Friday 09 May 2003 10:22:36 am
I understand that some people don't like the "CamelCase" syntax, but if you generalize a bit the idea is very useful. In fact many wikis today allow simple delimiters to designate a named object reference e.g. ((name of object)). Others allow nested namespaces e.g. ((start object name : child name : grandchild name)). [my comments below are based on incomplete understanding of eZ, so please bear with me ...] The main intent is easy creation of links between objects, which requires easy ways to refer to other objects. Lots of applications want ways to easily link objects. If some object structure is present (as in eZ), it can be used as a namespace, and we can use object names or attribute names to easily refer to objects e.g. without having to go through URL details. Moreover, with an eZ-like explicit object structure, one might even be able to refer to paths to objects by sequences of attribute names or object names [e.g. xPath-like names]. Also there is no intrinsic reason why the target of a reference must exist before you create the reference. This is one of the wonderful things about Wikis, but the idea is much older. e.g. only the most restrictive programming languages require definition (exist) before use (create reference). The better ones allow declaration (assert existence) before use; and the most flexible ones allow easy mixing in an incremental and interpreted environment, with the option for checking dangling links etc. This is where Wiki's "pending reference" idea is extremely helpful to the smooth and fluid creation of an interlinked object graph. Thoughts?
|