Hi Mike,
   Thanks for the comments. They're helpful.

On Fri, Mar 15, 2019 at 2:12 PM Michael Kay mike@xxxxxxxxxxxx <
xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx> wrote:

> I can't think of any persuasive use cases. If your original data fits in
> memory, then there is little reason to make a snapshot copy of it, you
> might as well use the original. The point about snapshot is that it is
> designed to capture the content of a node and its immediate "relatives"
> (the ones you are most likely to need access to) without the memory
> overhead of holding the whole document; so it's all about memory use.
>
> I guess there could be cases where the output of snapshot() happens to be
> exactly the data that you want to put into a serialized message and send to
> some other system, but it's more likely in general that you would want to
> control the message content more precisely than this.
>
> Michael Kay
> Saxonica
>




-- 
Regards,
Mukul Gandhi