I just got done taking the MIT Intro to CS class on edX. I did this to get a sense of where I stand.

I could teach it. Hands, down.

I am going to take the second class in the sequence for the same reason. My guess is I know most of it.

While in school I got as far as Linear Algebra. So, I'm OK with most math. Although, frankly, I haven't really had to use Calculus very much over the years.

My approach has always been to take deep dives into whatever I've had to tackle. When faced with having to do image processing work on FPGA's and write convolution filters I simply bought every relevant book I could find and jumped right in.

When I came across a problem that could benefit from applying a Genetic Algorithm training a Neural Network, same thing. I have yet to find something I cannot tackle.

You have to keep in mind that what was taught in colleges and universities back in 1984 or thereabouts is stone age when compared to today. I took a C class one semester. I nearly fell asleep. I had already done more on my own than the professor was covering in this class. I was already working with multidimensional arrays when he was covering how to print to the console.

The other reality is that nearly nothing I've done over the last fifteen years even existed back when I was in school. The only way I could have done multi-gigahertz FPGA work was to learn it all on my own. When I was in school I don't even think PAL's existed. I designed and built my first computer. That means, design, buy the chips, wire-wrap the boards and use a dumb terminal and a hex keypad to bootstrap the thing. Different times.

With regards to your question about my grasp of software engineering. Well, you know that movie where the kid goes "I see dead people". I see code. After having done so much work across a number of domains most solutions develop in my head intuitively without thinking. Again, the MIT course I just finished being an example. I really didn't have to think much about any of it. I just sit down and write code. Good, clean and efficient code. Comments and all.

So it sounds like you have as much technical talent as anyone who's launched a successful tech venture...so will a com sci degree help? No, you're obviously hitting intractable institutional biases here...looking back at college, I see a lot of its value in networking and exposure to new ideas, but I don't see non-traditional students benefitting to the same degree...so why waste time and money.

Regardless of your technical aptitude, the fact is is that, particularly in web dev, companies are all too happy to hire novice bootcamp coders for near 6-figure. Maybe there really is a huge scarcity of Rails developers, and perhaps it's worth focusing on that to at least get your foot into the modern development industry, but again, it's hard to anticipate how many walls will be unfairly in your way because of the age discrimination you've described.

So I guess the short way of saying it is that pure coding skills alone won't be your edge, and a degree will not change that. At least when it comes to getting recognition from recruiters and employers. However, code can be your edge in helping you pursue, produce, and promote entrepreneurial ideas, but "Just have and execute a really good idea!" is not terribly helpful advice.

But I wouldn't feel insecure at all, if I were you, about whether your coding experience is relevant today. I was in college less than a decade ago and we started with C and most of my engineering labs were doing things like building and designing the kind of adder circuits built decades ago...That kind of thought process and desire to tinker with things doesn't go obsolete with new programming languages.

> So it sounds like you have as much technical talent as anyone who's launched a successful tech venture

Without any ego at all, I'd say, yes. I may not be absolutely up to speed on every aspect of the cadre of web technologies out there but that's not an issue. I'm comfortable with HTML5, CSS3, Javascript/Jquery and PHP. I am now switching to Python and Django due to a number of reasons not relevant to this discussion.

Can I code all of these technologies like a possessed monkey? No. My style is different. I rely on reference material to support what I need to do. I know what needs to be done and I get it done. In other words, to throw out an example, I know when a state machine is the best way to solve a problem, I know how to design it and code it right the first time. During that process I might need to look up language details here and there because in working with so many languages I decided a long time ago it was pointless to try and remember everythign about every single one of them. I get to job done. Now, if I am working on, say, Objective-C daily for weeks the reference material moves to the background pretty quickly as I don't have to "task switch" between five different technologies in the course of a week. Not sure if this is how young coders work at placed like Google. Perhaps they do rock at the tools required for the entire stack, from server to client-side. Or maybe they specialize and only do JS. Don't know.

> so will a com sci degree help?

I can't see how other than, perhaps, if employers or investors place value in the credentials. I understand that a degree is useful to establish a baseline. At the same time, I can code FizzBuzz just fine and, apparently most CS graduates can't code their way out of a paper bag. Currently crunching through ProjectEuler.net as a neat way to get better at Python.

The one area that's of great interest to me is machine learning and AI. I've dabbled in it here and there but my work never landed me solidly in projects where I had to learn the material. With all the new-found interest in robotics I find myself thinking that this is an area where I could do very well. I mean, I am expert level at Solidworks and fairly decent at mechanical design and can also do all of the electronics and software. I could see formal learning in ML being a good approach now that I am not busy trying to run a business from day to day. Yet, I find myself asking if that's reason enough to enter into a four+ year program when I could acquire the knowledge from one (or all!) of the excellent ML courses available online today. I'd have to bone-up on statistics, calculus and linear algebra before I could jump in head first. The difference is that I could do this on my own and only on the stuff I don't remember rather than having to endure a year and a half of math towards a degree.

As you can see, I am fairly biased against the whole idea of going for a degree at this juncture. Still, I want to keep an open mind and listen to reason before making a decision. Because my wife is not an engineer I don't think she understands what the landscape looks like for engineers my age, degree or not, any more than I might understand the nuances of her medical practice.

> I was in college less than a decade ago and we started with C

Right. Most anything an engineer (software or hardware) will do today probably wasn't taught in school ten years ago. If you are not learning new things on a constant basis you'll quickly become irrelevant. As an interesting point, I do have friends around my age who have not kept up as I have and are pretty much done with engineering as it's known today. One of them actually stopped doing hardware and software about fifteen years ago and shifted into management. He got laid off. He is unemployed. Nowhere to go.