source: content/en/posts/web-stacks.org@ 4c9c7af

#+TITLE: Tech stacks for earning money as web developer
#+DATE: <2022-07-27 Ср>
#+LANGUAGE: en

* Tech stacks for earning money as web developer
#+BEGIN_abstract
In this article I'm going to categorize languages, frameworks, libraries
and software by a criteria of ease of starting to earn money. It is
biased and irrational, be aware.
#+END_abstract
I work in IT field as web developer for many years, but cannot claim
myself as professional. I don't like to work in general, I don't like
corporate IT culture and I don't like web development in particular, but
I has an experience in that field and it is much easier to dive in in
order to earn money for a living than other IT fields. I hope this text
will be useful for someone who is outside of field of web development
and wish to start earning money in that field.

* What is tech stack
#+BEGIN_abstract
Tech stack is the mix of technologies (software, frameworks, libraries)
that project uses at your current or future company.
#+END_abstract

Knowing everything from your techstack is not the only key for
success. It can sound obvious, but social skills can play more important
role than your programming skills. Programming is very social activity,
everything you are going to use is built by other people, sometimes
something breaks and you need to contact with people that provide you
software you use, it is also important to communicate with your
company's team and not bring yourself into conflicts.

Next I present the list of technologies you should know to work as web
developer. The list is not mandotary and you are not required to learn
it on expert level to get your job offer.

* Fundamental technologies
Some things you will use in most of your projects, so you are better to
familarize with them anyway. What I've discovered, that most knowledge
of some tech is transfarable to other techs.
** Text Editor
Not sure if I need to write it, but you need development environment
where you will write code, run tests and compile your project.

*Most popular choices*
- [[https://gnu.org/software/emacs/][Emacs]] / [[https://www.vim.org/][Vim]] (I recommend them for a long run)
- Jetbrains IDE (proprietary)
- VSCode (proprietary)  

*Less popular*  
- Sublime (proprietary)
- [[https://www.eclipse.org/][Eclipse]] / [[https://netbeans.apache.org/][Netbeans]] / [[https://www.kdevelop.org/][Kdevelop]]
- other editors and IDEs

Proprietary means that source code is not available and probably it
costs money to use and you as a user don't have control over it. You can
learn more why it is important on [[https://gnu.org/philosophy/][GNU website]]. Such important tool as
text editor should be free for everyone.

[[[https://en.wikipedia.org/wiki/Text_editor][Wikipedia: Text editor]]] |
[[[https://en.wikipedia.org/wiki/Integrated_development_environment][Wikipedia: Integrated development environment]]]

** SQL
Or it should be called persistance, but mostly you will use something
like [[https://www.postgresql.org/][Postgres]] or [[https://mariadb.com/kb/en/][MariaDB]]/[[https://mysql.com/][MySQL]]. Rarely it can be nosql like
mongodb.

The database is the main thing aroud which you will develop your
application. I would say that the code is secondary compared to the data
you store and get from the database, so it is extremely good idea to
learn SQL well especially for a long run. For the start it is okay to
know a little and get going along the way, because you will probably
work with SQL through ORM.

I recommend to use [[https://dbeaver.io/][Dbeaver]] as IDE for running SQL related code. It eases
the complexity of writing and running SQL.

[[[https://en.wikipedia.org/wiki/SQL][Wikipedia: SQL]]] |
[[[https://en.wikipedia.org/wiki/Relational_database][Wikipedia: Reletational Database]]]

*** ORM
Each tech stack provides ORM. It is quite rare, where you need to work
with pure SQL or the project doesn't use ORM at all. ORMs might be
something from the list:
- [[https://hibernate.org/][Hibernate]]
- [[https://docs.djangoproject.com/][Django ORM]]
- [[https://www.doctrine-project.org/][Doctrine]]
- [[https://sqlalchemy.org/][SQLAlchemy]]
- [[https://guides.rubyonrails.org/active_record_basics.html][Rails Active Record]]
- and many others

[[[https://en.wikipedia.org/wiki/SQL][Wikipedia: SQL]]] | [[[https://en.wikipedia.org/wiki/Object%E2%80%93relational_mapping][Wikipedia: ORM]]]

** In-memory database
Some data needs to be stored in RAM for better performance. For that
cases you need to learn about Redis or similar technology.

