When I hear “Senior Developer” I think of someone who has mastered programming. I think of a person who can design, code and test a system. They can talk to system architecture or component design. They understand and use design patterns. This person can anticipate the performance bottlenecks, but knows not to pre-optimize. This person will leverage asynchronous programming, queuing, caching, logging, security and persistence when appropriate. When asked they can give a detail explanation of their choice and the pros and cons. In most cases they have mastered object-oriented programming and design, this not an absolute other languages such as javascript, F#, scheme are powerful and are not object-oriented at heart. They are adept in risk management and most important of all they can communicate the before mentioned to their peers.
What is mastery? There is a generally accepted idea, that to master ANY one skill it takes 10,000 hours of repetition for the human body and mind to grasp and internalize a skill. This is written to at length in Malcolm GladWell’s book Outliers.
Some examples of in Malcolm GladWell’s Outliers are:
Mozart his first concerto at the young age of 21. Which at first seems young, but he has been writing music since he was 11 years old.
The Beatles were initially shunned. They were told they did not have the mustard and should consider a different line of work. They spend 3 years in Germany playing about 1200 times at different venues, each time being 5 to 8 hours in length. They re-emerged as The Beatles we know and love today.
And lastly, Bill Gates at age 20 dropped out of Harvard to found Microsoft. To some this might seem foolish, but considered at 20 he had spent nearly half of his young life programming. In 1975, only maybe 50 people in the world had the experience he did. His experience gave him the foresight to see the future in Microsoft.
Peter Norvig also discusses the 10,000 hours rule in his essay “Teach Yourself Programming in Ten Years”.
In the book Mastery by George Leonard, great detail is given on how to master a skill. One must practice the skill over and over and over again. The more the repetition, the more you become aware of the differences in each repetition. Only with this insight can you become better.
The software industry’s titles (Junior, Mid-Level and Seniors) are misleading and inconsistent from organization to organization. I’ve worked with companies, who defined a Senior Developer as someone who had 5 years or more of experience. There is no mention to the quality of the experience, just that they have sat in from of a computer for 5 years. In working with these folks many of them had not yet grasp object-oriented programming, yet they were considered Senior Developers.
There must be a better more objective way to measure the skill set of a software engineer. John Haugeland has come up with a computer programmer’s skills matrix. It’s gives a common, objective way to measure a programmer’s skill level, which otherwise is left mostly to gut feeling.
When looking at software engineers I see 4 tiers of skills: Luminary, Senior, Mid-Level and Junior.
Luminary (10+ years) is one who has mastered a skill and has set about improving their respective discipline. Some examples include: Ted Neward, Uncle Bob Martin, Donald Knuth, Oren Eini, Peter Norvig, Linus Torvalds. This is change depending on your skill-set.
Senior (7 to 10+ years, Level 3) is one who has spent the last 10,000 hours programing in a specific genre. There is a strong understanding of design patterns, They leverage asynchronous programming, queuing, caching, logging, security and persistence when appropriate.
It’s very possible that a Senior will never reach Luminary. Luminary’s are often found speaking and writing. They are actively trying to impact their discipline.
Mid-Level (4 to 6 years, Level 2) is one who understands day to day programming. They work independently and create robust solutions. However, they have yet to experience creating or maintaining large or complex systems. In general Mid-Level developers are great with component level development.
Junior (1 to 3 years, Level 1) is one who understands the basics of programming. They either have a degree in software engineering or computer science or they are self taught. Their code is continually reviewed. Guidance is given in regard to algorithms, maintainability and structure.