New programmer, language question

[Guest]

26 minutes ago, reniat said:

I'm getting a lot of corporate-softeware cynicism. I think one of the most valuable things you can do as a developer has nothing to do with language: breeding culture. 

Some people are lucky to be at a job with a strong quality minded development culture, but even if you're not you can slowly bring positive change in that direction. A simple example would be helping the developers start using git if they don't have a good SCM in place, or getting people behind adding code reviews to the process of each task if they aren't already, etc.

There is a lot more that you can do to improve the quality of your co-workers code than you think, and doing so can have far more leverage than you writing quality code by yourself. If a company truly doesn't want to change, perhaps its time to find a place with better culture, but you'd be surprised at how much a bit of firm patience can be in introducing systemic changes.

Nah I'm just unoptimistic, but I suppose there are good examples too out there with also software engineering procedures during development In co-groups

 

26 minutes ago, reniat said:

Be careful with your wording here. the reason you almost reignited a language religion war in here is that you more or less called java developers code-monkeys. Also, as mentioned, I think you need to update your definition of script kiddie

Even if I do not like OOP in general, I actually like Java when is used correctly, but they are so rare as correct programmers in general. Nah I was a more generic than this, code monkeys are independent of the programming language, some of the time they are both script-kiddie and monkeys lol... It's just that java has one of the highest number of them
Oracle developers for example did a good job

A bad example was my high school professor who taught me imperative programming in Java and c#

[Dat Guy]

1 hour ago, Lukyp said:

A proficient programmer indeed can write all the languages correctly

There is a large difference between correctness and elegance.

[reniat]

1 minute ago, Dat Guy said:

There is a large difference between correctness and elegance.

Are you saying good developers can't write elegant code in all major languages?

[Dat Guy]

There are languages which make it syntactically impossible to write elegant code. In my very own opinion, Python is among them.

[Guest]

39 minutes ago, Dat Guy said:

There is a large difference between correctness and elegance.

17 minutes ago, Dat Guy said:

There are languages which make it syntactically impossible to write elegant code. In my very own opinion, Python is among them.

In this case I agree, other times you can do both

 

[reniat]

21 minutes ago, Dat Guy said:

There are languages which make it syntactically impossible to write elegant code. In my very own opinion, Python is among them.

I'm just curious, can you show an example of what you consider elegant code? To me, elegance is writing code that does a lot (is accurate, is performant, etc.), while also being clean and maintainable.

[Dat Guy]

Performance and maintainability are mostly language-related. Python won't let you do that. You cannot just add an "if {}" around a block without having to rearrange a lot of code and/or using a pseudo-intelligent IDE which does all the work for you.

 

I would argue that most Scheme applications are elegant according to your and my definitions because of the sanity of the language.

[reniat]

24 minutes ago, Dat Guy said:

Performance and maintainability are mostly language-related

I'm not sure I agree with that. Well written java can be faster than poorly written assembly, depending on the problem (an super basic example would be someone making an O(2N) solution in java vs a O(N²) solution in assembly where N is very large). Also, you can write clean code in any language, and really unclean code in any language. Therefore I would argue that performance and maintainability are mostly developer related (though specifically with performance there would obviously be additional weight towards the tech stack).

 

I don't agree that python "won't let you" write maintainable code that performs to specification. It absolutely CAN be written in a poorly maintainable style with bad performance (list comprehensions can be abused to hell and back in the clean code dept.), but that's true for literally every language.

 

24 minutes ago, Dat Guy said:

You cannot just add an "if {}" around a block without having to rearrange a lot of code and/or using a pseudo-intelligent IDE which does all the work for you.

 

i'm not sure exactly what you mean here. Are you just talking about lacking curly brackets? There is legitimate criticism in the whitespace design of python, but it doesn't result in python being unable to write clean/maintainable code, it just means a competent developer would make sure to keep their whitespace consistent as part of writing in that language.

[Dat Guy]

1 minute ago, reniat said:

a competent developer would make sure to keep their whitespace consistent as part of writing in that language.

