Re: XML* vs. eXpat

[Joe Pluta <joepluta@xxxxxxxxxxxxxxxxx>]

Jon Paris wrote:
> > The XML-INTO is pure DOM processing; you
> > tell it where in the document to get some data, and it pulls it out,
> > thank you very much.
> >     
>
> Unless I completely mis-understand DOM Joe, XML-INTO is not a DOM  
> parser.
Agreed.  I said it's DOM processing, not a DOM parser.

> The only thing you can tell it to do (via the path=... option  
> in the %XML BIF) is where to start in the document.  But that is a  
> long way from the DOM notion of random access.
I used the term "DOM processing" simply to contrast with SAX processing; 
you can point it to a specific spot in the document and pull out 
specific elements.  With SAX you have to handle every event, even if 
it's just to ignore it.

>   path=... will let me  
> specify to skip a chunk at the beginning before (for example) the  
> section detailing all the invoices - but it won't let me specify to  
> start at invoice X which is how I think of DOM.
>   
Even that description falls far short of a true DOM parser, Jon.  With a 
DOM parser, you can extract individual elements, perform iterative 
searches at different levels of the document, all of that.  Heck, a true 
DOM parser like Xerces lets you add nodes on the fly - just like the DOM 
processing capabilities in JavaScript.

> I do think that XML-INTO is a lot better than folks here have given it  
> credit for - it is very "RPG-like" and works very well.  It is not  
> perfect by a long way - but for straightforward schemas it is  
> wayyyyyyyy easier than any SAX approach.
>   
I agree 100% - it's a lot easier for simple documents.  But really I 
just wanted to point out another example of IBM's compiler teams doing a 
better job than they seem to be getting credit for of late.  I guess 
because I've actually written a parser, I appreciate the work involved.

Joe