- From: Chris Lilley <chris@w3.org>
- Date: Thu, 15 Jul 2004 17:42:12 +0200
- To: Bjoern Hoehrmann <derhoermi@gmx.net>
- Cc: www-svg@w3.org
On Thursday, July 15, 2004, 12:38:09 PM, Bjoern wrote: BH> Dear Scalable Vector Graphics Working Group, BH> As you are all aware, the complexity and implementation cost of BH> a technology is measured in terms of how many sheets of paper are BH> required to print the specification; I must thus ask you to reduce BH> the complexity of the SVG 1.2 specification in order to enable its BH> widespread adoption on the web. We have produced a microfilm edition in your honor. However, it needs a special machine to read it. BH> Please do all of the following: BH> * Remove all non-essential content such as illustrations, BH> examples, inline schema fragments, membership listings, BH> element/attribute/property summaries, and the TOC BH> * set the font-size in the style sheet to at most 9px BH> * set the line-height in the style sheet to at most 8px BH> * set the letter-spacing in the style sheet to less than 0px BH> * set all paddings and margins in the style to at most zero BH> Testings showed that already the last four steps reduce the complexity BH> and implementation cost of the SVG specification to less than one third BH> and I am confident that the removal of the non-essential content will BH> lead to a further significant reduction. I believe that less than 20% BH> of the original complexity are a realistic goal, enabling SVG to go BH> where no one has ever gone before. ;-) However you make a good and serious point - a specification that has full implementation details, precisely specified conformance clauses, a well specified grammar and abundant examples and use cases will be easier, not harder, to implement and more likely to have interoperable implementations than one that relies on shared unwritten understanding and terse statements of the maybe-obvious BH> For future SVG specification, there is room for even more improvements, BH> you can save a lot of text by less verbose prose descriptions of the BH> technical content. This technique as commonly used by other specs can BH> easily adopted by the SVG specification, for example the ridiculously BH> complicated discussion on filters can easily replaced by a much easier BH> to understand description like, "make it look nice, like in Photoshop". 'Do the best you can, implementations can do whatever they want' wooliness has no place in a technical specification. Content developers want consistent, repeatable results from different implementations. BH> And then you must start modularizing the specification so that you can BH> just reference other parts of the specification^W^W^W^W successful and BH> simple technologies, so that your specification would become *really* BH> simple, I could even think of just a single printed page! All of which demonstrates, most eloquently, that measuring implementation complexity cannot be done with a set of scales and a printed specification. -- Chris Lilley mailto:chris@w3.org Chair, W3C SVG Working Group Member, W3C Technical Architecture Group
Received on Thursday, 15 July 2004 12:04:37 UTC