Stefan Tilkov talk on "Why software architects fail - and what to do about it" came to my attention right after I finished the first chapter on Practical Object-Oriented Design in Ruby: An Agile Primer by Sandi Metz. Quite a coincidence, but not really. I've had Stefan talk in my Watch Later list on YouTube for quite a while, but never got to decide it was worth my time.
But it is. Oh boy... So much. Stefan has really eye opening insights, specially for the young and inexperienced software developer like myself. Software design is not a well taught subject at University (problem decomposition is never taught as central topic in Computer Science and related fields - at least in my Uni).
Lessons learned from his talk
Generic solutions can be a pain to work with;
Understanding the domain does not cause allergies;
Specialized solutions might be the answer, but it can be just as bad as a generic one;
Do not add to complexity, keep it simple;
Analysis paralysis is bad, really bad (as an anxious person, I can confirm that);
Rehab yourself from innovation addiction (you are not FAANG);
Enforcing an architectural approach that clashes with frameworks, libraries or tools you use;
Don't use tech because it is popular;
The hammer and nail thing: avoid it;
Sticking to awful things just because they are there (hello there, JavaScript);
Being detached from the reality of the system is a no-no.
How to be a better architect?
Dogma and rules: 10%
Experience: 20%
Pragmatism: 20%
Flexibility: 10%
Minimalism: 10%
Trends and future needs: 10%
Experiments and PoCs: 10%
Hands-on participation: 10%
Vendor advice: 0%