> > > Why not optionnal before and after, then processes querying metadata
> > > header would get a usefull information if available and everybody
> > > trailing checksum kind information.
> >
> > Because XML isn't a serial structure. A tag at the same hierarchy level as
> > TABLEDATA (inside TABLE) can go either before or after TABLEDATA, doesn't
> > matter which since its its the same level tag, but you can't have two
> > copies.
> >
> > For tables where you just want the meta-data the fact that the data in the
> > TABLEDATA is a whole sub-level down means you could just parse the table
> > to a certain depth and mis the actual data blocks, but grab the meta-data
> > telling you how much data there actually is...
>
> It can be tags with diff. names but with ~ same content;
> i.e. RowNumber and RowNumberChecking, or whateverelse.
But XML is a hierarchical, not a serial, format...
Despite the fact that some times (most times?) we will be using an event based API, i.e. SAX, rather than a tree based API, i.e. DOM, it doesn't mean we should break the XML paradigm!
Al. Received on 2003-10-23Z16:06:41