How many people code in C these days, and if so, why?

[IDEngineer]

And despite how I typed my examples above (which I did for clarity), I actually always type conditionals which contain a fixed value like this:

if (true == boolean)
if (5.3 == fVar)

Why? Because if I accidentally leave out one of the equal signs, the compiler will conveniently report it as an error. Typed the "traditional" way, the compiler will simply (and silently) execute the assignment and then resolve the conditional based on the zero vs. nonzero (false vs. true) numeric value of the assignment. This works in Java and other languages too. It "reads" a bit weird but is perfectly logical and lets the compiler check my work in an extra way for free.

I teach others to type this way and (especially for newbies who haven't internalized the whole "double equal sign" thing) it catches a LOT of mistakes.

[Siwastaja]

The problem is that in C 'if' can be used with any type.

IMHO, you should have a naming convention (explicit, or more subtle) so that you know the types, because the problem isn't limited to the if clause and comparisons.

Once you fix that, you get rid of needing the comparisons to other values just to demonstrate the type!

if(isDog) // any isSomething is bool
   printf("%d", numDogs); // num anything is int

or explicitly,
if(b_dog_exists)
   printf("%d", i32_n_dogs);

whatever your coding style or naming convention is, I don't think writing excess comparisons just to "demonstrate" the type is a good solution. This information should be available in some other way.

[Siwastaja]

But it's not particularly ugly. Some people may find it quirky, but that's largely questionable.

I agree, there's nothing ugly. We don't need to hide all low-level details in near-neurotic way; only abstract things when there is a clear benefit and good reason. Just use a simple comment to document dual-use as a flag and counter (I do that all the time, not only for performance, but because it makes sense.)

Similarly, I accept, and even prefer, to have my plumbing visible. It does not distract me, it has a purpose, it's needed, I use it, and when visible, I can see if it leaks. Some people prefer to hide plumbing inside concrete though because it's "ugly". Then replacing it is a huge project (it's known here as a personal tragedy, called "plumbing renovation", requiring evacuation for half a year and almost doubling the estate price.)

[SiliconWizard]

And despite how I typed my examples above (which I did for clarity), I actually always type conditionals which contain a fixed value like this:

if (true == boolean)
if (5.3 == fVar)

Why? Because if I accidentally leave out one of the equal signs, the compiler will conveniently report it as an error. Typed the "traditional" way, the compiler will simply (and silently) execute the assignment and then resolve the conditional based on the zero vs. nonzero (false vs. true) numeric value of the assignment. This works in Java and other languages too. It "reads" a bit weird but is perfectly logical and lets the compiler check my work in an extra way for free.

I teach others to type this way and (especially for newbies who haven't internalized the whole "double equal sign" thing) it catches a LOT of mistakes.

Yes, I don't like this style, but it's a pretty common one to avoid errors.

Note that many decent/recent compilers will issue a warning if you write 'if (a = 5)' (and you shouldn't ignore warnings!), and otherwise static analyzers will also spot that. But if you don't use any, then yes I agree.

[SiliconWizard]

But it's not particularly ugly. Some people may find it quirky, but that's largely questionable.

I agree, there's nothing ugly. We don't need to hide all low-level details in near-neurotic way; only abstract things when there is a clear benefit and good reason. Just use a simple comment to document dual-use as a flag and counter (I do that all the time, not only for performance, but because it makes sense.)

Agreed, but I don't quite agree with the low-level nature of this particular point.
If anything, I consider the two states of original C "booleans", zero and non-zero, to be a little more abstract than just 0 and 1. Weirdly enough, many people seem to have a problem with that. Guess we're all different.

[IanB]

Yes, I don't disagree with the fact that your if("string") example is valid C (equivalent code is valid in many other languages, as well; and equally not making any sense), as is your modified code in post #204, which nobody has commented about, I was just looking for the point you are trying to make with your messages, but now that you have admitted you have none, and didn't even intend to, it explains many things, and it is unnecessary to discuss any further.

Apparently I failed in my communication somehow. I don't know why everyone found it so objectionable.

I did have a point:

C does not have a boolean type. A C compiler will definitely complain if you try to assign a string  (char *) to a non-string type.

I was just checking this assertion, since later C standards did add a boolean type, and it doesn't in fact complain if you assign a string to it.

Instead of saying "oh, that's interesting" everyone started beating up on me. So OK, it's not important, nobody is bothered and it's not worth arguing about.

[Tomorokoshi]

For levity's sake this seems appropriate here:
https://xkcd.com/386/

Anyway, sure, C has its +'s, ++'s, -'s, and --'s. For practical reasons there are not a lot of alternatives for embedded work, a lot of code base is available, etc. I can't afford 100 MB to support "Hello World" with Java or other wanna-be bloat.

