Stephen E. Beller wrote:
> 
> I know how to (a) pre-arrange large data sets into particular arrays needed
> for rendering graphs, (b) store these arrays in a delimited text file (e.g.,
> CSV) for rapid transport, (c) quickly open the file client-side in a
> spreadsheet and (d) immediately render lots of offline-interactive
> (slice/dice-able) charts without intermediate data transformation. This is a
> very fast, simple, efficient process.
> 
> I'm trying to do a similar thing within an XML file for client-side browser
> rendering, but don't want the overhead of tags and XSLT parsing if possible.
> 
> Is there a way to do this?

You request is not clear yet.  If you do not want to use xml, then why 
are you using xml?  Even if you do use xml, you do not have to use xslt, 
although it might turn out to be useful anyway.

The overhead is only significant if it incurs a perceptible delay for 
the user, right?  Is your data set so big that this happens now? 
Anyway, how are you rendering charts in a browser?   Or is it not charts 
and graphics that you want to do?

Assuming that you are handed some kind of xml file, it is presumably 
record oriented, so some kind of SAX or SAX-like handling would be 
indicated.  That would be fast and should lend itself to similar 
handling to your previous files, in that each row could be quickly and 
easily split up into fields.

If you are going to use IE or Mozilla, each of them give you some access 
to their xml parser.  Otherwise, if you know your source xml data will 
always be in a simple format that does not use any of the more complex 
features of xml, there are a few javascript xml parsers out there that 
would probably work.  They do not implement all xml's features, but that 
would probably not matter for your purposes.

Your approach will probably also depend on whether it needs to work 
cross-browsers or not, and which browsers need to be supported.

Cheers,

Tom P