One would wonder, what are the responsibilities of these Architects?
- They are supposed to see the bigger picture
- They are supposed to guide the project in the right direction
- They are supposed to give good solutions to technical problems
A lot of architects that I have met, have stopped doing active programming. Most of their time is spend in attending meetings, calls and drawing some pretty diagrams.
You might think whats wrong with that?
Well, I think (and these are strictly my views), If someone is not in touch with programming then how can he/she suggest a sane solution to any technical problem?
If someone has left programming years ago then
- How can he/she solve any technical problem?
- How can he/she visualize that there is going to be a technical problem implementing a solution that he/she is suggesting?
And I am making these claims out of personal experience.
First example:
I have worked in a project where some architect decided that, there is no MVC framework in the Java world that suites our requirements!
And the project ended up have its very own brand new MVC framework. This was not some project in 1980 (BTW I was not even born by then!). This was a project executed in 2009.
I don't understand that in 2009 how can one decide to write a new MVC framework?
One might thing whats wrong with writing a new MVC framework? Most important problem is, re-invent the wheel!
Lots of open-source projects have invested a hell lot of time and money in making sure that, a MVC framework does what its supposed to do.
If someone write an MVC framework, may be he/she is not aware of all the possible cases an MVC framework should support.
End result, a half baked solution that solves the "todays problem" but its no good for the future!
Another thing to consider is the cost: Who is paying for this framework anyways? Its the client of-course.
End result, software becomes expensive!
I think such wrong decisions make software expensive then it should have been. Client gets a piece of software that he never needed in the first place.
I think there must be at least 25 MVC frameworks in the Java world in 2009. I am sure at least a few would have fit whatever requirements we had. Its just short slightness on the part of the Architect, to decide to write a new MVC Framework rather than finding one that fits the requirements.
Another example:
I had worked in a project where Architect's had decided that, we should not be using any types other than String's.
We had a few web-services to do various things in the project. One of them was to do payment. According to his suggestion, all arguments to the methods in payment web-service should be of String type!
Man! the first time I head of this decision, I was thinking, what the @#$%^&!
Imagine passing amounts as String and then in the web-service implementation converting them into proper types.
Imagine because of the fact that amounts were of type string, client's of web-service could pass "abcd" as the amount.
Hence, web-service implementation had to validate these values and throw an exception or return an error message stating that "abcd" is not a valid amount!
All that validation code and the mapping code is all waste! It's just a big big big waste! Something that was not required in the first place. Client paid for that code! (Its almost like distributing free money.)
Why did this happen?
Because the Architect was not in touch with Programming, he could not visualize the real problems that would appear because of this decision.
End result, maintenance nightmare, ugly code, overly complicated system.
I think in the end all I want to say is, Stay in touch with Coding. Coding is not a mundane task that anyone can do.
A person can become a good Architect iff (if an only if) he/she is a good Programmer!