[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [Elist Home]
Subject: Re: combine
> By the way, what is the practical advantage of making RELAX NG order > insignificant. The reason I don't like it is that when you use order to resolve conflicts, you can't tell when order is being used to resolve conflicts and when it is insignificant. For example, consider the situation now where inclusions are significant. Now imagine that I have ten <include> elements. Typically, the relative order of only a few of these is significant. But there is no way for a schema author to say that the relative order of these two is intended to be significant, but the relative order of these two is not intended to be significant. This makes it hard to maintain. Somebody modifying the schema doesn't know whether changing the order will break things. It also means that errors may go undetected. Suppose schema A and schema B in fact conflict but the schema author is not aware of this. When the schema author includes schema A and schema B, then one will necessarily be included before the other, and so one will override the other without any error being detected. If the schema author could accurately express his intent, the validator would be able to report an error. James
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [Elist Home]
Powered by eList eXpress LLC