PLUnG

Note to the Readers

Following text is based on ideological concepts which are put together to define a programming language which supposed to be optimal for programming in the e-commerce domain.

The text itself divided into four chapters,

1. Chapter 01: Introduction

2. Chapter 02: Technical Specification

3. Chapter 03: Discussion

4. Chapter 04: Reference and Annexures

In technical specifications which explain in Chapter 02 is generalized to the eCommerce domain in Chapter 03.

CHAPTER 01: INTRODUCTION

Introduction

Buying or selling, service or goods through the internet, simply can be called as eCommerce. At the end of 2019, the global eCommerce market had reached $3.5 billion of sales and as a share of global retail, it is 14%. Further, the expected growth end of 2020 is 16%. (Meyer, 2020) On the other hand, even if we could rhapsodize the growth of the eCommerce today as a whole, actual magic happens in underline technologies.

Today’s market there are several numbers of programming languages are available for eCommerce domain. Most of them were general-purpose languages and optimized for eCommerce domain. Some of them were a collection of related technology to address the requirement of eCommerce domain (for an example MEAN stack).

In following text, it will be discussed about an ideological programming language which is specially focus on eCommerce domain. Apart from defined features of a programming language which will be discovered later in this text, following high level features were took in to account,

• Security
• Responsiveness (interaction handling and interrupts)
• Memory management
• Interoperability
• Modularity
• Community and support

The language here in the following discussion should address the common vulnerabilities related programming language. Since this language ought to specialize in eCommerce domain responsiveness for events and interrupts is vital. Further, the language should have the capabilities to stack with other related platforms/technologies and provide end to end support. In other words, language should well-supportive for full-stack development. Language should have minimum dependencies to another relevant component such as libraries. Moreover, it is one of the major concerns of the founders of this language to let the opensource community to adopt the language.

Let’s introduce the programming language as PLUnG, “Primary Language of Unified Generation”. It is clear that someone can argue that PLUnG is a high-level abstraction of tradition PERL programming language because PLUnG is influenced by PERL programming and programming structures and syntax may have similarities. Rather than introducing an entirely new layout is considered as negativity to simplicity and easy to learn factors. As a matter of fact, PLUnG offers PRUnG^[1] (Primary Repository of Unified Generation) as their native repository while making CPAN (Comprehensive Perl Archive Network) which is PERL’s native library as a secondary repository. So, in this way, PLUnG refers a wide range repository.

Interaction handling is one of the main specialties of PLUnG. In the eCommerce domain, responsiveness to the interactions is the key to maintain precious relationships with the customers.

In following sections of texts, it will be discussed the technical aspect of PLUnG.

Since it’s need to be more significant with the differences regards to PERL, some of the explanation in the text may refer the PERL’s features.

 

CHAPTER 02: TECHNICAL SPECIFICATION

Compilers

There were major concerns regards to compilers, whether PLUnG should be a Compiled language or Interpreted language or like is patriarch language should it be a Hybrid language. Following factors were considered,

• Portability
• Maintainability
• Execution Efficiency and,
• Security

Compiled languages are less portable, since for every new deployment on new environment, it needs to be compiled. Compared to that, in interpreted context the source code may run directly on the machine which is more portable. In spite of that, Hybrid language user can use the byte code to run on their native Virtual Machine.

In maintainability aspect, compiled language can be considered as statically typed language and most of the errors are identified in compile time. Despite that, interpreted languages are opposite.

Both hybrid and compiled language source code is converted in to low level binary or bytecode. Therefore, unlike the interpreted program compiled code can be stored in secured manner.

Since interpreted languages reads each line of the source code prior to execution there is a significant delay on the execution compared compiled languages.

As discussed earlier in the text, Security and Efficiency is one of the major concerns of PLUnG because the domain which PLUnG address is eCommerce domain. Following the table, the summaries the general metrics. (Ampomah , Ezekiel , & Abilimi , 2017)

	Compiled	Interpreted	Hybrid
Portability	Low	High	Average
Maintainability	HIGH	Low	Average
Execution Efficiency	High	Low	Average
Security	High	Low	High

Table 1: Summary of Compiler metrics

More importantly, as PERL^[2] (to be more specific PERL 6+) PLUnG convert the source code to Parrot byte code (Torkington, n.d.). Rather than introducing a new Virtual Machine to run the byte code, the main reason to use Parrot byte code is to avoid the neediness of multiple Virtual Machines. In essence one VM for the entire stack. For instance, as mentioned in the Introduction section, one of the main specialties of PLUnG is to support full-stack development. In that case, PLUnG’s program ought to be integrated with other programs, since then these languages^[3] can share the same VM to convert their bytecodes to machine code. (Dev, 2001)

It is worth note that followed by mentioned criteria on Introduction section Hybrid compiler was the obvious choice.

PLUnG uses Flex as the Lexer which is considered as the successor to Lex and Bison which is considered as the successor to Yaac is used as the parser.

Data types

As discussed earlier in the text, PLUnG required to keep their source code simple as possible while maintaining the ambiguity at their minimum. PLUnG is a loosely typed language but provides fewer data types to programmers to declare. Further, adequate flexibility is provided to programmers to design their own types.

PLUnG associated with five major data types,

• Scalars
• Simple Arrays (Arrays of scalars)
• Maps
• Interaction Objects
• Abstract types (User-defined)
• Boolean

To be more specific, programmers are not required to explicitly specify the type of data, PLUnG itself discover them during the PCUnG^[4] process.

Numerical Data

When it’s come to the numerical representation, to be considered as an integer value programmer may not use any decimal places. When there is a decimal point present it will be initially considered as afloat by the compiler. If the value should not be represented using a float type compiler will automatically switch the type to double.

Floating point and double values are represented as a combination of an exponent and a fraction.

Strings

In order scalar to be a String, programmer can declare the Scalar value inside double quotation. Strings in PLUnG can comprised ACII or UNICODE and escape sequence. In order to provide the simplicity and reliability to the program, when a string assignment operator call for already defined Scalar,

1. It will create a new string Scalar and
2. Memory address will alter to use new string Scalar.

Arrays

PLUnG arrays can only contain same type of element. Since programmer does not explicitly defined the data type for numerical data, for numerical arrays PLUnG will induce a coercion process where all data types are converted to one particular type. For an instance, if a programmer trying to push a decimal element to an array which contain only integer elements, at the compile time all the elements in this array will be represent in double form.

Arrays in PLUnG can be,

1. a Static array which memory location and size are fixed prior to the execution

Figure 4: Static Array

1. a Fixed Stack Dynamic array where array size is fixed but memory allocation is taken place on the stack when specific subroutine is called

Figure 5: Fixed Stack-Dynamic Array

1. a Stack Dynamic array where neither size and stack memory location are not fixed (until particular subroutine call).

