I'm not making excuses or even a rationale, simply trying to provide a suggestion as to why so many people use camelCase even when it isn't required.
Underscore is certainly an acceptable character for identifiers in DTDs, RelaxNG, and many programming languages. However, the original poster didn't ask about using an underscore. The inclusion of the "-" and "." as identifier characters in SGML/XML was a somewhat unconventional practice, when considered in the domain of computer languages. When reading works from before XML, such as Goldfarb's annotated SGML reference, one gets the sense that avoiding too close an association between SGML identifiers and programming language identifiers was deliberate. Colin Paul Adams wrote: > > >>>>> "Mitch" == Mitch Amiano <[EMAIL PROTECTED] > <mailto:mamiano%40nc.rr.com>> writes: > > Mitch> No doubt it is that so many people use that convention in > Mitch> programming too. But consider that in many programming > Mitch> languages, zero-or-more would not be an identifier but an > Mitch> expression zero(minus)or(minus)more. So camel case allows > Mitch> you to think of the element and attribute labels as > Mitch> identifiers, without any other mapping. > > That's a very poor excuse, since you can use zero_or_more, which is > even closer to natural english syntax than zero-or-more (in as much as > _ resembles as space more). > -- > Colin Adams > Preston Lancashire > >
