Sign In:     


Forum: General Discussion

Topic: Writing Scripts
Virtual DJ has users all over the world, all who speak & read different languages. When someone in France writes a script can they write it in French or do they have to learn the English to write it?
 

Posted Wed 01 Jul 26 @ 10:11 pm
Answer = no (to French usage)

This is how all programming languages of current day work btw so this isn't something unique to VirtualDJ.

A programming language implementation is different from an AI model, which can be trained to accept different natural languages to do the job.
 

I believe it's English only.
 

script is its own language, it's based on English.
I worked with a blind guy who used a screen reader, and he only spoke Spanish.
I had to explain every verb name like

" 'param_cast' param is short for parameter , cast means we want to send a value somewhere else "

so his translator got it

hard work on my part, hard work on his too, but he got the fundamentals.
it's just how it is, I think most code is written in 'almost English' no matter where the developer is.
 

Tbh that's where AI is bridging the gap - you explain (in any natural language is trained to understand), and the model takes care of knowing how to create the programming language implementation.

It's sort of what the desire was for programming all along - being able to tell the computer what to do efficiently/in a language that the programmer is fluent in (with all the required translations being handled).

Until fairly recently, that meant understanding a computer (normally high-level) language syntax, benefits and even computer architecture specifics. Now, depending on the model used and its training, you can actually attempt to achieve that goal more directly (i.e. woth natural language) without requiring computer language specific knowledge.
 

DJ VinylTouch wrote :
Tbh that's where AI is bridging the gap - you explain (in any natural language is trained to understand), and the model takes care of knowing how to create the programming language implementation.

It's sort of what the desire was for programming all along - being able to tell the computer what to do efficiently/in a language that the programmer is fluent in (with all the required translations being handled).

Until fairly recently, that meant understanding a computer (normally high-level) language syntax, benefits and even computer architecture specifics. Now, depending on the model used and its training, you can actually attempt to achieve that goal more directly (i.e. woth natural language) without requiring computer language specific knowledge.


You describe this as something positive, but it isn't.

I'm not a gate keeper, but if you lack "computer language specific knowledge" then you're just going to crash and burn more quickly than anyone else..

AI makes mistakes. And it will keep making mistakes even while getting better.
Damn, humans make mistakes.
And then they spend hundreds of hours trying to debug the code they wrote.
So, imagine you trying to debug a code given to you by AI, without having "computer language specific knowledge"
It's not going to go anywhere..

Whether we like it or not, learning to program/script in a language, requires you to learn that language.
You don't have to know English to learn to use "concat" but it definately helps as it makes it easier to remember which function you need for a task by it's name.
 

@PhantomDeejay I'm actually neutral on this, but it is a positive depending on how you look at it.

Remember the history of programming languages essentially describes the quest to find more concise but precise ways to tell the computer how to do the job (Machine Language/Assembly Language to various higher level languages for which a language construct translates/compiles to many lines of assembly instructions).

People did initially reject higher level system languages too (and preferred to write in say, assembly), as the compilers initially produced inefficient code, but they did get better over time, to the point that current compilers would probably generate better assembly code in far less time than a seasoned assembly programmer would and no current programmer would initially start a regular project in assembly due to that.
It's the same with Generative AI, and the better the model knows the problem space/the better you describe the task (which is helped with knowledge of computer programming as you pointed out), the better its output would be, and that, combined with the knowledge of programming language historys is why I loosely describe it as the next iteration of programming in a sense.

AI will never replace the programmer (that's not the intent of the comments at all), but it has the potential to greatly help the programmer in his/her role.
 

I remember spending hours at my friends house years ago, as he had a computer (Atari 800 I think) and a book with pages of code that you could type in to create games.

I don't think we ever got one to work once we'd finished entering all the lines of code.

🤣
 

Yep there was no IDE wth intellisense back then 😂