Figure 6: Stack Dynamic Array

1. a Fixed Heap-Dynamic Array where array size is fixed before the execution and memory is dynamically allocate from the heap when special subroutine called

Figure 7: Fixed Heap Dynamic Array

1. a heap dynamic array where array is not fixed before the execution and memory is dynamically allocate from the heap when special subroutine called

Figure 8: Heap Dynamic Array

Maps

Maps are allowing programmers to associate key value pairs to their program. Implementation of Maps in PLUnG is similar to most generic implementations of Hash Maps.

Behind the abstraction, there is contiguous array which maps to the Map and element in this array are singly linked lists. When programmer required to add an entry to the Map, key is hashed and linked list entry pushed to the array.

It is important note that, values in Maps can be stored in any valid type but keys should be always uniquely defined using either numerical or string values.

Figure 9: Map

Boolean

Programmers may explicitly ($^) or implicitly defined the Boolean variable in PLUnG. Even both the definitions are valid, it is encouraged to use explicit definition to be more robust and acquire more readability to the source code.

Figure 10: Boolean Data type definition

Interaction Object

Interaction Object (__obj(type)) is a special data type which allows programmer to manipulate interaction of their eCommerce platforms. Programmer can install the interaction type using PRUnG¹ libraries and introduce them their program. This bring simplicity for the programmers by avoiding the neediness of the bulk coding to read and manipulate interaction’s meta data.

For an example, assume that when it receives a Facebook interaction, the interaction generally contain sender, timestamp, subject and etc., programmer can read this data to Interaction Object alter desired data or add new field to inject more information. Interaction Object can be easily convertible to Jason object or specify a REST API call or even convert it to a webhook to pass data to third party application. More importantly inside an Interaction Object there can be hundreds of fields predefined depend on the type⁷ which programmers can manipulate^[5].

Table 2: Genesys’s Facebook interaction meta data (Genesys, 2019)

Figure 11: Interaction creation using PERL and PRUnG native libraries

Note that, with the shebang statement programmers need to specify the compiler version to use. In case of an empty argument it will be used latest and stable release of the PCUnG compiler. Secondly, to import the libraries, programmers may specify either using import keywork or relative path to PRUnG configuration directory (If the library is outside of the PRUnG configuration directory it needs to specify using import keywork). If programmer intended use PERL libraries/binaries it is required to explicitly specify^[6].

Data type	Description	Example
Scalars	A single data unit. Can be number, character or string	$count = 87; $city=”Adelaide”;
Simple Arrays	List of Scalars or Objects	@list = [“foo”, “bar”]
Maps	Set of key value pairs	%books= [“Key1” => “value1”, “Key2” => “Value2”]
Boolean	Boolean values	$^keepalive=True
Interaction Objects	Specially introduced to store the interaction[7] details	__obj(type[8])= [“Attribute1”:”data1”, “Attribute2”: True]

Table : Summary of Data Types

Pointers and Aliasing

To provide indirect addressing and to facilitate programmers to manage dynamic allocation of the memory, PLUnG uses pointers and references. Other than many other languages, in order to aliasing to be safe and to maintain compiler optimization, PLUnG uses a slightly different approach.

As it’s the fact, pointers supposed to point a particular memory location. In PLUnG when a programmer declares a pointer, they may use additional control operators called qualifier. Qualifiers enforce memory safety, compiler optimization and readability of the pointers. (Hicks, Morrisett, Grossman, & Jim, 2004)

Type of qualifiers,

Qualifier	Description
@nullable Default	The pointer itself cannot be null
@notnull	Pointer itself can be null
@aliasable Default	For the same memory location there can be multiple pointers
@notaliasable	For the same memory location there cannot be multiple pointers
@killonchange	If the value of pointing memory location is altered without referring the pointer (ex. By another pointer) pointer itself will be nulled. This qualifier Cannot use with @notnull qualifier
@killontype	If the type of the actual memory location is altered at runtime, pointer itself will be nulled. This qualifier Cannot use with @notnull qualifier
@uncahnge	Pointer cannot be altered during the execution of the program
@taint	Garbage collector will not act on the memory location pointed This option ought to be use with care since this could leads to memory leakage

Table 4: PLUnG Pointer Qualifiers

In PLUnG pointer is can be declared by adding “::ptr” keyword to the scalar and “::addr” is used as the address referencing operator.

Table 5: Pointers syntaxes

Note that, if programmer does not defined any qualifier explicitly, PLUnG compiler will refer the default qualifiers mentioned in Table 4: PLUnG Pointer Qualifiers.

Type Checking System

PLUnG is Type Safe language. Both Static and Dynamic type checking mechanisms are available to PLUnG programmers. In essence, Static type checking which is carried out at compile time is mandatory while Dynamic type checking which is supposed to carry out in runtime is optional. For an instance, in an event of Down casting^[9] Dynamic type check is preferred unless it has not addressed in source code level. Note that as explained in Numerical Data section, PCUnG the compiler of the PLUnG will try to coerce the numerical data to suitable type.

By default, dynamic type checking may enable with the installation of the compiler. Programmers may disable the runtime type checking which is not recommended, by using toggle the environment variable “PCUNG_RTC_FLAG” or by explicitly defining it in Shebang section of the main program.

Figure 12: Runtime Type Check disabling using environment variables

Figure 13: Runtime Type Check disabling in source code level

Exception Handling

Generally, exceptions can be defined as events which are supposed to introduce during the execution of a program if the execution flow of the program deviate from the expected path. (Shankar, 2002) PLUnG programmers can observe following information related to an exception,

• Exception type
• Stack trace of the exception (Where the exception occurs)
• Error and State information of the exception

In order to provide robust compile-time and runtime error handling mechanism, PLUnG has introduced additional subclasses under their throwable main class.

Figure 14 Hierarchy of Exception Classes

In this text, it will be more focusing towards to the exceptions which are specific to eCommerce applications rather than ordinary exception.

According to Oracle definition, exceptions can be categorized in to three categories,

1. Checked Exception
2. Unchecked Exception
3. Error

Checked exceptions are supposed to check at compile-time prior to run time while unchecked exceptions are supposed to checked at run time. Despite that, errors are irrecoverable circumstances where more likely to out of the control from programmer’s perspective.

Runtime Exceptions in PLUnG categorize in to 2 sub classes. Common exception are basic runtime exceptions which are useful with general^[10] programing. For an instance, if user trying to divide a number by zero (0), it will trigger an ArithmeticException.

