- From: Jon Bosak <bosak@atlantic-83.Eng.Sun.COM>
- Date: Thu, 1 May 1997 03:34:07 -0700
- To: www-style@w3.org
[David Perrell:] | > ... most people would rather have the enforcement come | > from the machine. I'm speaking from direct experience of how | > publishing works in large corporations and the kinds of policies that | > have to be put in place in the absence of software support for | > enforcing stylistic consistency. | | Be they cascading or parameterized, stylesheets can be kept under | central control. But can a stylesheet be an enforcement mechanism? I | thought external stylesheets were about convenience and efficiency. | Without the cooperation of authors, no stylesheet will "enforce" | stylistic consistency. But with cooperation, they will "facilitate" it. You can't prevent anyone from doing anything. So what. Every medium- to large-size corporation invests major money in creating a corporate personality for their public documents and puts management systems in place to ensure that content producers within the company follow the style guidelines. Anyone can, in theory, ignore these controls, and then they will, in fact, get fired. The point is that some methods of specifying style constraints work more gracefully with this model than others. The parameterization concept should work more gracefully with the requirement for corporate style control than the cascading model because it allows permitted changes to the corporate style to be provided as a kind of API to the stylesheet and makes clear to authors the distinction between what can be changed and what can't be changed that in the cascading model would have to be provided as a set of verbal policies. There is nothing to prevent corporate style from being controlled through verbal policies; that's what we've been doing for quite a while. But it's been my experience that authors find it easier to work with policies when they are built into an interface than when they have to be memorized. All things being equal, it should be easier to build policies into an interface using a parameterization model than using a cascading model. Jon
Received on Thursday, 1 May 1997 06:34:33 UTC