Pharo

Introduction

Pharo is a pure object-oriented dynamically typed and reflective language. The stated goal of Pharo is to revisit Smalltalk design and enhance it.

The name Pharo comes from the French word "phare" (French pronunciation: ​[faʁ]) which means lighthouse. This is why the Pharo logo shows a drawing of a lighthouse inside the final letter O of the name.

Key features

• Optional fusion of developed program and development environment

• A pure object-oriented approach
• Simple syntax
• Immediate object identity swapping
• Resumable exceptions
• Live object inspection
• Dynamic inheritance
• Multiplatform virtual machine with JIT, combined generational garbage collector, ephemerons, forwarders
• Easy call stack manipulation
• Fast object enumeration
• Objects as methods
• Optional Green threads
• AST metalinks
• Customizable metaclasses
• Relatively low memory consumption
• Customizable compiler
• Optional complete object memory persistence
• Fast object serialization
• Easy use of proxy objects

Relation to Smalltalk

Pharo is based on general concepts of Smalltalk but seeks to improve on them so does not limit itself to them. The basic syntax of the language has a close resemblance to Smalltalk. However, the way classes are defined in Pharo differs from other Smalltalk dialects.^[how?]

Language syntax

Pharo syntax postcard

The Pharo syntax is based on Smalltalk-80 language syntax with several extensions. Some of these are common among modern Smalltalk dialects.

• literals for dynamic arrays. The expressions that specify the array content are evaluated in time of the program execution

{1. 2. 1+2}

• literals for byte arrays that can be composed only of integer numbers in the range from 0 to 255

#[1 2 3 4]

• literals for scaled decimals, a representation of fixed point decimal numbers able to accurately represent decimal fractions

3.14s2

• pragmas. In Smalltalk-80 the pragmas are used only for primitive methods. In Pharo they are fully capable method annotations

<gtInspectorPresentationOrder: 30>

• two double quotes inside a comment are interpreted as a single double quotes character that is part of the content of the comment

The Pharo language syntax is supposed to be very simple and minimalistic. The basic language elements are often presented on a single postcard. The grammar is classified as LL(1).

The language grammar does not specify directly how the code should be stored in files. Pharo uses Tonel as the preferred code serialization format.

History

Pharo emerged as a fork of Squeak, an open-source Smalltalk environment created by the Smalltalk-80 team (Dan Ingalls and Alan Kay). Pharo was created by S. Ducasse [1] and M. Denker in March 2008. It focuses on modern software engineering and development techniques.^{[citation needed]} Pharo is supported by the Pharo consortium (for legal entities) [2] and the Pharo association for physical persons [3].

Use of Pharo

Companies and consultants

Some companies use Pharo for their development projects.^[12] In particular, they use:

• Seaside for dynamic web development^[13]
• Zinc for server architectures^[14]
• Moose^[15] to analyse data and software from all programming languages
• Graphic libraries for evolved user interfaces
• Roassal to visualize data^[16]

The Pharo consortium^[17] was created for companies wishing to support the Pharo project. The Pharo association^[18] was recently^[when?] created for users wishing to support the project.

Performance and virtual machine (VM)

Pharo relies on a virtual machine that is written almost entirely in Smalltalk itself. Beginning in 2008, a new virtual machine (Cog) for Squeak, Pharo and Newspeak has been developed that has a level of performance close to the fastest Smalltalk virtual machine.^[19] In 2014/2015 the VM community is working on Spur, a new Memory Manager for Cog that should again increase performance and provide better 64-bit VM support.^[20]

See also

• Amber Smalltalk
• GNU Smalltalk
• Squeak
• VisualWorks

References

External links

• Official website