Real Time Controlled Exception are useful for the situation where a program manipulates real time interaction. Further, these exceptions are associate with interaction objects which have discussed earlier in this text. For an instance, assume that a program has written to handle VoIP^[11] interaction,

1. SocketException exception can be triggered by,
   1. Underline port is unreachable (server or client side)
   2. If it’s using a sock^[12] file to communicate, permission issues, SELinux issues or even file unavailability could cause this
2. ChannelException exception can be triggered due to various issues on the communication channel. For an example,
   1. Under line codec is not compatible
   2. If the program is supposed to convert one codec^[13] to another, codec conversion errors
   3. If signaling of the channel is unreadable due to encryption

Every time when a programmer refers/imports an interaction library, their related exception classes may attach to exception class tree. And these exceptions may or may not be runtime/unchecked exception. Considering above same VoIP example, assume that programmer user Asterisk^[14] interaction libraries. Set of exception which are related to Asterisk’s interaction will be added to exception class tree. In that case programmer may follow the manual from vendor, in this case “Asterisk Manual” to configure exceptions.

Following key words are to be used with exception handler,

Keyword
try	This keyword is associate with except key word. The coding block which specify under this keyword should contain exception code
except	Followed by try block, this is where exception is handled
fin	Purpose of the keyword fin is to execute its code segment without considering the execution of except segment
throw	Used to throw an exception
throws	This keyword is used to declare an exception

Table 6: Keywords with Exception Handler

Following code segment is used to alter the Time Stamp of calls. The program will catch DataExceptions where could be possibly triggers when it violates the allowed date time format. If an exception triggers, it will pass a waring log to syslog service and execute the fin block by logging interaction status with unchanged time stamp to syslog service.

If exception is never triggered, time stamp will be altered and by executing fin block, it will be logged an entry on syslog with altered time stamp and interaction status.

Figure 15: Exception Handling in PLUnG

Memory Management

Garbage Collection

Programmers are not required to explicitly clean the dirty memory location since PLUnG uses Virtual Machine’s (Parrot) garbage collection mechanisms. (Parrot Foundation, 2011)

As a default garbage collecting mechanism^[15], Mark and Sweep (MS) is used by PCRUnG. In MS,

Stage 1: Garbage Collector traces all memory object and test the health^[16] of these memory object. If a memory object is unhealthy it will be marked as dead.

Stage 2: Memory objects which are marked as dead will be Swept and reclaim.

Additionally, for programmers to allow more control over memory management, it is possible to explicitly defined the garbage collector mechanism in the program. Programmers may use the main program’s shebang line to defined the Garbage Collection mechanism. Note that, this is may affect to behavior of the running VM.

Figure 16: Explicitly define the Garbage Collector

Scope of the Variables

Visibility of the variable in PLUnG can be highly manipulatable. Basically, PLUnG programmers may declare their variable either private of global manner. Programmer are not required to explicitly required to declare private variable by default all the declarations are private. Special scenarios like use same name for private and global scoped variable, programmer may explicitly define the private variable.

Figure 17: Private and Global variable

Further, PLUnG as most of the modern programming languages support only Static/Lexical scoping.

Figure : PLUnG Static Scoping example

In above code segment, value which returns by function_01 is independent from its calling function (function_02) and function_01 is always refer the value of the global variable. Importantly, in these type of scenarios, same variable ($varX) is used in both global and private scopes. It is recommended to explicitly declare the scope in such cases. For an example, in order to alter the value of the global variable it may use our::$varX or $varX.

Data Abstraction

PLUnG supports Object Oriented Programming (OOP) by providing interface and abstract classes and related features to the programmer. The declaration of these classes may slightly different than other OOP languages.

Figure 19: PLUnG classes

Objects in PLUnG can be created as an Interaction Object or ordinary object. Even though syntaxes are different, there is no significant difference when it comes to actual object.

Figure 20: Polymorphism and Object creation

CHAPTER 03: DISCUSSION

Discussion

This section is dedicated discuss above mentioned features of the programming language in terms of eCommerce domain. In essence, previous chapter of the text describe the features of the programing language in technical aspect but in this section those concepts relate to eCommerce domain and try to evaluate those features against various matrices.

Evaluation against Readability metrices

Readability in Simplicity aspect

PLUnG itself is a loosely typed language and programmers do not need to explicitly define type of the variable. Even though this can be argued, since the exact primitive type is not explicitly defined in the source code the “it may cause some readability issue”, it is worth to mention that the ultimate goal is here to lift the burden from the programmers. For an instance, if a programmer defined a numerical variable, underline primitive type of the variable is to be cast in the compile time or during the runtime. Further, when programmers are required to deal with object, they may use the syntax which are native to interaction objects or ordinary object (refer: Figure 20: Polymorphism and Object creation). Further, with aliasing, it has been introduced an advance concept of qualifiers which is enhanced the distinct readability of the code.

Readability in Orthogonality aspect

Finest example here regards to Orthogonality in readability perspective is, PLUnG static scoping. According to Figure 18: PLUnG Static Scoping example, programmers are able to draw a fine line what variable should be accessible locally and how global variable are to be used.

Evaluation against Writability metrices

Writability in Simplicity aspect

As mentioned earlier sections, PLUnG provide adequate number of datatypes to the programmers. For an example, programmers do not need to defined primitive type of numerical scalars in PLUnG, it will be constructed by PLUnG itself.

On the other hand, PLUnG supports data abstraction, programmers may define their own types to use in the system. Despite that, regards to interaction objects, programmers can directly refer underline type and create Interaction Objects without having the burden of implementing it.

Further, rather than native repository, PLUnG support cross library support with PERL native repository which was introduced as CPAN, which provides programmers to wide range of support in terms of libraries.

As mention in Introduction sections, this whole PLUnG project ought to be adopt under open source community, due to that the syntax styles using in the PLUnG more likely to show a relativeness to LINUX and PERL style which opensource community may easily familiarize with.

Writability in Orthogonality aspect

In order to maintain the orthogonality matrices regards to working with libraries (native and CPAN), programmers are required to use different syntaxes. Which enhanced both writability and readability aspects (refer Figure 11: Interaction creation using PERL and PRUnG native libraries).

Another example is the usage of Interaction Object where programmers are able to separate ordinary object from Interaction Object.

Evaluation against Reliability metrices

PLUnG provide both static and dynamic type checking. Further, dynamic type checking is optional but recommended while static type checking is mandatory.

Exception Handling in PLUnG has been optimized by introducing additional exception classes which are native to real time communication which ought to be more useful in the eCommerce domain. (refer: Exception Handling)

Last but not least, with aliasing, PLUnG has introduced a concept where programmers can use qualifiers to avoid memory leakages while optimizing their code.

Ideality on ECommerce domain

Sole purpose of this language to provide maximum support to the eCommerce domain. Following discussion based on the feature and their suitability as a programming language which dedicated to eCommerce domain. The discussion is divided to separate sub section to be more subjective.

Data Types

In order to provide more simplicity, PLUnG is introduced as a loosely typed language. There is a tradeoff between simplicity and reliability in this context but the reliability has been achieved through mapping variable to most suitable primitive type during both compile and execution time.

