Last month, JS Monthly London‘s host Guy Nesher gave a talk titled “JS interviews”.
Guy stated that as “normal” developers, working in angularjs / react / backbone, we never really need to use things like closures.
If we use a linter / strict mode (or, honestly, just some common sense), hoisting is not something we’re going to encounter, and in general – all of that ‘advanced’ stuff – prototypes,
bind.. – that’s just stuff you need to know for your interview, and then you can forget about it and go actually do your job.
Guy likened this to a carpenter being asked at a job interview whether he can change a lock (no, I’m not a locksmith), or his opinions regarding some abstract wood-manufacturing techniques (I don’t care, I just cut the wood).
The underlying assumption was that
(complete slides from the talk can be found here, courtesy of Guy)
Now, Guy is not your average developer, having spent years in law before making the switch to software, so it’s understandable that he has some different views.
And when I say ‘different’, I mean ‘bloody infuriating’.
You can’t be a driver without being able to change tyres
If you’re anything like me, you probably have steam coming out of your ears by now.
How dare this guy claim that you can be a competent developer without a solid understanding of the technology you’re using, of general software engineering and CS concepts such as SOLID, data structures, performance..?
What the hell does he take us for, some mindless code monkeys?
Of course you need a deep understanding of the technology and concepts behind the code you’re writing, otherwise you wouldn’t be able to understand why your code behaves or performs in a certain way, and you wouldn’t be able to debug it in certain situations!
You also wouldn’t be able to apply well-known solutions / patterns where appropriate, or understand the cost / benefit of using one technology over another.
So at the Q&A portion of the talk I posed Guy with the following question-
“Suppose you were interviewing me for a position, not as a carpenter, but as a driver.
You would ask me ‘do you know how to drive a car?’, and I would answer ‘Of course! Right pedal is to go, left pedal is to stop, and you control the direction with the wheel.’
‘Great’, you would say, ‘And suppose that, while you’re driving along, you get a flat tyre. Do you know how to handle that?’
‘I don’t really know car engineering in depth.. all I know is Right pedal is to go, left pedal is to stop…’.
Would you hire me? I can still get your car from point A to point B.
However, if anything goes wrong, I’ll be stuck.”
“But software development isn’t something you do alone, like driving.” Was Guy’s answer.
“In software development, you’ll have a senior developer on the team who’d know how to ‘change tyres’. A kind of a ‘pit crew mechanic’.
But the rest of the time, the ‘junior driver’ is going to be cruising along just fine.”
Are we all such special snowflakes like we’d like to think?
Guy’s answer lead me to do some thinking.
Doesn’t a lot of what I do, day-to-day, consists of ‘more of the same’?
CRUD over a database, some validations, showing some aggregation to the user…
In front-end development it’s even easier to spot:
Some form, an AJAX call to a server API, displaying data.. there isn’t even any business logic involved (hopefully).
A lot of the routine is… well, routine.
I don’t use any ‘special’ or ‘deep’ knowledge when I do the above things.
Not all of my time is spent innovating or ‘Engineering’.
A lot of what I do actually isn’t that special.
Especially if I’m using very high-level languages (ruby, JS), especially when using frameworks on top of those languages (rails / angularJS) to abstract away the ‘scary’ SQL or network operations.
So maybe Guy is right? for the 80% of routine work, you don’t need to hire a ninja rockstar hacker (or whatever the stupid buzzword du jour is);
Just have one senior guy within a team that can handle stuff like architecture, coding standards, tech evaluations, helping the more junior members when they’re stuck, and let the others get on with the day-to-day.
Bootcamps and the rise of the junior developer
The notion presented by Guy ties in very nicely with recent trends in the professional software world.
As demand for software developers is expected to rise at a “Much faster than average” rate, universities don’t produce new graduate at an increasing rate.
This gives rise to “code bootcamps“: intensive, 6-24 week programs, designed to bring you from zero to web-developer hero (or, more precisely, to junior web-developer).
