[
Lists Home |
Date Index |
Thread Index
]
Bill de hÓra wrote:
> Playing about with RNG compact syntax, I've come to the conclusion that
> it's to harder to read than the XML syntax.
There are ways to write it that are readable and ways to write it that
are not so readable. The style that's most readable in one syntax isn't
necessarily most readable in the other syntax. In my experience, real
world schemas tend to make very heavy use of definitions/references,
partly for modularity and partly because with RELAX NG, as with DTDs,
redefinition is the main customization mechanism. If you look at
Docbook for example, the number of <ref> elements is much greater than
the number of <element> elements. For this reason, RELAX NG compact
syntax was designed to optimize the readability of schemas that use
definitions/references heavily. The examples in the compact syntax
tutorial were automatically translated (by trang) from the examples in
the XML syntax tutorial and haven't been optimized for readability in
the compact syntax (this should probably be fixed).
> Consider the + quantifier below:
>
> element html {
> element head {
> element title { text }
> },
> element body {
> element table {
> attribute class { "addressBook" },
> element tr {
> attribute class { "card" },
> element td {
> attribute class { "name" },
> mixed {
> element span {
> attribute class { "givenName" },
> text
> }?,
> element span {
> attribute class { "familyName" },
> text
> }?
> }
> },
> element td {
> attribute class { "email" },
> text
> }
> }+
> }
> }
> }
>
> when you could have this:
>
>
> element html {
> element head {
> element title { text }
> },
> element body {
> element table {
> attribute class { "addressBook" },
> element+ tr {
> attribute class { "card" },
> element td {
> attribute class { "name" },
> mixed {
> element? span {
> attribute class { "givenName" },
> text
> },
> element? span {
> attribute class { "familyName" },
> text
> }
> }
> },
> element td {
> attribute class { "email" },
> text
> }
> }
> }
> }
> }
The quantifier after the keyword is an interesting idea, but it weakens
the similarity to normal regular expressions. And the example still
isn't all that readable. A better solution is to use a style more
suitable for the compact syntax. I would probably write it something
like this:
start = html
html = element html { head, body }
head = element head { title }
title = element title { text }
body = element body { addressBook }
addressBook = element table { addressBook.class, card+ }
card = element tr { card.class, name, email }
name = element td { name.class, mixed { givenName?, familyName? } }
givenName = element span { givenName.class, text }
familyName = element span { familyName.class, text }
email = element td { email.class, text }
addressBook.class = attribute class { "addressBook" }
card.class = attribute class { "card" }
name.class = attribute class { "name" }
givenName.class = attribute class { "givenName" }
familyName.class = attribute class { "familyName" }
email.class = attribute class { "email" }
> Also, attributes don't need to be bracketed; by nature they don't have
> any structure on the rhs.
Although an attribute cannot contain attributes or elements, the pattern
for the value of an attribute can be quite complex:
attribute value {
list { xsd:int { minInclusive = "0" maxInclusive = "100" }+ }
| "#none"
}
James
|