In this context, the main feature that PLUnG provide in terms of Data Types is Interaction Objects. Programmers are not required create a series of classes to handle interactions. Associate the related Interaction library and just use the Interaction Object in the program. (Refer: Interaction Object)

Exception Handling

One of the major hurdles in eCommerce domain regards to programming is handling Realtime data from various channels. As described in Exception Handling section, it has introduced additional exception classes to the language. The events related to these exceptions are considered as common in eCommerce applications. Further, programmer defined exception and exceptions from third party plugins can easily pluggable to throwable inheritance tree.

Compiler

In Compilers section it has been explained well enough why PLUnG uses a hybrid compiler which was introduced as PCUnG.

Since eCommerce solutions are frequently deal with Realtime interactions the speed of the execution was considered and also reusability or portability of the source code is considered. The other, concern was the security of the source code. Security indeed ought to be considered severely, it is assumed that there are many other factors ought to be taken in to account, rather than code interpretation to secure the source code. Either way, as mentioned in the text, the security feature can be considered as a bonus feature.

Further, today’s programming is not bounded to one single programming language or technology but they are stacked with other related technologies. In this case, the language here discussed also considered to be used similarly. That is one of the reasons where the selection of an existing Virtual Machine preferred over developing an entirely new one.

CHAPTER 04: ANNEXURES AND REFERENCES

REFERENCES

Ampomah , E. K., Ezekiel , M. M., & Abilimi , G. C. (2017). Qualitative Assessment of Compiled, Interpreted and Hybrid Programming Languages. New York: Foundation of Computer Science FCS, New York, USA.

Dev, P. T. (2001). introductions. Retrieved from docs.parrot.org: http://docs.parrot.org/parrot/latest/html/docs/book/draft/ch01_introduction.pod.html

Genesys. (2019, 09 06). Metadata. Retrieved from Genesys docs: https://docs.genesys.com/Documentation/GWC/latest/WidgetsAPI/WebChatMetadata

Hicks, M., Morrisett, G., Grossman, D., & Jim, T. (2004). Experience With Safe Manual Memory-Management in Cyclone. Vancouver: ACM. Retrieved from Cyclone.

Meyer, S. (2020). Evolving ECommerce: 14 Trends Driving Online Retail In 2020. Retrieved from bigcommerce: https://www.bigcommerce.com/blog/eCommerce-trends/#weighing-your-options-deciding-if-a-trend-makes-sense-for-your-business

Parrot Foundation. (2011). PDD 9: Garbage Collection Subsystem. Retrieved from Parrot Design Documents: http://parrot.github.io/parrot-docs3/3.5.0/html/docs/pdds/pdd09_gc.pod.html

Shankar, A. U. (2002, 11 14). Exception Handling in Perl. Retrieved from Perl: https://www.perl.com/pub/2002/11/14/exception.html/

Torkington, N. (n.d.). perlcompile. Retrieved from https://perldoc.perl.org/: https://perldoc.perl.org/5.8.9/perlcompile.html

ANNEX 01: GLOSSARY

Asterisk	An opensource IP PBX system/server which use to handle communication (calls)
Bison	A general-purpose passer generator
CPAN	Comprehensive Perl Archive Network
Flex	A lexical analyzer
Genesys	A proprietary IP PBX system/server which use to handle communication (calls)
Interaction	Some form of communication which involve two or more parties
MEAN Stack	MongoDB, Express.js, AngularJS, and Node.js. MEAN is an end-to-end JavaScript stack
Parrot VM	A process virtual machine to run dynamic languages
PBX	Standards for Private Branch Exchange where calls are handled in an organization
PCUnG	The native compiler of PLUnG
PERL	A high-level programming language
PLUnG	Primary Language of Unified Generation
PRUnG	Primary Repository/registry of PLUnG
SHEBANG Line	Use to let operative systems to know which interpreter to use. May used differently depending on the context
Virtual Machine (VM)	A process which has different namespace to host other processes
VoIP	Voice Over IP. This is how voice communication (calls) happens through the IP network

ANNEX 02: FIGURES

Figure 1: Execution Life Cycle 6

Figure 2: Memory representation of Floating 7

Figure 3: Memory representation of Double 7

Figure 4: Static Array 8

Figure 5: Fixed Stack-Dynamic Array 8

Figure 6: Stack Dynamic Array 8

Figure 7: Fixed Heap Dynamic Array 9

Figure 8: Heap Dynamic Array 9

Figure 9: Map 9

Figure 10: Boolean Data type definition 10

Figure 11: Interaction creation using PERL and PRUnG native libraries 11

Figure 12: Runtime Type Check disabling using environment variables 14

Figure 13: Runtime Type Check disabling in source code level 14

Figure 14 Hierarchy of Exception Classes 15

Figure 15: Exception Handling in PLUnG 17

Figure 16: Explicitly define the Garbage Collector 18

Figure 17: Private and Global variable 19

Figure 18: PLUnG Static Scoping example 19

Figure 19: PLUnG classes 20

Figure 20: Polymorphism and Object creation 20

ANNEX 03: TABLES

Table 1: Summary of Compiler metrics 5

Table 2: Genesys’s Facebook interaction meta data (Genesys, 2019) 10

Table 3: Summary of Data Types 11

Table 4: PLUnG Pointer Qualifiers 12

Table 5: Pointers syntaxes 12

Table 6: Keywords with Exception Handler 17

1. Native Repository of PLUnG ↑

2. PERL 5 and older versions are considered as interpreted languages ↑

3. Parrot’s hostable languages: Tcl, JavaScript, Ruby, Lua, Scheme, PHP, Python, Perl 6, APL, and a .NET ↑

4. PCUnG: Native compiler of PLUnG ↑

5. Predefined attribute can be manipulated but cannot be deleted ↑

6. Refer Line 5 of Figure 11: Interaction creation using PERL and PRUnG native libraries ↑

7. Interaction: A call, Facebook message, WhatsApp message etc. ↑

8. Interaction type. Can be Contains: Counters, Elapsed times, Timestamps, and Boolean state flags ↑

9. Downcasting: casting downwards to a subtype along the inheritance tree ↑

10. In the context general programing referred to programming without associating real time user interactions ↑

11. Voice Over IP ↑

12. These are socket files where the endpoints of communication pipes ↑

13. Codec define, encoding or decoding characteristic of a media stream ↑

14. An opensource IP base communication system (PBX) ↑

15. Parrot VM’s Garbage collection mechanisms: MC, Tri-color MC, Copying collection and etc. ↑

16. Health: Whether a particular memory object is used by the program ↑

Note to the Readers

Following text is based on ideological concepts which are put together to define a programming language which supposed to be optimal for programming in the e-commerce domain.

The text itself divided into four chapters,

1. Chapter 01: Introduction
2. Chapter 02: Technical Specification
3. Chapter 03: Discussion
4. Chapter 04: Reference and Annexures

In technical specifications which explain in Chapter 02 is generalized to the eCommerce domain in Chapter 03.

