Ajith Ranabahu wrote:
Hi there,
Just as Chinthaka was saying OMAttribute was a child of the OMNode
(OMNamedNode to be exact). However this was changed later with the
immutable attribute concept (Samething that happened to OMNamespace).
Since the removal of the OMAttribute from the hierarchy some other
changes are still pending
1. OMNamedNode is not needed now
2. Certain Node types are not used anymore
so it owuld be good to remove any mention of unused classes from code.
However I am not very happy with this change. Yes, it may have reduced
the size but OMAttribute is no longer a node now. This means that it
cannot be put into a node link list and we are stuck with old
ArrayLists to hold the attributes!
you could use LinkeList or any *List - ArrayList is pretty efficientthough
My guess is we should think of these changes in a more serious manner
so that our structures remain consistent and elegant
interfaces (API) is what matters and if they can be implemented os
performence is good andeven more important memory footrpint is not too
big (compared to DOM ...)
alek
On Tue, 22 Mar 2005 09:06:52 +0600, Eran Chinthaka
<[EMAIL PROTECTED]> wrote:
Hi Ashutosh,
Sent this mail, but could not see it delivered, so resending. My
apologies if you receive multiple copies.
Hi all,
I see that we have a separate OMAttribute class, and OMAttribute is
treated differently from other node types.
OMAttribute was a derived class of OMNode, earlier. That was a wrong
decision and later we changed that to the current state. There was a thread
on that in this mailing list.
But, at the same time in
OMNode class also we have a field to set the Node Type as
ATTRIBUTE_NODE.
I think we should remove that. This was a thing slipped through the last
change.
-- Chinthaka
How are these two different and why can't we just go
with the OMNode class?
Ashutosh
--
The best way to predict the future is to invent it - Alan Kay