The number of bootcamp graduates has been growing significantly in recent years.
This trend, if continues, means that we’ll be seeing more and more developers who, like Guy, have never studied algorithms, data structures, or anything else that might be ‘under the hood’ in day-to-day work.
It will be extremely interesting to follow these developers over the next few years, and see how well they’re able to make the transition from junior to mid to senior, and what are the differences in their performance compared to the other two large groups of developers – the formally educated and the self-taught.
What do I do with all these juniors?
Whether we like it or not, it seems that Guy’s (and the bootcamps’) vision is here to stay – more and more developers with little or no ‘deep’ knowledge in programming will be joining the workforce in the coming years.
Numbers alone dictate that – there is, and will be, a huge demand for developers, inevitably leading to lowering the entry barriers for newcomers, and making experienced developers that much more expensive.
That means that it’s very possible that you will end up on a team that is some sort of variation of what Guy has described – a few ‘drivers’, with one or two ‘pit stop mechanics’ to help them along.
It seems that the thing to do right now, instead of looking at these people down our collective noses, is to come up with an effective method for integrating and mentoring these newcomers.
Whether they’re interested in eventually becoming ‘mechanics’, or are content to just stay ‘drivers’, they’ll need our help.
I’ve personally been part of a couple of teams which included bootcamp graduates.
And, of course, I was once a junior myself.
In all these situations, I never thought the team had enough awareness for the need of junior developers to be mentored:
The assumption of the team was that after a suitable period of training, these developers are ‘ready’, and were therefore thrown into the deep end and were treated as any other team member.
In my opinion, mentoring / training of junior developers is better as a sustained, consistent process, as opposed to a ‘one and done’ job.
Here’s my $0.02:
- Juniors need to receive feedback and advice on their work often, and over a long period of time.
Instilling ideas and ways of thinking is a lengthy process.
- Teams need to understand that having junior developers on the team doesn’t only mean that they (juniors) will be performing slower, due to their inexperience.
It also means that more senior team members will be performing slower, due to the fact that they also need to be assisting their team members.
- Having different skill levels should be reflected in the work being done by team members – some tasks are more complex, or require greater knowledge and experience, so they shouldn’t be done by a junior.
- It needs to be official: Managers need to let team members know that guiding / being guided is part of their jobs descriptions.
This will help to avoid friction from juniors who are perhaps too ‘proud’ to accept guidance, and from seniors who can’t be bothered to guide.
- Making it official will also guarantee that the subject of training new team members is not forgotten or abandoned as projects and deadlines get more hectic.
This post began with a question – do you need to be intimately and deeply familiar with the tools that you’re using in order to be an effective developer?
The answer, in my opinion, is “No, but up to a point”.
You can be extremely productive in a lot of scenarios, not having a broad knowledge base.
If you want to progress beyond the ‘junior’ label, however, I think you need to expand your knowledge.
However, regardless of my, or anyone else’s, opinion on such developers, the reality is that we’re going to be seeing a lot more of that type of devs in the coming years;
Rising demand for developers, coupled with the rise of the “bootcamp” concept, mean that these junior developers are going to be coming into your team.
The question then becomes how to utilise these developers in order to produce the best quality (and quantity) of work?
People have been trying to answer this question for a while now.
However, it seems that the assumption is always that the individual is responsible for her own training, or, at most – that training is something that’s internal to the team.
I don’t think this is enough; companies need to understand that the success of their projects and their organization is dependent on the success of the juniors.
Therefore, there has to be a management commitment to making these people successful.
Training and supervising these guys takes time and resources, from both the junior and senior members of the team.
It also requires a slightly different work process where junior members are involved.
Also, how will this play out with other factors like the high pace of the industry, and the relatively high turnover rate in our field?
Will a company that needs a project done today be willing to invest in training an employee who might not be there tomorrow?
I guess we’ll see that soon enough.