Contents

Note to the Readers i

CHAPTER 01: INTRODUCTION 1

Introduction 2

CHAPTER 02: TECHNICAL SPECIFICATION 4

Compilers 5

Data types 7

Numerical Data 7

Strings 8

Arrays 8

Maps 9

Boolean 9

Interaction Object 10

Pointers and Aliasing 12

Type Checking System 14

Exception Handling 15

Memory Management 18

Garbage Collection 18

Scope of the Variables 19

Data Abstraction 20

CHAPTER 03: DISCUSSION 21

Discussion 22

Evaluation against Readability metrices 22

Evaluation against Writability metrices 22

Evaluation against Reliability metrices 23

Ideality on ECommerce domain 23

CHAPTER 04: ANNEXURES AND REFERENCES 25

REFERENCES 26

ANNEX 01: GLOSSARY 27

ANNEX 02: FIGURES 28

ANNEX 03: TABLES 29

CHAPTER 01: INTRODUCTION

Introduction

Buying or selling, service or goods through the internet, simply can be called as eCommerce. At the end of 2019, the global eCommerce market had reached $3.5 billion of sales and as a share of global retail, it is 14%. Further, the expected growth end of 2020 is 16%. (Meyer, 2020) On the other hand, even if we could rhapsodize the growth of the eCommerce today as a whole, actual magic happens in underline technologies.

Today’s market there are several numbers of programming languages are available for eCommerce domain. Most of them were general-purpose languages and optimized for eCommerce domain. Some of them were a collection of related technology to address the requirement of eCommerce domain (for an example MEAN stack).

In following text, it will be discussed about an ideological programming language which is specially focus on eCommerce domain. Apart from defined features of a programming language which will be discovered later in this text, following high level features were took in to account,

• Security
• Responsiveness (interaction handling and interrupts)
• Memory management
• Interoperability
• Modularity
• Community and support

The language here in following discussion, should address the common vulnerabilities related programming language. Since this language ought to be specialize in eCommerce domain responsiveness for events and interrupts is vital. Further, the language should have the capabilities to stack with other related platforms/technologies and provide end to end support. In other words, language should well supportive for full stack development. Language should have minimum dependencies to other relevant component such as libraries. Moreover, it is one of the major concerns of the founders of this language to let opensource community to adapt the language.

Let’s introduce the programming language as PLUnG, “Primary Language of Unified Generation”. It is clear that someone can argue that PLUnG is a high-level abstraction of tradition PERL programming language because PLUnG is influenced by PERL programing and programing structures and syntax may have similarities. Rather than introducing an entirely new layout is considered as a negativity to simplicity and easy to learn factors. As a matter of fact, PLUnG offers PRUnG^[1] (Primary Repository of Unified Generation) as their native repository while making CPAN (Comprehensive Perl Archive Network) which is PERL’s native library as a secondary repository. So, in this way PLUnG refers wide range repository.

Interaction handling is one of the main specialties of PLUnG. In the eCommerce domain, responsiveness to the interactions is the key to maintain precious relationships with the customers.

In following sections of texts, it will be discussed the technical aspect of PLUnG.

Since it’s need to be more significant with the differences regards to PERL, some of the explanation in the text may refer the PERL’s features.

 

CHAPTER 02: TECHNICAL SPECIFICATION

Compilers

There were major concerns regards to compilers, whether PLUnG should be a Compiled language or Interpreted language or like is patriarch language should it be a Hybrid language. Following factors were considered,

• Portability
• Maintainability
• Execution Efficiency and,
• Security

Compiled languages are less portable, since for every new deployment on new environment, it needs to be compiled. Compared to that, in interpreted context the source code may run directly on the machine which is more portable. In spite of that, Hybrid language user can use the byte code to run on their native Virtual Machine.

In maintainability aspect, compiled language can be considered as statically typed language and most of the errors are identified in compile time. Despite that, interpreted languages are opposite.

Both hybrid and compiled language source code is converted in to low level binary or bytecode. Therefore, unlike the interpreted program compiled code can be stored in secured manner.

Since interpreted languages reads each line of the source code prior to execution there is a significant delay on the execution compared compiled languages.

As discussed in earlier in the text, Security and Efficiency is one of the major concerns of PLUnG because the domain which PLUnG address is eCommerce domain. Following table, the summaries the general metrics. (Ampomah , Ezekiel , & Abilimi , 2017)

	Compiled	Interpreted	Hybrid
Portability	Low	High	Average
Maintainability	HIGH	Low	Average
Execution Efficiency	High	Low	Average
Security	High	Low	High

Table 1: Summary of Compiler metrics

More importantly, as PERL^[2] (to be more specific PERL 6+) PLUnG convert the source code to Parrot byte code (Torkington, n.d.). Rather than introducing a new Virtual Machine to run the byte code, the main reason to use Parrot byte code is to avoid the neediness of multiple Virtual Machines. In essence one VM for entire stack. For an instance, as mentioned in Introduction section, one of the main specialties of PLUnG is to support full-stack development. In that case PLUnG’s program ought to be integrated with other programs, since then these languages^[3] can share the same VM to convert their bytecodes to machine code. (Dev, 2001)

It is worth note that followed by mentioned criteria on Introduction section Hybrid compiler was the obvious choice.

PLUnG uses Flex as the Lexer which is considered as the successor to Lex and Bison which is considered as the successor to Yaac is used as the parser.

Figure 1: Execution Life Cycle

Data types

As discussed earlier in the text, PLUnG required to keep their source code simple as possible while maintaining the ambiguity at their minimum. PLUnG is a loosely typed language but provides fewer data types to programmers to declare. Further, adequate flexibility is provided to programmers to design their own types.

PLUnG associated with five major data types,

• Scalars
• Simple Arrays (Arrays of scalars)
• Maps
• Interaction Objects
• Abstract types (User defined)
• Boolean

To be more specific, programmers are not required to explicitly specify the type of data, PLUnG itself discover them during the PCUnG^[4] process.

Numerical Data

When it’s come to the numerical representation, to be considered as an integer value programmer may not use any decimal places. When there is a decimal point present it will be initially considered as a float by the compiler. If the value should not be represent using float type compiler will automatically switch the type to double.

Floating point and double values are represented as a combination of an exponent and a fraction.

1

Exponent (8 bit)

Fraction (23 bit)

1

Exponent (11 bit)

Fraction (52 bit)

Figure : Memory representation of Floating

Figure : Memory representation of Double

Strings

In order scalar to be a String, programmer can declare the Scalar value inside double quotation. Strings in PLUnG can comprised ACII or UNICODE and escape sequence. In order to provide the simplicity and reliability to the program, when a string assignment operator call for already defined Scalar,

1. It will create a new string Scalar and
2. Memory address will alter to use new string Scalar.

Arrays

PLUnG arrays can only contain same type of element. Since programmer does not explicitly defined the data type for numerical data, for numerical arrays PLUnG will induce a coercion process where all data types are converted to one particular type. For an instance, if a programmer trying to push a decimal element to an array which contain only integer elements, at the compile time all the elements in this array will be represent in double form.

