At what point do syntaxes in programming languages become important?

Question

I’m in the mood to discuss the subject of when a syntax of an N programming language becomes important for the production of systems.

For example: A language with simple syntax and easy to learn, interferes with the choice of which language I will build the system I want?

Is it also possible for a systems developer to exchange his programming language even with a well-developed system, for reasons of some great urgency? If so, what kind of emergency could these be?

Answer 1

We do not have discussions here, we give objective or subjective answers based at least on the person’s experience.

I am a "scholar" of programming languages and for a long time I was a lover of syntax. Time made me see that only semantics really matter. The syntax has the function of making the programmer’s work pleasant, a good syntax is a good UX for the programmer, but it rarely influences the final result. Of course some syntaxes may favor a more correct code or encourage a more performatic code or influence at other points, but it is not alone that defines this.

You know that thing where you use the best tool for the problem? The syntax of a programming language is never an important point in this, or even the language as a whole makes little difference except in extreme cases (it will not write an operating system in Javascript or make a front web with C, gives but is not suitable).

You can decide whatever criteria you want to decide what language to use. Being simple and easy to learn are two different things, many think it’s the same thing. Simple is a somewhat ambiguous term, I imagine it’s easy to use. Always prefer the easy to use because most of the time you uses the language, little time you learn, so being easy to learn is a desirable feature by naive or because one will use it once and throw away.

I don’t take it to iron and fire but the best tool is the one you know. If it makes little difference use the one you dominate (and must dominate, not only know). Of course just because you know PHP will not make a desktop software, or make web what should be desktop (this last case is change the problem mistakenly so the language fits your taste).

But ease of use and learning should be 2% syntax and 90% semantics and 6% other things (numbers kicked). Syntax people choose by taste.

For example, PHP is taking a turn for its evolution and the way people use that should change language, PHP is becoming a Java with PHP syntax, it makes no sense, if you want to program in Java programe in Java. If you don’t like it, go to C#. If you still really want PHP syntax go to Hack.

Changing the language in the middle of the project does not make much sense for all that I have said above. You should say what a rush this is, I can’t see any.

So I wouldn’t advise going into low-community languages, but it’s still not usually tragic. There’s Cobol, there’s still a lot of people using Clipper/Harbour modern form, and even Delphi, which is a proprietary language and people are running from it, still lives well.

It is very rare, not to say impossible, but if you really need to change a language because something has not been feasible can still do it only at a specific point and maintain what has been developed. I just hope this doesn’t serve as an excuse to adopt microservices because there is a way to solve this in a simpler way.

We are, or should be, engineers, we are problem solvers, there is always a way to solve, some better, some worse. And when we can’t solve, we call those who can. Complex things must be solved by seniors. Junior might try, but he’ll risk doing worse.

Solutions should be given by observing the concrete case, there are no magic answers that solve everything, if it were so would not need so much developer.

Worry about the semantics of the language, its general philosophy and how it is implemented, and what may in some cases be even more defining, its standard library and those of existing third parties (although this has become more linear in most cases).

I like C# because it better combines all these points for almost any kind of application (it’s even for any type to give up the maximum suitability), and has syntax of good taste :). Each has to find his own way and live with that choice, or exchange if necessary.