To be fair, function call overhead (and other classic stuff like conditionals or integer division or FP operations) is an issue when you are concerned with nanoseconds. For example, in the inner loop of a (software) renderer or anything else that is run several million times a second. Otherwise, just forget about it. There are much bigger fish to fry here.

Anyways, as a wise person has said “Premature optimization is root of all evil.” It follows directly that uninformed decisions about performance are nothing more than superstition and just as correct.

0: Well, that’s absolutely true, but what the fucking fuck about myself?! I have to see this shit day in and day out. So I’m basically doing this for myself, and the reason that I posted about it is the fact that most people (close to all) would not know it unless I said something.

1: Maybe not for you, but arguably there are some people who care enough that they normally take measure like using Tor and scrubbing their UserAgent headers, etc. Here I’m saying that I’m not letting Google have that information about anyone. Of course, my webserver still generates a log of every access, and the log contains user IP address and agent along with the URI they visited, but I’m not doing anything with that data unless there is a problem with my server (or some kind of attack.) They are there for a month, and they are deleted.

2. I don’t know what they are. I’ll check it out.

3. Thanks! “You see everyone? There’s a good blog reader! “

As for MatGill, nicknames are generally chosen by others. And I do miss those days. I really think I would have been an awesome hardware engineer if I went that way. Now I’m only a mediocre AI guy with no stamina to work.

Hi there, my name is Eduardo and I am a journalist for a newspaper in Brazil called O Estado de S. Paulo. You can reach us at http://www.estadao.com.br

I found a few weeks ago about Garshasp and thought that you guys could fit perfectly into our tech/digital life section.

So I’d like to ask a few questions by email about how Garshasp was developed, how you got together and started the project and what it was like to develop a game in Iran.

I’d really appreciate your answer. Me and my editor find the story it’s really interesting.

And congratulations for Garshasp! The videos and teasers I saw made me very curious about it.

Thank you very much!

Obrigado!

Eduardo Roberto

Also, wouldn’t you agree that even in MATLAB, knowing the way memory and cache behave, and having some knowledge about the ISA of the underlying CPU, and being intimately familiar with the workings of the MATLAB compiler/interpreter/whatever-kind-of-shit-it-has will help you?
Of course, these days the hardware is so complex and the behavior of even the lowest-level programs so unpredictable, that you already have to rely really heavily on profiling. That’s even more true in higher-level environments. But a knowledge of the hardware, the operating system, and the MATLAB compiler and runtime will definitely help you understand the behavior you observe.

“it might even be true that the better they know this mess, the better they are”

I would argue that this is not strictly a cause-and-effect scenario. Good programmers seem to also be interested in the details of the platform. But it’s not entirely essential to know about the platform to be a good programmer. Now let me add that some knowledge of the platform will help develop a “programmer’s mentality”. But the wider the gap between the platform and the software, the less useful such knowledge becomes. This is coming from a guy who once coded a RISC CPU on an FPGA and wrote a pong game on it with keyboard/mouse/monitor interface, then moved on to develop web applications with java and later moved to the fucking MATLAB. If I was a game developer, then the detailed knowledge of the GPU would have been useful, but when you have to code your shit in MATLAB, then even God himself can’t help you.

Some technologies are like leprosy; everyone just learns enough to identify them on sight and keep the hell away. But of course there is a catch. Sometimes you have to deal with it (like me) because you have to work on Windows, with Visual C++ and many third-party libraries that also use VC CRT library. The good thing is that at least most of what I use at home and at work are opensource libraries and I don’t get into much trouble that cannot be solved with a rebuild of those.

Basically WinSxS was supposedly a centralized system-wide database-like place to manage multiple versions of DLLs. In practice, it meant that you could not just distribute some of the most important shared-object-files with your application and have them just work on all versions of Windows. Your life became miserable just because some idiot did not keep her system libraries up-to-date, and even worse, you could not easily help her without going to some length in installing the right libraries in the right places and telling that to Windows. Your applications also needed manifests to declare exactly which version of the DLL they were going to link to. However, in typical Microsoft fashion, the old (almost well-understood) way that applications declared what DLL they would need (and what functions from that DLL,) i.e inside the binary’s headers, were not deprecated and were still used.
All gods of torture would have to help you if your manifest and your binary headers were not saying the same thing!

By the way, the above is not a scientific treatise on the subject. I do not, and don’t want to have any more knowledge on and contact with this subject.

But they really are.

@eam:
Do you want me to get myself killed?! And the Zippo is safe and sound, thanks for asking. Actually, it was working about two minutes after the incident.

@Amin:
I actually wouldn’t recommend this method for that purpose! It does have some effect of your hair, but not enough to balance out the fact that you are actually setting fire to your hand!