Thank you.

That's interesting that collect: returns different classes of collections based
on the class of the source collection.  When would you want that behavior?

I see in this particular method, the next line allocates and Array explicitly. 
Is there a reason we want to let collect: decide the class of collection in the
first line but use an Array in particular in the second line?  

	allText _ pages collect: [:pg | OrderedCollection new].
	allTextUrls _ Array new: pages size.

Would you say we should change the second line to:

        allTextUrls _ pages collect: [:pg | nil].

As an old C programmer I'd be tempted to go the other way and force them both
to be Arrays for compactness since I know they're not going to grow.

-- Dan

--- Bijan Parsia <bparsia at email.unc.edu> wrote:
> 
> On Fri, 20 Oct 2000, Dan Winkler wrote:
> 
> > A Smalltalk beginner's question...
> > 
> > In BookMorph>>getAllText I found this curious line:
> > 
> >         allText _ pages collect: [:pg | OrderedCollection new].
> > 
> > That seems to iterate over all the pages in the book, but not actually look
> at
> > what it's iterating over.
> 
> Yep.
> 
> >  It just wants a collection of empty collections, one
> > for each page.
> 
> Indeed.
> 
> >  Wouldn't it be cleaner to use a loop based on "1 to: pages size
> > do:"
> 
> Certainly not :) This is a degenerate use of #collect: but perfectly
> reasonable. Think about the extra cruft you'd have to add. First, you'd
> need to determine *what kind* of collection pages is (right now it could
> be anything!), then make a new one, save it to allText and then,
>  in your block, add: a new OrderedCollection each iteration, well, if it
> *supports* #add, which it may not. It could be an Array, in which case you
> might be better off setting up a stream...
> 
> Bleah. Fragile and yucky and unclear!
> 
> Now say, in some future version, you want to do *more*, e.g., you want to
> add the title of the page to each fresh OrderedCollection. Now you have a
> *mess* of rewriting to do, since you have to figure out how to get ahold
> of the page. As opposed to:
> 	pages collect: [:pg | OrderedCollection new add: pg title]
> 
> 
> 
> > instead of locating and then passing around each page without actually
> > using it?
> 
> Er...This is very very very very very very cheap. They're just
> references. Plus *you* aren't doing any locating or passing, so what do
> you care? ;) A future compiler might even optimize this.
> 
> >  Or is this a Smalltalk idiom?
> 
> Remember: the enumeration protocal is my friend, the enurmeration protocol
> is my friend. #do:, #collect:, #select: etc. are almost always prefereable
> to some #to:do: loop and they're *always* cleaner. (There *may* be some
> cases where inner loop logic takes over, but they're super rare.)
> 
> Cheers,
> Bijan Parsia.
> 
> 


=====
+
+
+-+-+-+ +-+-+-+-+-+-+-+
|D|A|N| |W|I|N|K|L|E|R|
+-+-+-+ +-+-+-+-+-+-+-+
+
+

__________________________________________________
Do You Yahoo!?
Yahoo! Messenger - Talk while you surf!  It's FREE.
http://im.yahoo.com/