Valentina 2.0. -- What is your 3 DREAM features?

[Ruslan Zasukhin]

on 1/31/04 6:42 PM, erne at ernestogiannotta at tiscalinet.it wrote:

>>> editable multi-table cursors,
>>> Vale should know which BO & recid a cursor field belongs and allow for
>>> editing then all occurrences of that record field in the cursor would be
>> updated
>> 
>> No, Erne. Problem not that Valentina did not know.
>> Problem in ambiguity.
>> 
>> I not agree that this is good idea update all values in cursor.
>> 
> 
> why not? after all that's just what that change implies
> and the very meaning of relational model
> 
> if I must implement this on my own, that's what I'll need to do,
> right now Vale does this automatically since changes are dinamically seen in
> all living cursors - but this will no more be true if we going to have tmp
> for cursors
> 
> ok... this all falls back to old discussion
> if we can keep cursors as the cool data storage commodity also in a
> multi-user environment
> or we must change our way and use them solely as data retriaval tool, store
> their values somewhere and release them ASAP
> 
> this can be time & memory consuming and dangerous as well because
> how will my app know that data I'm viewing have been changed by other apps?

Erne, 

But this is the MAIN RULE OF TRANSACTION !!!

    * You MUST NOT see any action of other users.
    * You should work like any other user exists.

This is IDEAL.

If you do not want this, you need do DIRTY READ mode
    and periodically refresh cursor if it is TMP or snapshoot

Postgre, Access provide feature -- change on other user update.
I think we will consider this.
But developer must perfectly understand when he can use this, and when this
will be killer of system.

 
> will we have some DataChangedNotification event?

We will think about this.
But I think not in 2.0

> If yse where, in the VDatabase object, in the VBAseObject object, in the
> VCursor object? 
> 
> And what about implementations of Vale that don't handle events?
> 
> I think I'm making too much confusion here,
> what I really mean is that would be nice if most if not all these issues
> would be handled by Valentina at a kernel level with a standard transaction
> model

-- 
Best regards,
Ruslan Zasukhin      [ I feel the need...the need for speed ]
-------------------------------------------------------------
e-mail: ruslan at paradigmasoft.com
web: http://www.paradigmasoft.com

To subscribe to the Valentina mail list go to:
http://lists.macserve.net/mailman/listinfo/valentina
-------------------------------------------------------------