Arrays in PLUnG can be,

1. a Static array which memory location and size are fixed prior to the execution

Figure 4: Static Array

1. a Fixed Stack Dynamic array where array size is fixed but memory allocation is taken place on the stack when specific subroutine is called

Figure 5: Fixed Stack-Dynamic Array

1. a Stack Dynamic array where neither size and stack memory location are not fixed (until particular subroutine call).

Figure 6: Stack Dynamic Array

1. a Fixed Heap-Dynamic Array where array size is fixed before the execution and memory is dynamically allocate from the heap when special subroutine called

Figure 7: Fixed Heap Dynamic Array

1. a heap dynamic array where array is not fixed before the execution and memory is dynamically allocate from the heap when special subroutine called

Figure 8: Heap Dynamic Array

Maps

Maps are allowing programmers to associate key value pairs to their program. Implementation of Maps in PLUnG is similar to most generic implementations of Hash Maps.

Behind the abstraction, there is contiguous array which maps to the Map and element in this array are singly linked lists. When programmer required to add an entry to the Map, key is hashed and linked list entry pushed to the array.

It is important note that, values in Maps can be stored in any valid type but keys should be always uniquely defined using either numerical or string values.

Figure 9: Map

Boolean

Programmers may explicitly ($^) or implicitly defined the Boolean variable in PLUnG. Even both the definitions are valid, it is encouraged to use explicit definition to be more robust and acquire more readability to the source code.

Figure 10: Boolean Data type definition

Interaction Object

Interaction Object (__obj(type)) is a special data type which allows programmer to manipulate interaction of their eCommerce platforms. Programmer can install the interaction type using PRUnG¹ libraries and introduce them their program. This bring simplicity for the programmers by avoiding the neediness of the bulk coding to read and manipulate interaction’s meta data.

For an example, assume that when it receives a Facebook interaction, the interaction generally contain sender, timestamp, subject and etc., programmer can read this data to Interaction Object alter desired data or add new field to inject more information. Interaction Object can be easily convertible to Jason object or specify a REST API call or even convert it to a webhook to pass data to third party application. More importantly inside an Interaction Object there can be hundreds of fields predefined depend on the type⁷ which programmers can manipulate^[5].

Table 2: Genesys’s Facebook interaction meta data (Genesys, 2019)

Figure 11: Interaction creation using PERL and PRUnG native libraries

Note that, with the shebang statement programmers need to specify the compiler version to use. In case of an empty argument it will be used latest and stable release of the PCUnG compiler. Secondly, to import the libraries, programmers may specify either using import keywork or relative path to PRUnG configuration directory (If the library is outside of the PRUnG configuration directory it needs to specify using import keywork). If programmer intended use PERL libraries/binaries it is required to explicitly specify^[6].

Data type	Description	Example
Scalars	A single data unit. Can be number, character or string	$count = 87; $city=”Adelaide”;
Simple Arrays	List of Scalars or Objects	@list = [“foo”, “bar”]
Maps	Set of key value pairs	%books= [“Key1” => “value1”, “Key2” => “Value2”]
Boolean	Boolean values	$^keepalive=True
Interaction Objects	Specially introduced to store the interaction[7] details	__obj(type[8])= [“Attribute1”:”data1”, “Attribute2”: True]

Table : Summary of Data Types

Pointers and Aliasing

To provide indirect addressing and to facilitate programmers to manage dynamic allocation of the memory, PLUnG uses pointers and references. Other than many other languages, in order to aliasing to be safe and to maintain compiler optimization, PLUnG uses a slightly different approach.

As it’s the fact, pointers supposed to point a particular memory location. In PLUnG when a programmer declares a pointer, they may use additional control operators called qualifier. Qualifiers enforce memory safety, compiler optimization and readability of the pointers. (Hicks, Morrisett, Grossman, & Jim, 2004)

Type of qualifiers,

Qualifier	Description
@nullable Default	The pointer itself cannot be null
@notnull	Pointer itself can be null
@aliasable Default	For the same memory location there can be multiple pointers
@notaliasable	For the same memory location there cannot be multiple pointers
@killonchange	If the value of pointing memory location is altered without referring the pointer (ex. By another pointer) pointer itself will be nulled. This qualifier Cannot use with @notnull qualifier
@killontype	If the type of the actual memory location is altered at runtime, pointer itself will be nulled. This qualifier Cannot use with @notnull qualifier
@uncahnge	Pointer cannot be altered during the execution of the program
@taint	Garbage collector will not act on the memory location pointed This option ought to be use with care since this could leads to memory leakage

Table 4: PLUnG Pointer Qualifiers

In PLUnG pointer is can be declared by adding “::ptr” keyword to the scalar and “::addr” is used as the address referencing operator.

Table 5: Pointers syntaxes

Note that, if programmer does not defined any qualifier explicitly, PLUnG compiler will refer the default qualifiers mentioned in Table 4: PLUnG Pointer Qualifiers.

Type Checking System

PLUnG is Type Safe language. Both Static and Dynamic type checking mechanisms are available to PLUnG programmers. In essence, Static type checking which is carried out at compile time is mandatory while Dynamic type checking which is supposed to carry out in runtime is optional. For an instance, in an event of Down casting^[9] Dynamic type check is preferred unless it has not addressed in source code level. Note that as explained in Numerical Data section, PCUnG the compiler of the PLUnG will try to coerce the numerical data to suitable type.

By default, dynamic type checking may enable with the installation of the compiler. Programmers may disable the runtime type checking which is not recommended, by using toggle the environment variable “PCUNG_RTC_FLAG” or by explicitly defining it in Shebang section of the main program.

Figure 12: Runtime Type Check disabling using environment variables

Figure 13: Runtime Type Check disabling in source code level

Exception Handling

Generally, exceptions can be defined as events which are supposed to introduce during the execution of a program if the execution flow of the program deviate from the expected path. (Shankar, 2002) PLUnG programmers can observe following information related to an exception,

• Exception type
• Stack trace of the exception (Where the exception occurs)
• Error and State information of the exception

In order to provide robust compile time and runtime error handling mechanism PLUnG has introduce additional subclasses under their throwable main class.

Figure 14 Hierarchy of Exception Classes

In this text, it will be more focusing towards to the exceptions which are specific to eCommerce applications rather than ordinary exception.

According to Oracle definition, exceptions can be categorized in to three categories,

1. Checked Exception
2. Unchecked Exception
3. Error

Checked exceptions are supposed to check at compile-time prior to run time while unchecked exceptions are supposed to checked at run time. Despite that, errors are irrecoverable circumstances where more likely to out of the control from programmer’s perspective.

Runtime Exceptions in PLUnG categorize in to 2 sub classes. Common exception are basic runtime exceptions which are useful with general^[10] programing. For an instance, if user trying to divide a number by zero (0), it will trigger an ArithmeticException.

