Youtube rarely makes great recommendations anymore, but this time it did.
Yday found this recording of a CppCon presentation from 2016 (slides) in my feed and made some time to watch it. It's pretty long, due to an extended Q&A section at the end which I mostly skipped, but well worth watching. The presentation itself is great and masterfully explores the reasons why we're still using C instead of C++ (especially if C++ is such a better alternative?) from a psychological/sociological point of view. The presenter quickly shows that technical arguments are just not effective, and also explains why.
I've greatly enjoyed it because it aligns with some of my leadership beliefs, and also because I don't understand (or I guess I don't have the same frame of reference) people who pick C as their preferred programming language. I've done a bit of C programming while I served in Mapserver PSC maintaining the Swig/Java bindings and to sum my experience up: what's to like about a programming language in which you have to allocate, deal with and free every single character string. The cognitive load is significant and obviously distracting. If this is the deal with strings (a relatively simple concept), imagine everything else.
To help my own understanding, I've picked a few slides out of the presentation that I thought were really interesting and commented them below.
Framing the problem
First, let's frame the problem that Dan Saks wants to explore. I think this
slide sums it up eloquently, no need to comment it:
If you're arguing, you are losing
This
slide is actually before the one above in the presentation, but I chose to reorder it because I thought it would make more sense to focus our attention on this LPT "after" the framing of the problem. In the presentation it works the other way around because the presenter probably wants to build momentum.
This is a lesson I've come to appreciate in my various leadership roles. In the context of leadership/mgmt I often rephrase it as:
relying on power to get people to do things is a slippery slope.If you use power once, you are putting yourself in a position in which you are forced to use power more and more to keep them going. Eventually you complain about people not being "committed", when you're in fact the reason they are not "committed". Arguing is one of the many forms of power.
The slide has great advice, which we'll see is backed up by further slides later on.
Frame (of reference)
Such depth I would never have expected from a programming conference!
The point made by Dan Saks is that people see the world through a set of beliefs (a frame) and use that frame to makes sense of the world. What is important to note here is that
the frame makes complete sense to the person holding it, or they would not be using it.
The key to get them to change or update their frame of reference is to understand it first. Negating it (that doesn't make sense! or that's stupid! or that's wrong! or yes, but... ) and then going on a long tirade pushing your frame of reference will only push you further from your goal as your interlocutor will go into defensive mode, closing up instead of opening up.
Programming in C really means: debugging
I laughed out loud when
Dan said this, but in hindsight and my admittedly short experience as a C programmer while working on Mapserver, this seems intuitively true. Maybe this is what people mean when they say: I like C because it gives me full control!
(Like you could manage memory better than a compiler or runtime (my sarcastic thought every time I hear this "full control" story)).
What IMHO C programmers really when when they say "it gives me full control" is: I can look at the code while it's running (with a debugger) and understand it. Now that we're closer to understanding their worldview, we can make hypothesis on the best way to change it or influence it. And here Dan has another pearl of wisdom.
Inception
Like in the
movie, the best way to have someone execute something at the best of their conviction is to make them think it was their idea in the first place. Who would not do their best to execute their own ideas?
A corollary worth mentioning is that if someone comes to you with an idea and you start making suggestions for changes, etc you've just cut their committment in half. Just shut up.
Conclusion
Dan Saks then goes into a longer section in which he explains one possible way to illustrate why C++ is a better choice (not a better programming language!) than C. Personally, this is not very interesting to me, because I was more interested in the first non-technical part. Additionally, I think that in 2023 we have even better choices that C++, such as
rust and
swift (another
swift link). Unfortunately vendor support is thin to non-existent, but I believe that the first vendor to add suppor for one of rust of swift will have a huge advantage. Maybe. If C programmers get to adapt their frame of reference first.
