ODULARITY or compartmentalisation, at least in the context of programming, are important and fundamental concepts. They accommodate for easy extension or change of any piece of software. In the context of Web design, elements to be dealt with apart are layout and content, as the former virtually ‘wraps’ the latter.
Design for the Web and typesetting must be treated just like programming. Almost no exceptions. Why rule out WYSIWYG paradigms and interfaces? Because Abstraction is lossy. Since even bloated application can never faithfully read one’s mind, they result in mistaken outcomes which make assumptions that should not be made. The issues are also ambiguity-driven. Below lie a few examples:
Cascading Style Sheets (CSS): What is the motivation? The idea of maintaining one single file to rule them all — a magic wand to manipulation of files, which facilitates consistent change of layout in a variety of pages. Separate content from presentation is the core rationale. This accommodates for flexibility that is inherent in themes. Styles adhere to a given standardised mechanism/syntax, much like API‘s in the programmatic domain.
If you develop a software product, you can have someone else extend your code, add themes or utilise your services, e.g. Google Maps API, which are now trailed by Yahoo equivalents. Also worth noting are the number of Firefox and Thunderbird themes and extensions (plug-ins). WordPress provides a genrous number of plug-in ‘hooks’ as well, which encourages outside contribution. All extensions are attributed to good support to foreign involvement and modularity which separates the trunk of an application from its add-ons.
LATEX: TeX styles are a similar scenario where consistent styling is easily achieved. Change can be applied to entire document where semantic structure has been embedded by the author. This, in fact, is why many books and eclectic conference proceedings are assembled using LATEX.
Fortunately, more and more people begin the comprehend the value of this approach and the mantra which is not only inter-operability-motivated. It is also guided by the emerging importance of structural composition of documents, which must not differ. In the absence of serapate styles, many files need to be changed consistently, which can become a labour-intensive task. Such operatings are possible, e.g. using recursive search and replace, yet they should be considered nothing but a fallback option.