Cart
[Home] [By Thread] [By Date] [Recent Entries]
Jonathan Borden wrote:
> alternatively, one could wrap the filesystem in a DOM/XPath accessor and let
> the filesystem code perform the access checks for you. I think it would take
> less code to wrap the filesystem *BUT* one could always munge Xerces to
> provide ACL behavior.
>
> My gut feeling is that using a filesystem designed for lots of small files
> will give the proper level of concurrency and access control. Which do y'all
> think would be the most efficient?
IMHO, this will have to be a DBMS, not simply a filesystem. The underlying data
store (which for many reasons should be native XML) will require an enclosing
engine to:
-- manage document versioning
-- maintain orthogonal extensions of the same base, as for example raw data
and various interpretations of it
-- identify various sources of documents with the standard transforms which
must be performed first in the pipeline of processing them
-- identify various output forms with the standard transforms by which they
are rendered
-- (most crucial of all) enable the definition of, and then implement, a
processing model which--by instance document, or by document class, or by user,
or user type, or user role, or by type of request--applies defined links,
transforms and other processing in a rational and appropriate order.
Respectfully,
Walter Perry
|
