think deletions in-stream are an example of case (3).  You need to decide what use cases become more difficult, if not entirely impracticable, and how tolerable that is.  At the moment, I am preoccupied with ways to ensure verifiable digital signatures on different stages of the material.  I'd like to make that as straightforward as possible, since it is challenging to get signing of XML documents right.

Another technical consideration is with respect to cross-cutting deletions.  (E.g., the end of one element content and the beginning of the following element content are deleted.  So an end-tag and a start-tag are caught in the deletion, and the effect extends to the end tag whose start tag is removed.  Done properly, that probably also means there are attributes on the surviving start tag that also reflect something about the end tag having been transformed.)  How one represents all that “in-line” becomes interesting. 

 - Dennis

 -- Original Message --

From: Robin LaFontaine [mailto:robin.lafontaine@deltaxml.com] 
Sent: Friday, May 31, 2013 02:31 AM
To: public-change@w3.org
Subject: Deleted content in situ or separate? (was: Generic Change Tracking draft spec)

One of the issues we came across was where to hold deleted content. There seem to be two choices:

1. In situ: Keep it where it is and mark it as deleted, e.g. by wrapping it in an element

2. Separate: Move it somewhere else and put a marker in to show where it was originally

It would be good to have any comments on this. I think that in situ seems simpler and avoids the dreaded ID or similar markers which can make life more complicated when processing.


-- -----------------------------------------------------------------
Robin La Fontaine, Director, DeltaXML Ltd  "Experts in information change"
T: +44 1684 592 144  E: robin.lafontaine@deltaxml.com      
Registered in England 02528681 Reg. Office: Monsell House, WR8 0QN, UK
Received on Friday, 31 May 2013 17:29:21 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 19:11:22 UTC