Real Time Controlled Exception are useful for the situation where a program manipulates real time interaction. Further, these exceptions are associate with interaction objects which have discussed earlier in this text. For an instance, assume that a program has written to handle VoIP^[11] interaction,

1. SocketException exception can be triggered by,
   1. Underline port is unreachable (server or client side)
   2. If it’s using a sock^[12] file to communicate, permission issues, SELinux issues or even file unavailability could cause this
2. ChannelException exception can be triggered due to various issues on the communication channel. For an example,
   1. Under line codec is not compatible
   2. If the program is supposed to convert one codec^[13] to another, codec conversion errors
   3. If signaling of the channel is unreadable due to encryption

Every time when a programmer refers/imports an interaction library, their related exception classes may attach to exception class tree. And these exceptions may or may not be runtime/unchecked exception. Considering above same VoIP example, assume that programmer user Asterisk^[14] interaction libraries. Set of exception which are related to Asterisk’s interaction will be added to exception class tree. In that case programmer may follow the manual from vendor, in this case “Asterisk Manual” to configure exceptions.

Following key words are to be used with exception handler,

Keyword
try	This keyword is associate with except key word. The coding block which specify under this keyword should contain exception code
except	Followed by try block, this is where exception is handled
fin	Purpose of the keyword fin is to execute its code segment without considering the execution of except segment
throw	Used to throw an exception
throws	This keyword is used to declare an exception

Table 6: Keywords with Exception Handler

Following code segment is used to alter the Time Stamp of calls. The program will catch DataExceptions where could be possibly triggers when it violates the allowed date time format. If an exception triggers, it will pass a waring log to syslog service and execute the fin block by logging interaction status with unchanged time stamp to syslog service.

If exception is never triggered, time stamp will be altered and by executing fin block, it will be logged an entry on syslog with altered time stamp and interaction status.

Figure 15: Exception Handling in PLUnG

Memory Management

Garbage Collection

Programmers are not required to explicitly clean the dirty memory location since PLUnG uses Virtual Machine’s (Parrot) garbage collection mechanisms. (Parrot Foundation, 2011)

As a default garbage collecting mechanism^[15], Mark and Sweep (MS) is used by PCRUnG. In MS,

Stage 1: Garbage Collector traces all memory object and test the health^[16] of these memory object. If a memory object is unhealthy it will be marked as dead.

Stage 2: Memory objects which are marked as dead will be Swept and reclaim.

Additionally, for programmers to allow more control over memory management, it is possible to explicitly defined the garbage collector mechanism in the program. Programmers may use the main program’s shebang line to defined the Garbage Collection mechanism. Note that, this is may affect to behavior of the running VM.

Figure 16: Explicitly define the Garbage Collector

Scope of the Variables

Visibility of the variable in PLUnG can be highly manipulatable. Basically, PLUnG programmers may declare their variable either private of global manner. Programmer are not required to explicitly required to declare private variable by default all the declarations are private. Special scenarios like use same name for private and global scoped variable, programmer may explicitly define the private variable.

Figure 17: Private and Global variable

Further, PLUnG as most of the modern programming languages support only Static/Lexical scoping.

Figure : PLUnG Static Scoping example

In above code segment, value which returns by function_01 is independent from its calling function (function_02) and function_01 is always refer the value of the global variable. Importantly, in these type of scenarios, same variable ($varX) is used in both global and private scopes. It is recommended to explicitly declare the scope in such cases. For an example, in order to alter the value of the global variable it may use our::$varX or $varX.

Data Abstraction

PLUnG supports Object Oriented Programming (OOP) by providing interface and abstract classes and related features to the programmer. The declaration of these classes may slightly different than other OOP languages.

Figure 19: PLUnG classes

Objects in PLUnG can be created as an Interaction Object or ordinary object. Even though syntaxes are different, there is no significant difference when it comes to actual object.

Figure 20: Polymorphism and Object creation

CHAPTER 03: DISCUSSION

Discussion

This section is dedicated discuss above mentioned features of the programming language in terms of eCommerce domain. In essence, previous chapter of the text describe the features of the programing language in technical aspect but in this section those concepts relate to eCommerce domain and try to evaluate those features against various matrices.

Evaluation against Readability metrices

Readability in Simplicity aspect

PLUnG itself is a loosely typed language and programmers do not need to explicitly define type of the variable. Even though this can be argued, since the exact primitive type is not explicitly defined in the source code the “it may cause some readability issue”, it is worth to mention that the ultimate goal is here to lift the burden from the programmers. For an instance, if a programmer defined a numerical variable, underline primitive type of the variable is to be cast in the compile time or during the runtime. Further, when programmers are required to deal with object, they may use the syntax which are native to interaction objects or ordinary object (refer: Figure 20: Polymorphism and Object creation). Further, with aliasing, it has been introduced an advance concept of qualifiers which is enhanced the distinct readability of the code.

Readability in Orthogonality aspect

Finest example here regards to Orthogonality in readability perspective is, PLUnG static scoping. According to Figure 18: PLUnG Static Scoping example, programmers are able to draw a fine line what variable should be accessible locally and how global variable are to be used.

Evaluation against Writability metrices

Writability in Simplicity aspect

As mentioned earlier sections, PLUnG provide adequate number of datatypes to the programmers. For an example, programmers do not need to defined primitive type of numerical scalars in PLUnG, it will be constructed by PLUnG itself.

On the other hand, PLUnG supports data abstraction, programmers may define their own types to use in the system. Despite that, regards to interaction objects, programmers can directly refer underline type and create Interaction Objects without having the burden of implementing it.

Further, rather than native repository, PLUnG support cross library support with PERL native repository which was introduced as CPAN, which provides programmers to wide range of support in terms of libraries.

As mention in Introduction sections, this whole PLUnG project ought to be adopt under open source community, due to that the syntax styles using in the PLUnG more likely to show a relativeness to LINUX and PERL style which opensource community may easily familiarize with.

Writability in Orthogonality aspect

In order to maintain the orthogonality matrices regards to working with libraries (native and CPAN), programmers are required to use different syntaxes. Which enhanced both writability and readability aspects (refer Figure 11: Interaction creation using PERL and PRUnG native libraries).

Another example is the usage of Interaction Object where programmers are able to separate ordinary object from Interaction Object.

Evaluation against Reliability metrices

PLUnG provide both static and dynamic type checking. Further, dynamic type checking is optional but recommended while static type checking is mandatory.

Exception Handling in PLUnG has been optimized by introducing additional exception classes which are native to real time communication which ought to be more useful in the eCommerce domain. (refer: Exception Handling)

Last but not least, with aliasing, PLUnG has introduced a concept where programmers can use qualifiers to avoid memory leakages while optimizing their code.

Ideality on ECommerce domain

Sole purpose of this language to provide maximum support to the eCommerce domain. Following discussion based on the feature and their suitability as a programming language which dedicated to eCommerce domain. The discussion is divided to separate sub section to be more subjective.

