Ask HN: Can you switch to a programming career despite nearing 40?

Fast forward 10, 20 years, some children, maybe a few wives/ex-wives, aging parents, you just don't have the time without some sort of moonshot mentality.

So I think a lot of it is circumstantial and not actually age directly, I am over 50, and design/write highly specialized software, the sort that ends up averaging a line a day across the life of the project maybe, or definitely that kind of metric for unique code.

I pick up new stuff all the time, I look for better ways and tools in general browsing etc almost every day, experiment with whatever I can get my hands on and find that any slow down in cognition is more than made with by experience and the ability to just "see" things at a glance other people cant.

Don't let age defeat you for no reason, believe in yourself and not what other people say, you will know what you can do fairly quickly, just be honest with yourself and target your strengths and work on your weaknesses.

Of course that's also coupled with the family constraints you mention as well.

- Requirements taking from business owners

- Data modelling for data warehouses

- Writing Python code for transaformation (from source table to data warehouse tables) and their scheduling as we are following ELT but EL parts are taken care by HQ

- Writing Python code for alerts, monitoring and dashboarding (using the excellent Dash library)

- Huge amount of SQL

If the answer is yes then I did switch to a programming career when I'm two years from 40. I'd really love to do more ETL but sadly HQ took away most of the job.

Avoid competing with 20-somethings coming out of local coding bootcamps. Here there are multiple groups churning out React + (node or ruby on rails) web devs. Just learning Angular instead would set you apart.

Pick one thing and focus on it. It can be so exciting to learn a bunch of new things but you will end up spread too thin to be of much value. Now that "one thing" can be a full stack of a couple technologies, but pick one of each and specialize in it.

Don't be too niche for your market. If you're in a small market, there may not be a lot of demand for machine learning or rust. A solid .NET dev is likely more employable. Look at open job postings and use them to build up a list of common needs there.

There's a pattern I've seen a number of people successfully follow. Find a somewhat newer technology with good potential value in your market. Learn it. Blog about what you learn. Develop a talk about it and give it a couple of times (Meetup groups are usually desperate for speakers). Produce a video of your talk, either from an in-person event when we can again or one from your desk environment. Start pinging publishers about writing a chapter or reviewing a book on your chosen tech. If possible, get your name on a book cover, even if it's with other authors. From then on, you are "Zero Balance, author of 'Getting Started With Golang Embeds'". Even if you're applying for jobs that aren't closely related, you project a lot more authority than the average coder.

So the question isn't really if someone at 40 can learn to code, it is more if someone, at any age, who hasn't thought in terms of systems can grasp it.

I have met some people who can't code who understand systems in a way that would totally blow away most people, SV, London, Peru, wouldn't matter where you were.

I have also met people that can code endlessly, but couldn't architect a novel system arrangement if their family was being held to ransom, it's a way of thinking.

It used to be that a systems analyst was a profession all of it's own, not really at all any more, and it is my observation this is the really root cause of so many software projects failing.

The first step in any formal analysis is Problem Identification, but it is surprising how often people will spend tens, hundreds of millions of dollars and never formally go thru this essential first step, and are then basically better off buying lottery tickets than hoping for a successful project. (and then go on to blame the consultant, or the market, or the fickle users, when really it was them all the time but they just never realised it).

So code, phht, that is not even the entry ticket, more the uniform.