Which largely increases the workload for even the simplest refactoring measurements.

Curly brackets are not the problem - "whitespace is syntax" is a horrible mistake though and it can't be fixed. Python is considered broken by me.

 

2 minutes ago, reniat said:

you can write clean code in any language, and really unclean code in any language.

Some languages make one of those things inherently complicated though.

[reniat]

Just now, Dat Guy said:

Which largely increases the workload for even the simplest refactoring measurements.

Curly brackets are not the problem - "whitespace is syntax" is a horrible mistake though and it can't be fixed. Python is considered broken by me.

In my opinion you are putting far more value into this than it deserves. It really is not hard at all to keep things consistent. I really can't see how this would result in "large increases to your workload for even the simplest refactoring measurements".

 

Can you give me an example of this? A code snippet that is really difficult to refactor in python?

[Dat Guy]

I don't know what you consider to be "really difficult". As I said, the simplest example would be something like:

foo = True  # can be a runtime variable or whatever
# 2000 lines of code or something
# now execute these 2000 lines of code only if foo == True...

A language that makes simple tasks like this one unbearable in Windows Notepad is a language that is broken. YMMV (and it probably does).

[Hi P]

Do you guys mind some questions? @Lukyp @reniat

 

1.- What makes a programmer a good one and not a script kiddie?
(generally speaking) does it mean any programmer that doesn't know Assembler, C and C++ is a script kiddie?

 

2.- Realistically speaking, what advantages does someone with knowledge in Assembler and C have over those who don't?

 

3.- Is large code always bad even if separated by modules?
(I don't know what modules are called in other languages, that's for Python)

 

4.- Are programs written in different languages or only one?
This is really confusing for me, because for example, if I write scripts in Python (which is slow) what benefits would there be by using C++ along with it? if at the end of the day Python would still be slow, I hope I made sense.

 

Thank you

[Franck]

31 minutes ago, Hi P said:

1.- What makes a programmer a good one and not a script kiddie?
(generally speaking) does it mean any programmer that doesn't know Assembler, C and C++ is a script kiddie?

A good programmer knows what the code he write does, or at least have a very good understanding of what it does and he is constantly learning in a said specific language to make his code better. (this is not language specific. You can apply to all languages). In short, never stop learning.

 

31 minutes ago, Hi P said:

2.- Realistically speaking, what advantages does someone with knowledge in Assembler and C have over those who don't?

Real world wise, not much other than you can code in Assembly and C if a project require. C has some more common use if you code in other languages and a library is coded in C. Then knowing about pointers will help you interface with those libraries without hitting problems. But nothing a quick google search can't teach you in an hour or so.

 

31 minutes ago, Hi P said:

3.- Is large code always bad even if separated by modules?
(I don't know what modules are called in other languages, that's for Python)

Having one big file just make it more difficult to go through and find yourself in it. There are other reasons like each module has a concept or a task it is meant to do. If you have a "module" that does file sorting it doesn't have anything to do with displaying graph so it should be in separate "module". There are lot's of other reasons as for why a modular approach is favored but you can read a lot on the web about this. As for large code base it's not really a thing if that is the minimum required to make it work. When you work on large project you can have very large codebase that do very simple things and size grows rapidly even if you refactor as much as you can. As long as it make sense as a whole there is "usually" no problem there.

 

31 minutes ago, Hi P said:

4.- Are programs written in different languages or only one?
This is really confusing for me, because for example, if I write scripts in Python (which is slow) what benefits would there be by using C++ along with it? if at the end of the day Python would still be slow, I hope I made sense.

That's a 2 part question. Are program writing in different languages, to that, from personal experience i would say that all our projects written by the developer at 99% of the time are coded in a single language. But that is written by the dev, nearly every single time those project use external libraries that are made with other languages so it depend on how if you consider the whole final project as a whole or just the part developed by the programmer. Although this depends on the language, chances that if you are working on assembly you will not use anything else other than assembly ever.

