- From: Michael Kohlhase <m.kohlhase@jacobs-university.de>
- Date: Sun, 30 Mar 2008 11:18:42 +0200
- To: Henri Sivonen <hsivonen@iki.fi>
- CC: Bruce Miller <bruce.miller@nist.gov>, David Carlisle <davidc@nag.co.uk>, ian@hixie.ch, public-html@w3.org, www-math@w3.org
Henri Sivonen wrote: > > On Mar 30, 2008, at 05:07, Bruce Miller wrote: >> I think there's some confusion here --- tho' it may be mine. >> I personally think the most compelling case for annotations, >> especially in a web context, is to provide presentation MathML >> for display to humans, along with the corresponding content >> MathML (when available) for export to applications (or perhaps >> for audio rendering, or ...). For any non-trivial math, for >> software to infer the meaning from the presentation is really >> just a wild guess; humans do somewhat better. Perhaps the >> notion of a "semantic web" has lost its popularity, but this >> use case seems to be exactly what a semantic, and accessible, >> web needs. >> >> I do somewhat share the concerns about using annotations to embed other >> applications' internal forms, leading to bypassing the MathML >> entirely, and increasing chances of the forms getting out of sync. > > So shouldn't Content MathML be developed to be rich enough to > encompass the semantics products need to round-trip but in a > vendor-neutral way? You could view the work we are doing in MathML3 to do just that. In an ideal world, there would be enough content dictionaries (the new addition to MathML3 from OpenMath) to express all information that is locked up in vendor-specific ways into a vendor-neutral format. This is certainly possible and we forsee that some of this will happen. BUT, encoding vendor-specific information in a new format (CDs) is hard work and would disclose information that some vendors would consider strategic, therefore I am not sure that this will happen for commercial entities. Nonetheless, the possibilitity to do this will be here with MathML3. >> But even so, _if_ there is the means for a browser to skip over a >> content mathml annotation, that mechanism could as easily skip >> other annotations as well --- provided it can find the end of >> the annotation, which of course is not necessarily trivial >> once you're outside of the XML world. > > Implementing skipping is not the problem. The problem is that *some* > products might not skip the annotation thereby getting different data. > That can be bad because the different data is out of sync or because > the different data is better and everyone else is missing out on the > better data. This is actually the *feature* and not the *bug*: some applications that can deal with the better data should be given access to it. We will have to deal with the possibility of data getting out of sync some other way. The proposal of disallowing other aspects (and I have argued that they are non-redundant) of the data is like solving the problem of air traffic control by just allowing one plane in the air at a given time. While the danger of collisions provably goes away, the price we pay in terms of quality of service (reduced traffic bandwidth) is too high. Therefore we propose to keep <semantics>/<annotation-xml> with all of the sync problems. Michael -- ---------------------------------------------------------------------- Prof. Dr. Michael Kohlhase, Office: Research 1, Room 62 Professor of Computer Science Campus Ring 12, School of Engineering & Science D-28759 Bremen, Germany Jacobs University Bremen* tel/fax: +49 421 200-3140/-493140 m.kohlhase@jacobs-university.de http://kwarc.info/kohlhase skype: m.kohlhase * International University Bremen until Feb. 2007 ----------------------------------------------------------------------
Received on Sunday, 30 March 2008 09:19:18 UTC