Vernacular

HTML Made Special

Hard to Memorise URLs

This applies mostly in case you’re defining a vocabulary for use with RDFa/Microdata.

Good language design should make it harder for people to make mistakes. One of the most basic part of that is using regular, easy to remember identifiers so as to avoid typos.

With that in mind, something in the following non-exhaustive list should strike one as wrong:

The W3C used to be an easy target in this area (they later switched to a better naming scheme, but most namespaces were minted before that), but they are by no means alone:

By contrast, consider the case of schema.org which is called “schema.org”, resides at http://schema.org/, and uses http://schema.org/ as the prefix for all of its types and properties. Granted, you still might hesitate between http and https, but overall you’re on pretty safe grounds where mnemonics are concerned.

There are arguments in favour of impossible to remember URLs, but as anyone who’s had to produce an XSLT style sheet outputting XHTML, SVG, XLink, XML Events, and MathML can attest these arguments are not grounded in pragmatic reality. They are too easy to get wrong, and — in part due to broken tools — lead to bugs that are sometimes difficult for users to uncover.

↖︎ Back to list