Computer Science Degrees - What's Lacking

I've been continuing to take classes despite working as a full-time developer. I work in the .Net world, but my heart is more in Python. Still, I wanted exposure to plenty of others and I'm currently in my second class on C++.

What I have noticed in the various courses I have taken at institutions of higher learning is that they don't teach anything remotely close to the real world. I know this is a common claim and people cry out against the undergraduate program not preparing people for the real world. I get that. But I have always dismissed them.

Being a better, more rounded thinker is more important than job training, is it not? Surely it is.
I'm not sure why no one slapped me. I've had that discussion a number of times with people. I should have been slapped. Now don't get me wrong. I value and love my education. I have a degree in Politics from the University of New Hampshire. I am versed in various literature and can hold discussions on a number of topics (at one point I could almost do so in French).

But this is a false dichotomy. It's as if the premise only allows well-rounded education or job training. That's ridiculous.

I believe what Computer Science majors need is less Assignment-based instruction and more Project-based instruction. Random assignments that are isolated from each other, meaning accomplishing one has no effect on accomplishing another, don't provide the type of reinforcement that maximizes the mind we need molded in academia. Completing a larger project with multiple parts is far more useful to the student. These parts interact with one another. One of the best lessons to learn early on is that changing code in one place can have small or large effects on other parts of code.

It's difficult to learn that in assignments that are 50-150 lines. But if you're building on top of the product/project you're working on all semester, you will naturally run in to those types of consequences and suddenly your student will begin to think very differently. I value those that can think holistically about a product rather than the frustratingly common linear thinker that says to the team "Well, I'll just change this and then it'll be fixed." I love it when I hear someone say "Well, if we change that there could be implications to these other parts."

Admittedly, I think there's a lot of cruft in undergraduate coursework these days. Taking art history doesn't make you a better thinker per se. You don't experience art in an art history class. You learn what you need to say about a particular painting/sculpture/etc. in order to demonstrate knowledge about the piece. We cannot say that a mathematics major is more well-rounded for taking such a course. Well-rounded according to whom? And how pompous to believe that one semester of being told what to think about paintings and what is meant to be conveyed by them makes an individual more well-rounded.

There's value in education for the sake of education. I don't argue against that and in fact, as a serial student, I'm quite fond of education for its own sake. However, the prototypical pedagogy associated with the usual assortment of subjects and course material falls short when it comes to Computer Science. Programming is a craft. It takes years to master. There are many other disciplines where this is similar, so why not change the teaching style?

Students experience a couple of introductory courses and then are expected to understand high level concepts. In my opinion, going through academia has the primary purpose of learning how to think which implicitly means application. Anyone who argues against that should be beaten with a stick. Socrates explored the meta, but he focused his knowledge on teaching and applying - even if the application was theoretical (see The Republic). If our goal is to teach people how to think, let's make use of the discipline. Sorting arrays is great. It's very helpful to understand a binary versus bubble sort and why one is faster or better in certain scenarios.

But we're a community of people that build things. We command the electron to do our bidding. We engage in digital alchemy. We are the citizens of an industry governed by Moore's law and yet we tacitly accept that there's not a better way to gain an education in our medium?

I reject that.

Let's build better coursework by building projects with students. When a student ends a semester with a (probably) working application that has been built in stages with a decent feature list there is a sense of accomplishment and a demonstrable grasp of the concepts this project instilled.

Motivations, Projects, and Resolutions

The new year is a time to reflect and decide on what to focus your energy. Or, in some instances, reflect on what projects you didn't finish. I like to look back and evaluate what projects from the past year deserve more energy, deserve to be let go, or need more development before they're viable. 

I keep a small notebook that fits comfortable in my jeans pocket for writing down ideas. Some I'll flesh out and try to find the minimum viable product (MVP). Some I simply like to flesh out and think about the 'cool' factor. Things that increase cool factor:

Real-time

"Post and response ... is kind of boring. I think most people have become accustomed to see the spinning wheel load your results."

This has got to be a trendy thing, but it also makes software so nifty these days. Things like 'real-time collaboration' or sharing or feedback. They all make software feel not only faster but engaging. Post and response, where you post data and are then rendered a results view, is kind of boring. I think most people have become accustomed to see the spinning wheel load your results. 

