Hi,
You are probably going to hear this a lot: it depends.
You'll find a lot of people doing 1, you'll find a lot doing 2... 3 4 5
6 etc.
For example, (just a few examples, books could probably be written on this
subject), an animation of a car driving down a road, sliding, going through
corners etc:
- could be an animation on a timeline
- could be coded using positions, vectors etc
- could be coded using displacements maps
- could be coded by recording all the positions of an object during a tween
and playing it back (altered or not)
- ..?
Which is the most appropriate to your project? Do the animations have to be
very specific, are they updated a lot? When they do have to be updated, do
you have to go to heaps of code, just becoz u have a timelinephobia? Etc etc
These are all questions u could ask yourself. I took the example of an
animation, but you could ask these kinds of things for databases as well
(which database? One or more? Do we have to switch between dbs? Etc).
No matter what I think u should be able to answer the 'why?' questions for
yourself. Why do I do it like this?
For example, why do I use a timeline animation in case A, and a coded
animation in case B?
Well because in case A the animation is well defined and very simple, and I
want to be able to update it quickly, in case B the animation is very
complex, it depends on factors like gravity and the like, and I don't want
to go there timeline wise.
What? So you use different options at the same time! Sometimes yes ;). The
beauty of grey ;). Note that I don't mean to sound like a know-it-all,
because the issues you raised are issues I think a lot of us struggle with
everyday (I know I do!), so I don't claim to have all the wisdom, I'll just
try to give you one view (my view) of the story.
Anywayz, back to the example...
So different options at the same time... The animation example is taken from
real life, from a project I'm working on as we speak. I always try to think
of it as, if I don't have a clue about HOW I'm going to do it (99%) I stop
and remind myself of the WHAT (as I should do all the time :)). So.. We have
an abstraction: an Animation class.
What does it do? It plays, it stops, etc. How? No clue, an Animation
subclass could be an animation built on top of Fuse, an movieclip with a
timeline, a handwritten piece of code, I don't care, as long as it plays and
stops and....
So I don't think this falls within either option 1 or 2, but within option
x... And I wish I knew how to describe option X :).
With regard to design patterns and MVC... First of all, I think design
patterns are a means to an end. Design patterns have multiple
goals/attributes:
- they solve (part of) a difficult architectural problem for you
- they increase reuse, robustness, maintainability
- they help documentation your code, instead of describing class a,b,c,d you
say, I used this pattern, and these classes play these roles
- they might increase the complexity of your application
Looking at the second point, reuse, robustness etc, the first patterns that
come to mind, might not even be considered as patterns.. Low coupling, high
cohesion, that kind of thing. Even without learning dozens of patterns, a
lot of times the questions are as simple as:
- if at one point I want to incorporate this into my programm, what happens?
- if I pull this string, what happens?
- if ...
- if ..
- if .
If the answers to these kind of questions are: well, then I'd have to throw
it all away and start over, some of the basic patterns may have been
overlooked. On the other hand the answer might as well be: listen, you are
trying to fit a square peg into a round hole, you are treading outside of
the scope this code was meant to handle. One more note on design patterns,
one of the most difficult things I encountered is that sometimes these
patterns have not different goals but OPPOSING goals, one might increase
reuse but break encapsulation, one might improve encapsulation but break
reuse etc.
With regard to the MVC pattern, I often collapse the controller and view.
Most important for me are clear cut responsibilities, a model with no
dependencies on the view. If another view has to be created, I might
refactor my collapsed view/controller. Lot of times there is even discussion
on whether it IS collapsed or not. For example 2 options:
- I write a movieclip, put a few components in it(view) , and write a class
(controller) and connect it to the movieclip in the property dialog. Is it
collapsed?
- I write a class which creates all the components by code. (this might be
clearer)
If I am certain up front there will be multiple views, I might put more time
into the view/controller issue. In any case its just another black box to
the rest of the program (let's assume a perfect world here for a moment).
Lot of the different approach to MVC even stem from names issues, for
example all the code in two examples might be the same, but in example A the
view is called MyView and in the example B the same thing is called
MyController. Go figure ;)...
Damn... Can't quit.... Can't quit...
Sorry for the longwinded answer, maybe someone else could provide you with a
simpler/shorter response ;)
It's not really the answer you were looking for maybe, but I hope it
contributes to your thought building process about these kinds of issues...
I'm looking forward to the other responses, as I'm always wondering about
these kinds of issues as well...
Greetz
Hans
________________________________
J.C. Wichman . Software Engineer
www.TriMM.nl
TriMM interactieve media . PO Box 1208 . 7500 BE Enschede . The Netherlands
Oldenzaalsestraat 141-143 . tel +31 (0)53 - 48 00 480 . fax +31 (0)53 - 43
53 027
-----Original Message-----
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of
[EMAIL PROTECTED]
Sent: Saturday, January 28, 2006 12:58 PM
To: flashcoders@chattyfig.figleaf.com
Subject: [Flashcoders] Q: All OOP or ??????
Hi
I was curious what approach most developers are using for their projects:
1 doing ALL of there development for every project within an OOP based
framewrok, ie everything, from drawing and/or placing of objects,
transitioning, interacting with server special effects, is done with a
custom Class based framework and is entirely OOP based.
2 Or do most developers currently use a more 'hybrid' approach, with a lot
of assets placed on stage at authortime and referenced via an '#include' if
code is needed to be outside the FLA file. If third party Class code
already exists (ie tween engines) then that is used rather than re-creating
it from scratch. Custom classes are developed ONLY if there is some need for
code re-use.
Related to the above discussion there has been a lot of talk about use of
design patterns, and , specifically the Model View Controller 'meta'
pattern. Having just finished 'Head First Design Patterns' it all makes
sense now. But there seems to be more than one way to interpret MVC,
especially when it comes to whether the controller should be separated from
the view
3 Which ways do developers implement the MVC pattern?
All the best
Jim Bachalo
[e] jbach at bitstream.ca
[c] 416.668.0034
[w] www.bitstream.ca
--------------------------------------------
"...all improvisation is life in search of a style."
- Bruce Mau,'LifeStyle'
_______________________________________________
Flashcoders mailing list
Flashcoders@chattyfig.figleaf.com
http://chattyfig.figleaf.com/mailman/listinfo/flashcoders
_______________________________________________
Flashcoders mailing list
Flashcoders@chattyfig.figleaf.com
http://chattyfig.figleaf.com/mailman/listinfo/flashcoders
