1) Whatever language you are most proficient in, that is often the most efficient tool for _you_ to get the job done right.
2) PHP is largely C for people who should never program in C (or PHP ;-). Depending on the nature of what you are doing, and if you use some basic support libraries (and the author mentions a few), your code needn't be that much more work to get done than any other language. In reality, it's not the typing in the code that tends to matter anyway.
3) If you are, for whatever reason, interfacing with a lot of systems components or C API's, there can often be a pretty big win using C as you avoid all the pain of language impedance mismatch.
4) Depending on what it is measured against and the nature of the problem, C can often be significantly more efficient, not just in terms of CPU, but the all important memory consumption these days. Depending on the problem, that can be highly highly relevant.
5) Portability. Okay, you can stop laughing now. ;-) Seriously though, if you talk to mobile game developers who've tried it, C and C++ tend to be the best languages for targeting a plethora of platforms. That goes double or triple for embedded systems. It's really not that hard these days to have very portable C/C++ code, and platforms that insist on making C a second class citizen tend to do themselves more harm than good (even Android ultimately relented).
Now, that's not to say that I think everyone should be doing their web projects in C, but depending on the context (the problem, the available tools, and the parties involved), it might well be a good choice.
Sometimes it is hard to beat just doing it in C.
The attraction of C to this mindset is that it can be transliterated (vs translated) into machine code. This transliteration is so straight forward that if you know C and are familiar with the basics of the instruction set you can do it by hand.
The survivalist mindset fears dependence. VMs and interpreters are the essence of dependence. So those with this mindset always seek find a way to do it in plain C when they can.
Programmer writes an entire operating system from scratch in case all the copies of Linux in the world are deleted.
A web developer writes his entire application in x86 assembly to prepare for an event in which all the world's compilers and interpreters disappear.
And more...
Except modern processor architectures are no longer a one-to-one correspondence between Assembly code and C.
Nowadays not any longer if you want to write code that takes advantage of branch prediction, speculative execution, cache lines, vector units, GPGPU ...
Just watch this Going Native talk on how sometimes generating code that is 4x bigger than the direct translation can yield up to 30% performance increase.
http://channel9.msdn.com/Events/GoingNative/2013/Compiler-Co...
I'll concede there's more fine variations to think about than with an 8-bit system, though, with the padding, etc.
I find that an odd perspective. If you can rewrite your VM/interpreter from scratch, are you really dependent on it?
I guess I see it less as "use some black-box language runtime to solve this problem" and more as "I will use plain C to solve this problem... by writing an interpreter in plain C for a good DSL to solve this problem in, and then writing the solution in that DSL. But, you know, someone else already did the first part. I'd do it myself if they hadn't, though."
Unless you have lots and lots of time to spend micro-optimizing everything, you'll get better performance writing in Haskell.
> Maybe there are good C libraries that solve your problem
Maybe, but most of the original post is about how lacking the library ecosystem is. I can well believe that you might have some useful domain-specific libraries for the backend, but that calls for writing a backend service in C (and using something like thrift to call it), not "web programming in C".
> Maybe you just want to learn C better.
If you try and learn on a fake problem you'll pervert your learning. Web development is not amenable to idiomatic C (Cello is a good example - it's not appropriate to the sort of problems C is appropriate for, and if you learn C based on using Cello it will not make you a good C programmer).
If you're an experienced C developer you know what's fast and what not (and you know about things like cache coherency). You don't spend that much time on 'micro optimization' because you know how to layout your memory in the first place, etc.
This is stating the case too strongly. Performance wise, the naive, unoptimized solution in Haskell (ex. using the default String type instead of Data.ByteString, number crunching using lots of iteration\recursion) is going to be much slower than than the naive, unoptimized solution in C or Lua. However, for a given subset of the range between "zero" and "micro" levels of optimization, for a given problem, Haskell may still represent the local maximum.
This is just not true. Haskell is terribly (3x,4x) slow comparing to C for the most trivial of computing tasks. You can probably get it to 2x or 1.5x by giving up all the lists and other default data structures as well as writing everything in procedural (by monads) way. If you do all this you can get close to C performance but you just gave up all the advantages of writing in Haskell in the first place. And you are still slow as hell.
And I don't know why it would need to be so horrible to program web in C, once you have abstracted out common web tasks, using them is as simple as using them in any language, just call a function.
Of course it's usually easier and safer to just throw more CPU at it, but at a certain scale that might get too expensive...
No, seriously, why? As far as I can tell, his entire argument is "web development in C is terrible and limiting, but not as quite as terrible and limiting as you might think." I guess that's good news if terrorists are forcing you to write websites in C, but I don't see even an attempt at explaining why you would choose to do this.