[[[https://en.wikipedia.org/wiki/In-memory_database][Wikipedia: In-memory database]]]

** HTTP Server and HTTP protocol
You don't actually need to know everything from the start, but since you
are going to work as web dev you need to know what you are using. It is
essential if you do backend APIs.

HTTP server is usually [[https://en.wikipedia.org/wiki/Nginx][Nginx]] or [[https://en.wikipedia.org/wiki/Apache_HTTP_Server][Apache]] (it is used rarely now). If you
work with Java, http servers can be written in Java such as [[https://en.wikipedia.org/wiki/Jetty_(web_server)][Jetty]] and
others.

[[[https://en.wikipedia.org/wiki/Web_server][Wikipedia: Web server]]] |
[[[https://en.wikipedia.org/wiki/Hypertext_Transfer_Protocol][Wikipedia: Hypertext Transfer Protocol]]]

** Fulltext search
You might need searching functionality. In basic cases it is okay to
include google search, but in many cases it is not acceptable choice. To
do a search like google does you will need fulltext search engines. You
can use your database for that or specific solutins like ElasticSearch
for such purpose.

[[[https://en.wikipedia.org/wiki/Full-text_search][Wikipedia: Full-text search]]] |
[[[https://en.wikipedia.org/wiki/Elasticsearch][Wikipedia: Elasticsearch]]] (proprietary)

** Message broker
RabbitMQ, ZeroMQ, Kafka and others. You need those to create queues and
other stuff that needs to be delayed.

[[https://en.wikipedia.org/wiki/Message_broker][Wikipedia: Message Broker]]

** Unix-like OS
You will probably work on Linux, variants of BSD system or macs. You
need to familiarize with the system you use, different tasks require
different degree of that skill, but knowledge of what you are using
every day will be handy.

There are many ways to learn it, but I recommend try to install it on
your machine and learn it along the way. I recommend to start with
[[https://debian.org/][Debian GNU/Linux distribution]]. You can read [[https://www.debian.org/doc/manuals/debian-faq/][FAQ]], manuals about
[[https://www.debian.org/releases/stable/amd64/index.en.html][installation]] and [[https://www.debian.org/doc/manuals/debian-reference/][how to use the system]]. Docs are very useful, they are
will be present in your filesystem you can read it by jumping into
/usr/share/docs folder or running ~info~ and ~man~ commands. I recommend to
read it and explore Debian website. Also for learning linux Arch Linux
wiki is good source of knowledge (Slackware and Gentoo wikis are good as
well).

[[[https://en.wikipedia.org/wiki/Unix][Wikipedia: Unix]]] |
[[[https://en.wikipedia.org/wiki/Linux][Wikipedia: Linux]]] |
[[[https://en.wikipedia.org/wiki/Linux_distribution][Wikipedia: Linux distribution]]]

** OOP
In corporate IT culture you will probably forced to use OOP
techniques. It is not something that really improves development that
much, but it is very common way to orginize codebase and it is kind of
great for working with many people.

Not everyone will degree with my definition, but patterns are needed to
express by your programming language object system, which is impossible
to express otherwise, because of different language limits. It is very
true for very static and not expressive language like PHP or Java and
less true for more expressive languages like Lisp and Haskell.

Besides basic principle of using OOP in your programming language of
choice you are going to need patterns to learn. The benefits of them are
questinable, but it is not bad to know and they are being asked on
interviews very often.

The best way to learn patterns are to use [[https://en.wikipedia.org/wiki/Smalltalk][Smalltalk]] and read the book
[[https://en.wikipedia.org/wiki/Design_Patterns]["Gang of Four"]]. Also there is very useful website about design patterns
with examples in many mainstream languages: https://refactoring.guru/

[[[https://en.wikipedia.org/wiki/Object-oriented_programming][Wikipedia: OOP]]]

** Debugger
You have a chance to work with others people code, so you have a chance
to get confused by a lot of stuff going on. Also it is possible that you
would make a mistake and will not understand what is going on, where is
an error. In that case you will need a debugger. Debugger is usually can
be the part of your text editor environment and it is usually included
in your programming environment alongside your interpreter /
compiler. Some languages don't include a debugger, so you need to
download it seperately.