 

For the part of the question about if Python call C++ would there be benefits then yes and "maybe" but mostly yes. C++ is one the fastest language there is. I did say "maybe" as i have been told that TensorFlow is supposedly faster in Python than C++. Personally i doubt it but until i actually test it myself or one of my guys start dipping into it i wouldn't make a definitive call on that.

[reniat]

32 minutes ago, Hi P said:

1.- What makes a programmer a good one and not a script kiddie?
(generally speaking), does it mean any programmer that doesn't know Assembler, C and C++ is a script kiddie?

Not knowing a single specific language DEFINITELY does not make you bad developer, and that's what i've been trying to avoid throughout this thread. You SHOULD absolutely be familiar with the concepts of assembly and how a processor works, but you don't need to be able to write complicated systems in low level languages at the drop of a hat if you don't actually intend to write low level systems.

 

There are LOT of definitions of a "good" developer, and a persons definition kind of says something about them as well. To me, a good developer has the following qualities:

1. They have a solid understanding of the language and environment they work in regularly. For example, I would expect a good front end web developer be very familiar with the virtual DOM, how javascript engines work, be deeply familiar with any REST apis involved with their website, etc..
2. They seek out knowledge to always be growing what they know, both in the space they work in as well as the space around it.
3. A dedication to systemic quality. By that I mean, seeking out places in their dev environment that could be improved to increase end code quality. This could be adding build steps that are more restrictive to warnings/errors, or encouraging everyone to adopt practices like mandatory code reviews if that isn't already in place.
4. Be able to learn domain knowledge. In other words, learning a codebase so that they can keep nuances and existing "codebase quirks" in mind to avoid defects that can come from working with older legacy components that aren't as intuitive.
5. Being able to mentor is another big one. this isn't REQUIRED to be a good developer, but it's SO much more valuable to your team if you can not only do a good job, but help others do a better job as well.

32 minutes ago, Hi P said:

 

2.- Realistically speaking, what advantages does someone with knowledge in Assembler and C have over those who don't?

Depends on what's being worked on. If you are having performance issues in Java related to garbage collection, having a great knowledge of the C memory model can really help in understanding the underlying issues (99% of the time it's copies of the same thing being made but never collected). You would also want to be familiar with how Java garbage collection works in that case, but even that is easier to grasp if you know what references/pointers are.

 

That said, if you are a front end web developer, low level knowledge won't help a lot. You are really abstracted from the system at that point, and what's more important is writing reusable clean components, how to leverage the virtual dom to increase performance by not re-rendering unnecessarily, as well as an understanding of UX design. 

 

It's never a BAD thing to know though, which is why every school that I know of will have a course on assembly/processor architecture.

 

32 minutes ago, Hi P said:

3.- Is large code always bad even if separated by modules?
(I don't know what modules are called in other languages, that's for Python)

Large code isn't automatically bad. Some systems are just large and need a lot of code. 

For example, I work in a large enterprise software company, and we have multiple codebases each with millions of lines of code. The important thing is not the overall size of the codebase, but how well it's laid out. Those millions of lines of code are spread across many components and sub-projects which let developers only worry about small pieces at a time. This comes down to having a solid architectural design that is maintainable, which is very important when dealing with OOP. It can be real easy to make an OOP systems that seems great now but is just a pain in the ass to extend in the future. When in doubt, try to break things up into logical chunks that are as small as possible without being so small that they don't make sense by themselves.

 

32 minutes ago, Hi P said:

4.- Are programs written in different languages or only one?
This is really confusing for me, because for example, if I write scripts in Python (which is slow) what benefits would there be by using C++ along it? if at the end of the day Python would still be slow, I hope I made sense.

Depends on the tech stack, and what the product is. Here's a couple examples:

 

A python script might use precompiled C modules to increase performance, so you may write in C/C++, and then use that module in python code.

 

Similarly if you are working on a desktop application, you might use something like Electron, which uses javascript for the front end pieces but you can write C/C++ modules which the javascript can consume. 

