Should there be a Notify in the AttachObserver method of an observer?

Hello everybody,

a few days ago I read an article which covered an example
implementation of the observer pattern. It alluded to the importance
of calling the Notify() method from within any AttachObserver()
method.
In C++ this would result in something like the following:

void Subject::AttachObserver(Observer* AObs)
{
if (ObserverList.Contains(AObs))
return;

ObserverList.Add(AObs);
Notify();
}


This invocation of Notify() guarantees that a newly attached observer
can immediately adjust to the current state of the observed subject.
Now that seems pretty reasonable to me, which led me to my first of
two questions:

Question 1: Does a reason exist why this is not mentioned in the GoF
book?


The problem I see with notifying all subscribed observers every time a
new observer gets attached is that this could well add up to a
veritable notification surge when a subject already has multiple
observers. If the attachment of several observers to one subject
occured at the same time, a possible solution for avoiding the
reduntant notifications could be a BeginUpdate/EndUpdate pair
enclosing the observer creation.
But I do not really fancy the BeginUpdate/EndUpdate mechanism since
faulty implementations of subject classes as well as imponderabilities
in multithreaded environments might lock the whole updating process.

Question 2: Can the BeginUpdate/EndUpdate mechanism be considered a
good solution to prevent premature notification of observers


Thanks very much for reading and kind regards,
Malte

---

The above e-mail address is not valid. To
contact me, please use my real e-mail address:

malte AT t DASH online DOT de

Just replace the capitalized words with the
corresponding punctuation marks.