- From: Dr. Olaf Hoffmann <dr.o.hoffmann@gmx.de>
- Date: Wed, 14 Aug 2019 10:50:06 +0200
- To: www-svg@w3.org
- Cc: stm@pdflib.com
Stephan Mühlstrasser: ... >How is an SVG processor supposed to distinguish a >SVG 1.1 file from a SVG 2 file? The usual suspects do not distinguish anyway, what is in practice already a problem due to a few incompatibilities between the versions 1.0, 1.1 first edition, 1.1 second edition, tiny 1.2. This is similar to problems with different versions of (X)HTML and their incompatibilities (there are far more as in SVG). Because interpretation is in practice not closely related to recommendations anyway, a precise version indication mainly allows to check files. Without a version indication this options becomes meaningless, there are no bugs anymore, not in documents and not in processors, whatever is written and however it is interpreted or presented ;o) Removing versioning is a lost chance to provide digital documents with precisely defined markup an meaning to conserve information for future generations. Such formats like HTML5 or SVG 2 seem to me more intended for disposable information, good for today and maybe tomorrow, but without a meaning for next year, next century, next millennium. But because in the current draft for SVG 2.0 almost all new features found once to be required for this version are removed anyway meanwhile. Why should authors (and implementers) care about SVG 2.0 at all? ;o) And of course, you can use for example Dublin Core Terms within a metadata element anyway to indicate the relation of the document to a recommendation. The same method works for example for HTML5 documents or for EPUB 3 with similar problems. This might be sufficient to get a defined meaning of the markup of a document for the future, as long as the linked recommendations are still available. Alternatively you can simply ignore, that the attribute is removed in SVG 2.0, it still belongs to the SVG namespace. This is called a willful violation of recommendations (as appear for example in HTML5.x recommendations as well), due to 'good' reasons of the violators ;o) Olaf
Received on Wednesday, 14 August 2019 08:50:33 UTC