Add to Chrome
Add to Edge
Add to Firefox
Add to Opera
Add to Brave
Add to SafariTypeScript vs Pharo
When comparing TypeScript vs Pharo, the Slant community recommends TypeScript for most people. In the question“What is the best programming language to learn first?” TypeScript is ranked 6th while Pharo is ranked 20th. The most important reason people chose TypeScript is:
Typescript has optional static typing with support for interfaces and generics, and intelligent type inference. It makes refactoring large codebases a breeze, and provides many more safeguards for creating stable code.
Specs
Question Ranking
Pros
Pro Optional static typing
Typescript has optional static typing with support for interfaces and generics, and intelligent type inference.
It makes refactoring large codebases a breeze, and provides many more safeguards for creating stable code.
Pro Strong typed language
Lot of benefits of it, you can read this.
Pro Strict superset of Javascript
Every existing Javascript program is already a valid TypeScript program giving it the best support for existing libraries, which is particularly useful if you need to integrate with an existing Javascript code base.
Pro First party Visual Studio support
As a Microsoft developed project, it has first party Visual Studio support that's on par with its C# support with features like syntax sensitive statement completion.
Pro Has a repository of high quality TypeScript type definitions for popular libraries
There are many ready to use and high quality TypeScript definitions for popular libraries including jquery, angular, bootstrap, d3, lodash and many-many more.
Pro Adds support for object-oriented programming
Typescript enables familiar object-oriented programming patterns: classes, inheritance, public/private methods and properties, et cetera.
Pro Polyfill for ES6 fat-arrow syntax
Typescript implements the fat arrow syntax, which always maintains the current context for this and is a shorter/more convenient syntax than traditional function definition.
Pro Great support for React, integrated typed JSX parsing
Strongly typed react components, so UI "templating" automatically gains type safety.
Pro Great support for editors (Sublime, Code, Vim, IntelliJ...)
Pro Works well with existing Javascript code
Both can call Javascript code and be called by Javascript code. Making transitioning to the language very easy.
Pro Compiles to very native looking code
Compiles to simple looking Javascript making it easy to understand what is happening and learn the language (if you already know Javascript).
Pro Built and supported by Microsoft
Being built by Microsoft, TypeScript is much more likely than most other similar open-source projects to receive continued long-term support, good documentation, and a steady stream of development.
Pro Ability to do functional programming
Pro Clear roadmap
TypeScript has a clear and defined roadmap with rapid and constant releases.
Pro Low number of logical errors brought in by built-in type annotations
TypeScript's built-in type signatures allow developers to fully document interfaces and make sure that they will be correctly compiled. Therefore, cutting down on logical errors.
Pro Works well with Angular 2
Angular 2 is built using TypeScript and applications built using it can make use of that (or not).
Pro Object-Oriented
In Pharo everything is an object. Compiler - object, network - object, method - also an object. And objects communicate with messages. No operators, no control-flow statements. Just objects and messages. Few things to learn, but you can learn OOP well.
Pro Easily learnt
There is good, free documentation including several books written by experts with extensive examples. There is an online MOOC. There are many tutorial videos. Supportive conferences and community. Even a professional support option if desired.
Pro Live updates
The nature of Pharo being a "live" environment allows you to perform live updates to your system without requiring to restart it. You can upgrade/modify classes while serving requests at the same time.
Pro Highly productive
Pro Seaside
The framework for developing sophisticated web applications in Smalltalk is developed in Pharo. Seaside lets you build highly interactive web applications quickly, reusably and maintainably.
Pro Remote debugging
Pro Beautiful coding patterns in your IDE
No need to search google for compact beautiful examples of how to do things, your live environment source is available and you can easily live search, see how it works and copy how the masters would do it (examples most languages still copy too).
Pro Glamorous toolkit & GTInspector
Most languages are still copying the Smalltalk tools of yesterday - GTInspector (written in Glamorous) takes live exploration of code/running objects to a new level. It's really slick, and better yet, you can easily write your own inspectors in 10 lines of code.
Pro Code can be run on rock solid GemStone environment
Pro 64 bit support as of Pharo 7
Use 32 bit or 64 bit versions of Pharo on Windows, Mac & Linux.
Pro Advanced code analysis tools
MOOSE environment provides extensive, easily leveraged and class leading tools for code analysis and improvement.
Pro Can run headless for production
Pro Really simple networking and REST with Zinc
Pro Graphics, graphing and visualisation framework - Roassal
Roassal and Mondrian provide fantastic and easily used frameworks for graphics, graphing and advanced visualisations (comparable to D3.js) but with much less code. Visualisations can be rendered into web friendly graphics (SVG, .png etc.) without additional work.
Cons
Con Too similar to Javascript
Presents some advantages compared to Javascript, but because it is designed to be a superset of Javascript, it means all the bad parts of Javascript are still present.
Con Type checking not enforced by default
You have to use compiler flags to make sure it catches flaws like usage of implicit any, etc.
Con Type inference coverage is incomplete
The default type when declaring and using a variable is any. For example, the following should break but does not:
function add(a:number) { return a + 1 }
function addAB(a, b) {return add(a) + b}
addAB("this should break but doesn't :(", 100)In order to avoid this, you have to declare type signatures for every variable or parameter or set the flag --noImplicityAny when running the compiler.
Con Requires "this" for field access
Even in cases were there is no ambiguity, you still have to use "this.fieldName" instead of just "fieldName".
Con Syntax is too verbose
Con No support for dead code elimination
Typescript compiler does not remove dead code from generated file(s), you have to use external tools to remove unused code after compilation. This is harder to achieve, because Typescript compiler eliminated all type information.
Con No support for conditional compilation
There is no clean way to have debug and release builds compiled from the same source, where the release version removes all debugging tools and outputs from the generated file(s).
Con Awful error messages
Comparing to Elm or Rust for example, TypeScript's error messages won't say you very much. For example if you change method of interface which your class implements it won't say your class have incorrect implementation. Instead it'll show error in usage of instances of class. In some cases it can spoil hours of your work trying to figure out why your parameters are incorrect.
Con Technical debt
As consequence of not enforcing type checking.
Con No Java-like package structure
If you prefer a Java-like approach of partitioning your code into different packages, the module system of typescript will confuse you.
Con Small community
Con No option to declare that a function throws errors
Con Small community
But they are very friendly and supportive. Best help comes through the mailing lists so not always easily googlable. There is also a Slack community where help is nearly instantaneous.
Con Odd language
Requires a different mindset. Much harder to apply what you know from popular or conventional languages . Switching over from or between other languages is more difficult.
Con Single threaded
Pharo's VM only ever uses one CPU core. If you want to write code that uses more than one CPU core, you need to jump through hoops such as running multiple VMs and synchronising your data.