Debugger allows you to step through your code and see the values of
stack and heap, basically you get an access to everything evaluated by
your interpreter/compiler. It is very common practice to use such a
thing when errors happen and I personally think, that it is great way to
learn how your language of choice works internally.

[[[https://en.wikipedia.org/wiki/Debugger][Wikipedia: Debugger]]]

** Profiler
Profiler is similar to debugger, but it helps you to detect which part
of your programm is running slow, it also helps you detect memory
leaks. Sometimes you or others write functions and does everything as
intended, but it is running slow. To understand which part of your code
is a bottleneck you need to run it in a profiler. Then profiler will
show you an diagram with execution time and memory consumption of your
functions, which is helpful even when everything works fine.

[[[https://en.wikipedia.org/wiki/Profiling_(computer_programming)][Wikipedia: Profiler]]]

** Frontend
If you are not exclusively backend developer, you will probably need to
work with templates. Templates represent the outer view of your website,
which represents backend's inner workings. Frontend part can be done
using template engine or can be done as Single Page Application with
lots of client-side javascript coe.


*** HTML
HTML is being used for representing your website information to your
users. It contains only structural data and very basic formatting. If
you see something that looks like old document, that is probably plain
html used.

[[[https://en.wikipedia.org/wiki/Web_template_system][Wikipedia: Web Template System]]] |
[[[https://en.wikipedia.org/wiki/Single-page_application][Wikipedia: Single-page Application]]]

*** CSS
CSS is being used for creating styling for your html pages. The role of
CSS is to make your html layouts pretty, to make your buttons beatiful
and grid is clear.

[[[https://en.wikipedia.org/wiki/CSS][Wikipedia: CSS]]]

*** Javascript
Javascript is being used for creating interectivity with your HTML
pages. If you see dropdown menu, carousel or a slider, there is probably
javascript involved. Javascript is big and complex programming language,
so you as web developer will be required to know only subset of it
except if your main programming language is Javascript.

[[[https://en.wikipedia.org/wiki/Javascript][Wikipedia: Javascript]]]

*** Template engines
Template engines are needed to transfer your backend data to your
user. It provides the way to generate HTML with embedding variables and
different function for developer's ease. Template engine is usually
built-in in your framework of choice.

[[[https://en.wikipedia.org/wiki/Template_processor][Wikipedia: Template Processor]]]

** Docker
Your team probably will use docker for deployment or development (or
both). Anyway you might want to get your project working in a container
for better accessability by your team. The point of docker as
development tool is to let your team quickly to get the same environment
as you. The point of docker as deployment tool to deploy the same
environment that you developing in, its important to have reproducible
production environment, because catching bugs in 3 different
environments (dev, staging, prod) suck a lot.
** API
There are standard ways to build APIs. If you are going to focus on that
part of application, only backend work, you should learn it.
- RestAPI
- JSONAPI
-     
** Type hierarchy and standard library

** Framework or(and) set of libraries
Every job sticks on their own stack. Very often it sticks to one
specific web framework, that you are required to learn. It can be
something like Django, Symfony, Spring, Ruby on Rails or others.

Frameworks are usually the biggest part of your codebase. The difference
between framework and library: Framework runs your code with your
classes, but libraries are being used by you explicitly and you execute
all the code from them. The concept of reverting the control over the
code to your framework is called [[https://en.wikipedia.org/wiki/Inversion_of_control][Inversion of Control]].
** Programming language

You will need to 
* Most popular programming languages
Basically your choices are quite limited here. More jobs available, more
chances to get that job. there will be different results in different
areas. I think that in most areas you will have a choices between Java,
Python and maybe PHP or Ruby. If you want to end up insane, there is
also Javascript, but it is not that mature for backend as other
technologies, so I will not include it (there is only "big" framework
called NestJS, its demand should be low).

