I can remember the first time that I read TOGAF. I was really amazed from what I read. After spending time to understand Zachman framework, TOGAF looks like a mature and practical EA framework. Each paragraph that I read had a lot of sense and it looks like I just need this framework book on my desk to start an Enterprise Architecture journey. This was almost 10 years ago and before I had any enterprise architecture experience.

I  will never forget the experience that I had with implementation of TOGAF as well. My admiration turn into disappointment very fast, what seems to be great ideas turn into non-practice advices. Actually I found out later that TOGAF, as any other EA framework, is actually a collection of common sense advices collated by someone into a framework. Yes, I know that EA framework are reference models and they should be adjusted to each enterprise environment. but if my adjustment are more then 50 % of the basic Framework, can one still claim that a framework is a reference model?

Regretfully I found out that as I'm moving forward with enterprise architecture implementation, my adjusted framework is becoming more and more different from the vanilla framework. Actually after some time I always found myself with a framework that significantly different from the reference framework that I started from.

People might say, that's OK. The EA framework helped you to start and then you adjust it to your own needs. Regretfully, I found that this claim doesn't have any grasp in reality as well. Actually what I found out is that starting an EA journey without a reference EA framework will bring you to results faster then following EA framework. finding Incompatibility between my organization needs and the framework as well as the time needed to fix those Incompatibilities takes more time then to strart without a framework.

In my own opinion enterprise architecture can be learned just from experience. You can read as much as you want about it, but just real life experience will teach you how to do enterprise architecture. Learning EA is a long journey full of mistakes that you'll discover by your own, fix them as you moving forward and will avoid them next time.

The best allegory  to what I'm trying to describe is raising children. There are many books and methods about how to raise children. Even if you will read all of those books and you'll try to follow all of advices, just the journey of raising your own children will teach you how to raise your own children. You'll do a lot of mistakes the first time, learn from them but still you'll have new experience with any child that you'll raise. The same is enterprise architecture. Following a framework won't help and in the worst case it just will create more problems. You need to roll your sleeves and start to feel real life practice. That the only and the best approach that I can recommend.

You defenetly don't need EA framework to start your work ( and personally I'm responsible for one framework), but there other information that you should prepare in advance:

1) What your EA initiative is going to solve

2) which data do you need inorder to make your EA work (Meta model)

3) What are the deliverable that you're going to provide (approved by stakeholders)

4) CEO or at least CIO support

5) Team member from business and IT departments