RaaiVan stands out as my proudest and most significant work. Originally designed as an “Enterprise Knowledge Management” software, its flexibility and extensive features have allowed many of its users to leverage it as a “Document/Content Management System” as well.

Why My Greatest Achievement?

There are several reasons why RaaiVan stands out as my greatest achievement:

• Many enterprise customers purchased its licence and have used it for many years, continuing to rely on it today.
• Its performance has been proven under heavy load, supporting thousands of daily users.
• Its flexibility has allowed customers to use it for a wide range of requirements, even beyond its initial purpose.
• I was the driving force behind it, designing and developing most, if not all, of its features. Customers frequently praised my innovative solutions, often asking, "How on earth did you come up with such an idea?"
• I was involved in or led all aspects of its development and deployment.
• I played a key role in negotiating with customers and understanding their requirements.

Before My Involvement

The initial version of RaaiVan served as a tool for collecting and sharing “Lessons Learnt,” “Employee Skill-sets,” and “Documents.” It included forms for gathering information, enabling users to comment on these forms or send direct messages to one another.

A key feature of the software was its ability to store a comprehensive list of an organization’s assets and associate relevant information with each asset. This functionality allowed management to easily track the extent of documentation available for each project, process, or area of expertise.

My Journey with RaaiVan

I initially joined the team as a facilitator and teacher, helping customers integrate the software into their workflows. However, I soon noticed that the development team struggled to fix issues and implement new features requested by customers. Leveraging my programming skills, I began assisting the team with releases, adding new features, and fixing bugs.

My contributions in development were so impactful that within a few months, Khashayar, the CEO of InKnowTex, dismissed the existing engineering team and offered me a 50-50 partnership. In 2012, I co-founded the new InKnowTex with Khashayar, embarking on a dedicated journey to elevate RaaiVan to new heights.

Full-Stack Software Development

From the very beginning, I was the sole developer of RaaiVan, taking on the role of a full-stack developer from day one. However, I already had six years of non-commercial software development experience with C, C++, and C# from my time in school and university. Additionally, I was part of my university’s ACM programming contest team and participated in the West Asian programming contest twice, in 2006 and 2007.

At the time, it was common to develop the front-end with asp.net. I believed that the three main layers of software—Data, Back-End, and Front-End—should be decoupled. Later, I realised this approach was the MVC pattern, and I had been implementing it without even knowing!

I chose the following tech stack and began migrating all the existing code to it:

1. JavaScript for the front-end, obtaining all necessary data through ajax. Since there wasn't any great front-end framework available then, I developed my own framework and built the entire front-end of the software in plain JavaScript. When I say the whole front-end, I’m talking about more than 150,000 lines of plain JavaScript code. The result was outstanding, and it still works alongside newly developed React components.
2. C# for the back-end, as it was originally. The main change I made was removing all the asp.net front-end code and implementing API handlers instead. Later, I realised this approach was REST API, and I had been doing it for years without realising it.
3. MS SQL Server for data, as it was originally. The main change I implemented here was to use stored procedures with T-SQL instead of the Entity Model. The Entity Model had significant performance issues, especially with complex queries, so using stored procedures greatly improved performance.

Modules and Key Features

Form Builder

The idea of a form builder came to me when we started receiving numerous requests from customers with varying requirements. Each customer needed forms with different inputs.

Instead of creating individual forms for them, I decided to provide them with a form builder so they could create whatever forms they needed.

I developed the first version in less than a week. Initially, it had only four types of fields: Text, Number, Date, and Multiple Choice. Later, I enhanced its user experience and features, adding many more field types such as Binary, RichText, Email, Link, Table, Item Select, Autocomplete, File & Media, User Select, Multi-Level Select (e.g. Country → City), and Divider.

The benefits of this approach were:

• Customers could create as many forms as they needed with as many fields as they wanted.
• All the forms and their instances were indexed by the search engine, allowing users to search across all forms with a single search input.
• Advanced search filters were automatically created based on each form, enabling users to perform more specific searches.

Workflow Engine

Just like the form builder, I decided to develop the workflow engine to meet the diverse requirements of our customers. Another advantage was that it made the software more appealing to potential customers.

My workflow engine functioned as a state machine. The concept was:

• Whenever an item is generated by a user, it circulates through the team or between teams, with its state changing during the process.
• At each state, someone or a specific role is responsible for an action, such as editing a part of a form, uploading a file, answering questions, or making a decision.
• Based on the decisions made, the workflow follows different paths, triggering automatic actions at each decision point, such as publishing the item or calculating a value.
• At each state, people only need to focus on the task assigned to them. The workflow is pre-configured and, based on inputs, it determines the next state.

I implemented the visual representation of the workflow using vis.js, a JavaScript network visualisation library.

Wiki

The Wiki was used as a tool to manage unstructured content, allowing users to write rich-text content, add media or documents, and tag other documents or people.

To ensure the validity of the content, I implemented a simple, optional workflow that could be enabled in the system settings.

With this workflow, whenever a wiki was edited, subject matter experts were notified to review the changes. A single approval from a subject matter expert was sufficient to publish the update.

The key features of the Wiki were:

• The content was broken into multiple blocks, each of which could be approved independently.
• An automatically generated table of contents.
• A content validation workflow.
• Support for multimedia and file uploads.

I also used CKEditor as the Wiki’s WYSIWYG editor.

Search Engine

Search is a crucial element of any software, especially if the software deals with content. When designing the search engine, I considered the following requirements:

• Users should only need a text input to search through the items.
• Since each item could contain multiple sections such as wiki, form, and files, I needed to ensure that users could find an item by searching for something within any of its sections.
• Users should have the option to perform more specific searches if they wish. Therefore, we should provide them with advanced search options.
• As forms were a structured part of the items, we needed to account for the possibility that users might want to filter search results by various form fields. For example, someone might want to find all products within a specific price range.

I implemented the search engine using Apache Lucene and adopted an event-driven approach to keep the index updated. Whenever a section of an item was updated by the user, an event was triggered to update the entry.

I used Tika to extract the text content from uploaded files. However, some files, like text documents, didn’t require any third-party library.

Search based on text content

Advanced search

Other Modules

RaaiVan includes several additional modules:

• Social Network: A tool that enables user interaction on every page of the software, as well as instant messaging, which I implemented using SignalR.
• Core Network: This is RaaiVan’s brain, maintaining all the items and their relationships in a graph data structure. Everything in RaaiVan is an item within this network. I also implemented a visual representation of the network using the JIT library.
• Q&A: A module that helps users get quick and accurate answers to their questions by directing them to subject matter experts.
• Poll: I developed the Poll module on top of the Form Builder. A key feature of this module is its ability to run surveys periodically, monitor the results, and track trends.