This is a nice-to-have, but it came up at JATS-Con, so I want to record it here.
It would be nice if the messages that get reported could be interleaved with the actual source document. Off the top of my head, the one way I can think to do this would be to re-serialize the document on the fly, and keep track of where we are in the dom tree. When the xpath expression in the error message matches the node we are on, insert the message there.
The (big) drawback of this approach is that the serialized version of the document won't match exactly what they uploaded. There is no way for us to know, post-parsing, small lexical things like whitespace and, for example, whether single- or double-quotes were used for attribute values. Users might be confused by this.
Klortho
commented
9 years ago
This is a nice-to-have, but it came up at JATS-Con, so I want to record it here.
It would be nice if the messages that get reported could be interleaved with the actual source document. Off the top of my head, the one way I can think to do this would be to re-serialize the document on the fly, and keep track of where we are in the dom tree. When the xpath expression in the error message matches the node we are on, insert the message there.
The (big) drawback of this approach is that the serialized version of the document won't match exactly what they uploaded. There is no way for us to know, post-parsing, small lexical things like whitespace and, for example, whether single- or double-quotes were used for attribute values. Users might be confused by this.