Punko,
> I'm curious - what are the obstacles you and others are seeing
> that prevent OOT from coming to the fore?
O-O has come from being a programming paradigm and as such has had big success (C++, Java etc. and many other tools have adopted o-o ideas). There has also been backlashes: not living up to the promise of reuse, lack of standards (many notations, many API:s, storing objects in databases etc), and as you say a scarcity of talent.
What I think and hope is happening is the acceptance of OO as more of a modeling paradigm, where businesses and systems are modeled with objectoriented principles and then deployed using objectoriented technology (being architecture-driven, components etc). This needs the acceptance of a standard notation (UML) and also a set of different development processes (you can't standardize on process, they need to be different for different companies, domains and situations).
> what events must take place
1. Management must realize that OO is really a way to get control of software development on a higher level (industrialize, using "blueprints for software"), not a "programming tool".
It's not as much their fault as it is those who advocate OO. I've seen presentations for management where they talk about virtual inheritance, polymorphism etc. Naturally, management just shakes their heads. Also, management want numbers: how much faster ? how much better ? how much cheaper?
2. OO must be teached at all universities to create a new generation of system developers (well, you can train the current generation also :-) ).
3. Good, supportive, integrated tools must be available (hopefully we'll get there but we're not quite there yet).
4. Sometimes I feel that this OO world tends to be made overly complex, with a lot of OO "experts" trying to complicate things rather than simplifying them. All the basic ideas behind OO can be explained to and understood by a non-technician in 2 hours.
> what do you think the timeframe is?
I feel very sure that OOT (in some form) will be the dominant development paradigm in X years, and virtually undisputed. However, I have no idea if X is 5,10 or 15. Not before year 2000.
/Hans-Erik. |