 

If you are a full stack webdev, you might work on javascript for the front end, Java or node JS or something for the server itself and then maybe C/C++ for any supporting APIs (you could also use C/C++ for the webserver and Java for a rest API as well, this was just an example).

 

Maybe you work on compilers, and you might use a combination of C++, assembly, and an intermediary language that the compiler uses.

 

Maybe you work on a desktop app in C++, and have performance critical portions written in assembly.

 

The point is, there are TONS of ways you could inter-use languages. This is why it's far more important to understand the principles of good development rather than be hyper focused on the languages.

 

[straight_stewie]

Just to add some points to previous answers:

 

3 hours ago, Hi P said:

What makes a programmer a good one and not a script kiddie?

I think we are lacking a solid, unambiguous definition of "script kiddie":  "One who writes software by copy-pasting others solutions into their code base, especially with little or no understanding of how those solutions work".

 

3 hours ago, Hi P said:

2.- Realistically speaking, what advantages does someone with knowledge in Assembler and C have over those who don't?

One benefit will be that sometimes you will have the questions: "Why do these two seemingly equally as fast algorithms perform so differently", or "why is this part of the language the way it is, x would be easier than it currently is". Knowing how the processor and the things built on top of it work usually yields an answer to those questions.

The above is somewhat of a similar idea to how a race car driver probably does a better job if he understands some things about the car being driven. It's not required that the driver understand anything about it, but it can help them to make better decisions throughout a race.

 

There are other reasons to learn how processors work, but by and large you shouldn't be worried about those right now. Truth be told, I know assembly (above and beyond my microprocessors 101 course) solely because I'm just downright interested in how processors work. I'm just a hobbyist, so take this with a grain of salt, but I have never run into a situation where I had a dire need to know assembly other than to satisfy my insatiable curiosity.

That isn't to say that it's not of extreme importance that someone understand assembly. After all, someone has to write compilers.

 

3 hours ago, Hi P said:

3.- Is large code always bad even if separated by modules?

No. Commercial software is often in the millions of lines of code. "Lines Of Code" is an oft used but nearly useless software/developer metric. At the end of the day, the best program is the one that solves the problem it was written for within the constraints under which the program must execute. Finding such constraints is an advanced topic, more related to the software development life-cycle than learning how to program.

For a simple example, many compilers will output the same instructions for both programs below:

// program 1
int a = 5;
int b = 3;
int c = a + b;

// program 2
int c = 5 + 3;

That is, the compiler will notice that a and b are not used anywhere else, and simplify it into int c = 5 + 3. Then it will notice that it's a simple operation on literal values, and go ahead and precompute the result, so it will finally output int c = 8. In other words: The code the processor executes is identical for both programs above.

 

In general, you should write your software in such a way that it's easy for humans to read. Your definition of "easy to read" will change as you learn and grow as a developer.

[wasab]

12 hours ago, Lukyp said:

I think that discussion before is more related to: Is C widely used and has active important projects?
Absolutely, most of the operating systems, together with C++, some really big and crucial applications, etc...
And
Do companies prefer you to know Java and/or other "consumer grade" programming script kiddie languages commonly used by code-monkeys because that is what the market is interested on? Like the usual "android/iOS programming" "web development" rather than C, even if is better and could lead significantly efficient projects but it's more like intended to be used by people that actually know about programming? (I mean, you could do everything with C and even better, it's just that people are only interested in give products that do what they do, even if badly)


Probably, and that may be the reason why stack overflow has more common "javascript" and the others in that graph languages, not just because is better, it just become more "used" or to be precise more scriptkiddie/codemonkeys friendly interested in working for "companies", probably the quality of their code isn't that even high and they need to continue to ask questions

Actually, its because 

1) javascript has no competition. If you are gonna do web front end, you will be stuck with javascript. But if you wish to challenge the w3c standards and say fck you to the all the browsers which are coded to render javascript then go ahead and use another language like C.