I know at my day job I see that wheel all the time when I'm doing tests. We're in the midst of launching a new version of our UI for the main part of our product and while it's beautiful and super fast, there's one part of the software that actually now takes longer to load - the piece that hooks into a government integration (of course). The overhaul didn't cause this per se. It was very well designed and load times across the board are reduced. But its new architecture requires a certain flow and as such, all operations for this one piece have to be re-rendered frequently. To ensure data integrity on what's coming back from the government service, I have to query them (the government) to get the latest results. This all occurs in real-time but the service is slow and unfortunately doesn't provide a fast way to get the relevant data. It was totally over-engineered. At some point I plan to change that, but for now I'm kind of stuck. Some of the load times have doubled in the new system for this one piece. 

Social

"Social also means having a community. If you build a community, meaning forming a common culture around a topic, idea, belief, etc., then you're more likely to attract more people from that pool."

Ideas these days tend to have some social component. Sharing an achievement in a weight loss tracker or a ride you just completed. Not to mention the calls to action to share that you're using a given product. Moreover, it's almost more popular now to use social authentication instead of having a database containing usernames and passwords that you have to protect. This makes sense - now the concern is not on you if the password store is compromised, you simply send out an alert email to your affected customers saying "due to the recent ... at <social network>, we'd recommend you change your password to protect your data at our site. Please click <here> to view details about the announcement from <social network>."

Social also means havinthem g a community. If you build a community, meaning forming a common culture around a topic, idea, belief, etc., then you're more likely to attract more people from that pool. Word of mouth campaigns are more budget friendly for you than spending time and money trying to advertise your project to form communities of one. It's always a good idea to test your idea against the possibility of building a community with it. Not all ideas can work that way - some are targeting very specific use cases and that's ok.

Feedback

"Feedback isn't just about getting criticism and praise from your users - it's about getting data on how they use the product."

Having an excellent feedback system users want to utilize instead of avoid is valuable. For one, your feedback loop is tight and you understand the people using your product. Getting flooded about feature X having a bug or needing new functionality? Awesome. No money just went out the door to get a 1 in 50 response rate. Fix it and fix it fast and tell your users how much you appreciate their help in making the product they use better. This reinforces the community idea mentioned above. Empower users to be able to provide you feedback in an unobstrusive way - don't hound them. Let them choose.

Feedback isn't just about getting criticism and praise from your users - it's about getting data on how they use the product. Think about analytics early. When your users creates a new item using your service and sends an invite through the service to a friend's email, track it. And track whether that friend signed up. Now you've got analytics on conversion rates from first-party referrals.

Motivations and Resolutions

This year I have some simple goals:

1. Write more blog posts

I need to be more consistent. I have a number of article topics I have written down in my notebook I would like to complete. I keep listening to podcasts that mention you have to just sit down and write to form the habit, and that's something I need to do. I am out of practice at that. Now when I sit down at my computer at night, once my son is asleep and the house is quieter, I tend to go straight to projects. I need to start scheduling time to write at least once a week, specifically in the realm of tutorials or sharing experiences about how I solved a particular problem.

2. Engage on Twitter

I was one of those that used to make fun of Twitter's core concept - share in a limited character amount. I tend to be verbose. Sometimes overly verbose. I took almost every writing intensive course offered in the political science department at the University of New Hampshire because writing several papers a week was easy for me - I could always crank out more words. Twitter limits that and forces you to focus tightly on what you want to communicate. I've been using it for a few months more actively and I find wonderful and copious amounts of information on topics that interest me. I'd like to build a bigger network there of like-minded people so that I can continue to tap that resource for inspiration, help, and the comfort of being of member of that community.

3. Pay It Forward

I have received help from a number of community members as I have learned what I know. My formal CS education is extremely limited. I'm currently completing pre-requisites to enroll in a Masters in CS program, but I earned my software developer position as a self-taught programmer. This means I've accumulated a lot of knowledge by the generosity of others. So I'd like to return the favor. I plan to be making posts discussing implementation strategies, coding tutorials, and some design information.

4. Release a Web Application to the Public

I have been working on an application for a few months, mostly figuring out what I want it to do and doodling on my sketchpads, but I've been coding in small bits here and there over the past month. I'd like to finish this application and get it out to the public. Mainly, I'd like to be a user on this site. So I believe others will want to as well.

I hope those reading this reflected on the last year and have come to decisions about what they'd like to do in 2014. If you'd like to talk to me about code, or hire me as a freelancer, please contact me on Twitter, Google+, or at my website.