Regarding the curiosity from the link a few pages pack regarding the apparent rise in popularity of C and Python along with the drop in popularity of Java, it makes perfect sense given the benefits and synergies pointed out by others of using C and Python together.

"Using the right tool for the right job" requires a set of value judgements that don't fall nicely into macroeconomic categories. Many of those are bypassed for non-technical reasons, having to do with the reality and practicality of project cost and scheduling. With a lot of code base and engineering talent invested in C, choosing some other language on technical merits alone will rarely make it through until a lot of off-budget investment is made, such as through independent projects or some game-changing set of tools.

[engrguy42]

IanB, between you and me...word to the wise...

Fanboys. This place is crawling with 'em. Meters, scopes, power supplies, you name it. And even software. Go no, right?? Yeah it's true. And who would'a thought there were even C fanboys?? 

Anyway, it's a bit like going into a bar in Dallas while the Cowboys are playing in the Super Bowl, and shouting "Cowboys SUCK !!". They'll rip you a new one.

So my recommendation: Just apologize, post shit like "C RULES !!!  " and slowly back out and move on.

And watch your back..........

 

[bd139]

It was quite objective and interesting until a couple of members crawled in...

[engrguy42]

It was quite objective and interesting until a couple of members crawled in...

I apologize. C RULES !!!! 

[bd139]

Common Lisp rules   

[Karel]

What about Borland Delphi? 

[janoc]

FYI:

C is actually rising in popularity - mostly thanks to embedded development.

https://www.tiobe.com/tiobe-index/

Last month growth of C is almost 3% and it has landed back in the 1st spot, overtaking Java.

[SiliconWizard]

@janoc: yes, I think this link has been posted a number of times already in this thread. As I said, to be taken with a pinch of salt.

Otherwise, and especially for engrguy42: I don't know how surprising that would be that people mostly focus on talking about C in this thread. Overall, the posts looked pretty reasonable to me. And you can't really complain about people being on topic, since this is the very topic. It was not about "what programming language do you guys use these days?" (which is kind of a popular topic and has already been asked numerous times I think.)

As Tomorokoshi posted, some people seem to prefer trying to prove others wrong than staying on topic. Oh well.

[engrguy42]

Overall, the posts looked pretty reasonable to me.

Of course they do... 

[Wolfgang]

Overall, the posts looked pretty reasonable to me.

Of course they do... 

IMHO, religious computer language discussions are a complete waste of time -
better spent to solve a real world problem.

[nctnico]

What about Borland Delphi? 

For that we would need a vomit emoticon. I had the misfortune to inherit a larger project once. What a mess. Pascal with C-isms bolted on. I have used Borland Pascal for a while an dabbled in Delphi as well but quickly found out that Pascal is not really suitable for any real programming work where you need to restructure a program half way sometimes. Having 4 or 5 different string types also doesn't help. If you need to resort to C-isms anyway then better use the real deal (C or C++)... If you don't want to use C/C++ then look a C#, Java or Python.

[coppice]

Common Lisp rules   

Rubbish. Snobol rules.

[SiliconWizard]

PL/M rules.

[Wolfgang]

Its endless. With all the time and effort wasted here in this forum you could have coded a graphics package in COBOL. Just as futile, but at least *something* works in the end 

[coppice]

PL/M rules.

PL/M, MPL, and a few other derivatives of PL/1 were created by microprocessors vendors in the mid to late 1970s, to support bare metal programming for their products. If C had more presence at that time these things would never have existed.

[SiliconWizard]

PL/M rules.

PL/M, MPL, and a few other derivatives fo PL/1 were created by microprocessors vendors in the mid to late 1970s, to support bare metal programming for their products. If C had more presence at that time these things would never have existed.

 

[engrguy42]

Its endless. With all the time and effort wasted here in this forum you could have coded a graphics package in COBOL. Just as futile, but at least *something* works in the end 

Yeah, no kidding...the horse died like 10 pages ago.

Well I spent some time today buttoning up my data acquisition app by making a custom icon.

Oh, and my app has some booleans, and in spite of the incredible complexity surrounding booleans  they worked fine. I mean, like, what's to discuss? They're, like, true...or...false. 

"if (plotIt && !plotStop)"

On second thought, I'm sure that will start even more boolean muscle flexing but anyway...

[bd139]

Now write an installer in WIX and sign it with an EV code sign cert 

[engrguy42]

Now write an installer in WIX and sign it with an EV code sign cert 

Why?