 

2) desktop applications are slowly dying. Many major applications are being moved into the cloud. e.g. Microsoft offices suite, all google applications, heck, all the consumer software on chrome OS are web apps. Excluding android apps and linux apps now run on chrome OS of course

 

3) All the MONEY are in the cloud. Nobody pays for traditional desktop app anymore. Even for mobile apps, users expect them to be free. I mean when i go on google play store, i honestly do expect these apps to be totally free and wouldnt be paying a cent for any apps that arent. Note, games are exceptions.

 

All these lead to the current state of affairs in which javascript, ASP.net, Java EE,  sql, ect are the main software stacks for web development and cloud computing and thus high on the popularity list while the low end ones like C++ and C are pretty much niche into areas like OS and graphics-heavy 3D games, neither of which can be moved into the cloud, although if they can actually move an entire operating system or tipple A games into the cloud, that would be awesome. 

[reniat]

6 hours ago, wasab said:

Nobody pays for traditional desktop app anymore.

just because people aren't paying directly for it, doesn't mean developers aren't going to be  paid to make them. look at slack, discord, spotify, etc. those are all relatively brand new desktop apps installed on millions of desktops that are essentially "free". the desktop space is definitely changing (looking at things like Electron), but we are VERY VERY VERY far from PCs being more or less an OS that simply connects to cloud apps. There are far too many applications where it simply doesn't make sense to pay for and manage persistent connections, let alone cut off your target consumer in the event of poor connectivity. 

 

Also, I feel like we had finally begun to move on from the language religion war in this thread. The post you quoted had already been talked about and moved on from lol

[Franck]

12 hours ago, wasab said:

although if they can actually move an entire operating system or tipple A games into the cloud, that would be awesome. 

There was a video on that actually not long ago on the main channel. It was a service that actually does that. Someone can surely pinpoint the exact name of the service

[straight_stewie]

13 hours ago, wasab said:

Nobody pays for traditional desktop app anymore

Yes, they do.

Trivially, a webserver can be thought of as a "desktop" application that handles file and data processing requests.

Less trivially, lot's of industries use "desktop" applications for all sorts of things. If we look at even a single industry we can find TONS of what one could call "desktop applications". As an example, let's look at the ubiquitous convenience store:

• Point of Sale equipment, like cash registers
• Inventory management systems, which maintain a local database and front end for entering information
• Security camera systems, which maintain a frontend that allows you to select from many views, and a backup and storing system.
• Financial management software, which helps to generate financial reports
• Employee management and scheduling software.

[Dat Guy]

14 hours ago, wasab said:

Nobody pays for traditional desktop app anymore.

I do.

[wasab]

3 hours ago, straight_stewie said:

Yes, they do.

Trivially, a webserver can be thought of as a "desktop" application that handles file and data processing requests.

Less trivially, lot's of industries use "desktop" applications for all sorts of things. If we look at even a single industry we can find TONS of what one could call "desktop applications". As an example, let's look at the ubiquitous convenience store:

• Point of Sale equipment, like cash registers
• Inventory management systems, which maintain a local database and front end for entering information
• Security camera systems, which maintain a frontend that allows you to select from many views, and a backup and storing system.
• Financial management software, which helps to generate financial reports
• Employee management and scheduling software.

You do know much of them are free right? 

[wasab]

2 hours ago, Dat Guy said:

I do.

Why would you when they are free?

[Dat Guy]

Because good software is never free.

[wasab]

17 minutes ago, Dat Guy said:

Because good software is never free.

Linux and all majority of the softwares I used are free and open source so you're wrong. 

[straight_stewie]

22 minutes ago, wasab said:

You do know much of them are free right? 

Business process software ready for end user deployment at a serious business is almost never free.

As one example pulled from my list above, Intuits Quickbooks Point of Sale, one of the cheapest and easiest to get POS solutions, costs between $1200 and $1900 USD for a single license, and does not include the machine to run it on.

 

Would you develop seriously robust software that businesses rely on, take on all of the liability if something goes wrong and a business loses money, and then give it away for free, specifically as a tool for other people to use solely to make money? You've got to atleast try to make your arguments make sense...