** Java / C#
Probably the most popular programming languages out there especially
when it comes to working at companies. It is statically typed and quite
easy to learn languages. They are restricted to OOP only, so you
prorably will use oop design patterns here the most. Type system are not
that advanced as in ML dialects and actually it is unsound, but it
provides some merits of safety compared to nonsafe C.

Frameworks:
- Spring
- .Net

ORM:
- Hibernate

** Python
Second the most popular programming language. It differs from Java a
lot. It is dynamic typed, more object oriented (everything is an
object), it forces you to indent your code right.

Frameworks:
- Django
- Flask
- FastAPI

ORM:
- Django ORM
- SQLAlchemy

** PHP
PHP might be unpopular in some areas, but it is quite popular in
Russia. It runs lots of small businesses and it is probably the easiest
tech stack to get a job (but not the easiest language to work with). The
language has lots of caveats such as weak typing, unorganized standard
library, less developed OOP and type checking, lots of procedural php
legacy, most functions of the language is defined in C.

Frameworks:
- Symfony
- Laravel

ORM:
- Doctrine

* Less popular programming language
** Go
** Ruby
Ruby is mostly used solely for web development with popular Ruby on
Rails framework. The demand of it might be low in some areas, so
investigate your market yourself before learning it. Ruby is strongly
type object-oriented in smalltalk sense language. It has mixed lisp,
perl and smalltalk roots.

Frameworks:
- Ruby on Rails
- Sinatra

ORM:
- Active Record
- DataMapper  

* Great language, but bad for finding work
* How to get good
First of all don't buy any paid course or workshops, most information
are available for free for everyone. Just read wikipedia links I
provided in artible and follow links there to official documentation of
the thing you want to learn.

When it comes to programming in general I recommend to learn it doing
test projects or just playing with the code. For practice you can also
use services like Exercism or Codewars to practice pure programming
solving problems. Also there is great web-resource for complete
beginners called [[https://freecodecamp.org/][FreeCodeCamp]], it covers basics of web development and
javascript programming.

Also I recommend reading old academic books on programming, because many
concepts are the same in many programming languages, what you need is
the ability think abstractly using data structures and algorhythms. I
recommend anyone who wants to really grock programming to read:
- [[https://en.wikipedia.org/wiki/Structure_and_Interpretation_of_Computer_Programs][SICP]] (mit scheme)
- [[https://en.wikipedia.org/wiki/How_to_Design_Programs][HTDP]] (racket)
- [[https://www.cis.upenn.edu/~bcpierce/tapl/][TAPL]]
- [[https://en.wikipedia.org/wiki/Essentials_of_Programming_Languages][Essentials Of Programming Languages]]
- [[https://download-mirror.savannah.gnu.org/releases/pgubook/ProgrammingGroundUp-1-0-booksize.pdf][Programming from the ground up]]
- Also see [[https://www.gnu.org/doc/other-free-books.html][free books from GNU website]]

/Basically the rule of thumb is that if the book has wikipedia article, it is probably a good book and you should read it/.  
      
There are many free sources to learn you just need to invest your time
into it. The downside of learning LISP is that mainstream programming on
your job can feel very restrictive and less expressive. Also you will
probably learn, that many mainstream languages have lots of quircks or
plainly bad design decisions.

* Conclusion
It is possible to get a job with less popular stack, but if you are just
starting it might be reasonable to focus on popular one first. The list
of technologies is quite big, so newbies can be lost in new
information. Be patient and learn one thing at the time. I would
recommend starting with programming language documentation, then SQL and
then going to docs of your web-framework / library.

Demand for web developers is huge and in my opinion it is the easiest
way to get a job. The web development sucks in general (at least for
me), but that topic will be hold for another article. 

If to continue tech stack tier list sorted by popularity next one should be:

** Advanced tier
- Scala
- Clojure
- Erlang
- Rust

** Rare tier
- Common Lisp / Scheme
- Haskell
- SML / Ocaml
** Rare and legacy tier
- Perl
- Cobol