Data Types

In order to provide more simplicity, PLUnG is introduced as a loosely typed language. There is a tradeoff between simplicity and reliability in this context but the reliability has been achieved through mapping variable to most suitable primitive type during both compile and execution time.

In this context, the main feature that PLUnG provide in terms of Data Types is Interaction Objects. Programmers are not required create a series of classes to handle interactions. Associate the related Interaction library and just use the Interaction Object in the program. (Refer: Interaction Object)

Exception Handling

One of the major hurdles in eCommerce domain regards to programming is handling Realtime data from various channels. As described in Exception Handling section, it has introduced additional exception classes to the language. The events related to these exceptions are considered as common in eCommerce applications. Further, programmer defined exception and exceptions from third party plugins can easily pluggable to throwable inheritance tree.

Compiler

In Compilers section it has been explained well enough why PLUnG uses a hybrid compiler which was introduced as PCUnG.

Since eCommerce solutions are frequently deal with Realtime interactions the speed of the execution was considered and also reusability or portability of the source code is considered. The other, concern was the security of the source code. Security indeed ought to be considered severely, it is assumed that there are many other factors ought to be taken in to account, rather than code interpretation to secure the source code. Either way, as mentioned in the text, the security feature can be considered as a bonus feature.

Further, today’s programming is not bounded to one single programming language or technology but they are stacked with other related technologies. In this case, the language here discussed also considered to be used similarly. That is one of the reasons where the selection of an existing Virtual Machine preferred over developing an entirely new one.

CHAPTER 04: ANNEXURES AND REFERENCES

REFERENCES

Ampomah , E. K., Ezekiel , M. M., & Abilimi , G. C. (2017). Qualitative Assessment of Compiled, Interpreted and Hybrid Programming Languages. New York: Foundation of Computer Science FCS, New York, USA.

Dev, P. T. (2001). introductions. Retrieved from docs.parrot.org: http://docs.parrot.org/parrot/latest/html/docs/book/draft/ch01_introduction.pod.html

Genesys. (2019, 09 06). Metadata. Retrieved from Genesys docs: https://docs.genesys.com/Documentation/GWC/latest/WidgetsAPI/WebChatMetadata

Hicks, M., Morrisett, G., Grossman, D., & Jim, T. (2004). Experience With Safe Manual Memory-Management in Cyclone. Vancouver: ACM. Retrieved from Cyclone.

Meyer, S. (2020). Evolving ECommerce: 14 Trends Driving Online Retail In 2020. Retrieved from bigcommerce: https://www.bigcommerce.com/blog/eCommerce-trends/#weighing-your-options-deciding-if-a-trend-makes-sense-for-your-business

Parrot Foundation. (2011). PDD 9: Garbage Collection Subsystem. Retrieved from Parrot Design Documents: http://parrot.github.io/parrot-docs3/3.5.0/html/docs/pdds/pdd09_gc.pod.html

Shankar, A. U. (2002, 11 14). Exception Handling in Perl. Retrieved from Perl: https://www.perl.com/pub/2002/11/14/exception.html/

Torkington, N. (n.d.). perlcompile. Retrieved from https://perldoc.perl.org/: https://perldoc.perl.org/5.8.9/perlcompile.html

ANNEX 01: GLOSSARY

Asterisk	An opensource IP PBX system/server which use to handle communication (calls)
Bison	A general-purpose passer generator
CPAN	Comprehensive Perl Archive Network
Flex	A lexical analyzer
Genesys	A proprietary IP PBX system/server which use to handle communication (calls)
Interaction	Some form of communication which involve two or more parties
MEAN Stack	MongoDB, Express.js, AngularJS, and Node.js. MEAN is an end-to-end JavaScript stack
Parrot VM	A process virtual machine to run dynamic languages
PBX	Standards for Private Branch Exchange where calls are handled in an organization
PCUnG	The native compiler of PLUnG
PERL	A high-level programming language
PLUnG	Primary Language of Unified Generation
PRUnG	Primary Repository/registry of PLUnG
SHEBANG Line	Use to let operative systems to know which interpreter to use. May used differently depending on the context
Virtual Machine (VM)	A process which has different namespace to host other processes
VoIP	Voice Over IP. This is how voice communication (calls) happens through the IP network

ANNEX 02: FIGURES

Figure 1: Execution Life Cycle 6

Figure 2: Memory representation of Floating 7

Figure 3: Memory representation of Double 7

Figure 4: Static Array 8

Figure 5: Fixed Stack-Dynamic Array 8

Figure 6: Stack Dynamic Array 8

Figure 7: Fixed Heap Dynamic Array 9

Figure 8: Heap Dynamic Array 9

Figure 9: Map 9

Figure 10: Boolean Data type definition 10

Figure 11: Interaction creation using PERL and PRUnG native libraries 11

Figure 12: Runtime Type Check disabling using environment variables 14

Figure 13: Runtime Type Check disabling in source code level 14

Figure 14 Hierarchy of Exception Classes 15

Figure 15: Exception Handling in PLUnG 17

Figure 16: Explicitly define the Garbage Collector 18

Figure 17: Private and Global variable 19

Figure 18: PLUnG Static Scoping example 19

Figure 19: PLUnG classes 20

Figure 20: Polymorphism and Object creation 20

ANNEX 03: TABLES

Table 1: Summary of Compiler metrics 5

Table 2: Genesys’s Facebook interaction meta data (Genesys, 2019) 10

Table 3: Summary of Data Types 11

Table 4: PLUnG Pointer Qualifiers 12

Table 5: Pointers syntaxes 12

Table 6: Keywords with Exception Handler 17

1. Native Repository of PLUnG ↑

2. PERL 5 and older versions are considered as interpreted languages ↑

3. Parrot’s hostable languages: Tcl, JavaScript, Ruby, Lua, Scheme, PHP, Python, Perl 6, APL, and a .NET ↑

4. PCUnG: Native compiler of PLUnG ↑

5. Predefined attribute can be manipulated but cannot be deleted ↑

6. Refer Line 5 of Figure 11: Interaction creation using PERL and PRUnG native libraries ↑

7. Interaction: A call, Facebook message, WhatsApp message etc. ↑

8. Interaction type. Can be Contains: Counters, Elapsed times, Timestamps, and Boolean state flags ↑

9. Downcasting: casting downwards to a subtype along the inheritance tree ↑

10. In the context general programing referred to programming without associating real time user interactions ↑

11. Voice Over IP ↑

12. These are socket files where the endpoints of communication pipes ↑

13. Codec define, encoding or decoding characteristic of a media stream ↑

14. An opensource IP base communication system (PBX) ↑

15. Parrot VM’s Garbage collection mechanisms: MC, Tri-color MC, Copying collection and etc. ↑

16. Health: Whether a particular memory object is used by the program ↑