Some minor issues with latest editors draft

After a brief scan of the latest draft, I noticed the following minor issues:


Section 2.4.9.1 - typo, says "getGlobalizedTrigger" instead of "globalizedTrigger".


Section 2.5.2.3 - record element, target variable attribute.

Says that the 'variable' attribute can only use the cdl:getVariable function.

Just to clarify - is it therefore ok to specify a documentPath (query) for the target variable? What would happen if the CDL attempted to record information in a subpart of a document that has not been initialized yet?

Noticed that the same situation arises with the assign activity which shows some examples, but all of the target variable attributes do not include the documentPath. However, I could see situations where being able to record/assign information within the contents of a document would be required. Therefore, is this a case of documenting that an exception may occur, or do we leave this implicit?

The other issue with using the 'getVariable' function in this way, is that the description of this function states that it "Returns the information of the Variable with name varName as a node set containing a single node." However, in the context of the target variable attribute, we are actually after a reference to the variable.

My preference would be to expand the relevant fields into attributes on the 'target' node,

e.g. 
<copy>
    <source expression="cdl:getVariable(......)" />
    <target variable="ncname" part="ncname"? documentPath="xpath"? />
</copy>

documentPath and possibly role would be optional/unnecessary depending on context (i.e. the role would be inferred from the assign activity).

This would ensure that the semantics of the source and target are appropriate - i.e. the source is intended to derive information (possibly just using an 'expression' attribute), and the target is responsible for identifying the reference to the location where the information is to be stored.




Regards
Gary

Received on Tuesday, 16 November 2004 15:17:29 UTC