Devlog 5

Today I worked on finishing the last parts of MemoryMap.
I improved the dashboard UI. The dashboard now looks much cleaner, more organised and more complete.

I also worked on:
-better campus cards
-search, filters and sorting
-guest mode
-invite links
-mobile responsiveness

I also added the option to upload a custom cover photo for a campus. This makes every campus feel more personal and attractive.

I tested all the main features properly and checked the website on diff screen sizes as well.
The main issues with login and invite links are also resolved now!

And finally, I am done with the website now. And will ship it soon!!!!

Devlog 4

Today I worked on gettin MemoryMap much closer to a proper working product.
I first fixed and improved a lot of things in the dashboard, campus cards, delete option, login flow and guest mode.

A user can copy a campus link, send it to a frnd, and that person should either log in first or directly join if already signed in.

I also added the campus delete option. The UI part is done, but I spent a lot of time debugging why deletion works locally but fails on the deployed Vercel site.

Rn most of the main website is done, just gonna make few more changes and checking all features of the website.

Devlog 3

I made a way where anyone can try making a campus map as a guest, instead of signing in. And also improved the UI for setup page.

And added another option to delete a created campus, but i am unable to make permanent deletion work…. and have been trying since sooo longgggg, have been going from the code to vercel to firebase and back to the code. I spent over 1.5 hours on just vercel and firestore settings.
I am tiredddd…. hope this works soon.

Devlog 2

After finishing the login page, I started working on the complete MemoryMap flow.

Rn users can create a new campus by entering the place name. After that, they get a proper setup page where they can add floors, rooms and corridors & stairs. Rooms can be moved, resized and renamed, so users can design the campus in their own way. Sounds exciting right?!

I also made the completed campus view in a dark theme, where users can click on any room and open its memory panel.
Inside each room, users can add incidents in text form and also upload images.

I also added the invite system.
Campus owners can copy a private invite link and send it to their frnds. The shared campus also appears later on their own dashboard.

Along the way, I fixed many UI issues, modal layouts, navbar structure, text colours, Firestore permissions and campus loading errors.

Overall i am lovin this idea rn! and its working well tooo!!!

Devlog 1

Introducing MemoryMap

Today I started working on MemoryMap, a private digital place where memories are saved according to the place where they happened.

So, instead of searching through old photos or WhatsApp chats, you open your private school map, click a place like the Chemistry Lab, classroom, auditorium or basketball court, and see all the photos and stories connected to that place.

You can create your own private campus, add rooms, invite only your friends (so whatever memory map you create is private), and everyone can save memories inside the rooms. Basically, it’s like walking through your school again online and reopening all the moments that happened there.

I thought about this idea, when i myself felt when me and my friends pass out schl, we’d forget sooo many incidents that we laugh at daily! And to keep those memories safe in one place, this website will be the best thing! @satvikhardat @Churu @Harman

Right now, after brainstorming a lot, I started working on the landing page, login page, and UI of the website! And i have attached an ss of the demo of the dashboard that i have added on the main landing page. I am soo excited tbh for this website cuz i can also use this later to save soo many incidents and pics.

#Devlog 4 – ProblemFirst

1. UI & Experience Refinement
   Reworked the homepage layout to improve visual hierarchy and clarity and shifted to a text-first, minimal design approach to better reflect the product’s intent.

2. New Loading Experience
   Added a custom loading screen to improve first-time user experience.

3. AI Experience Improvements
   Refined how responses are displayed using structured cards, improved visual clarity of AI-generated insights.

ProblemFirst is now complete — a fully working, thoughtfully designed product ready to be experienced.

Edit - I changed the UI and make it into somethin more fun lookin too! can’t change the pic of homepage in the ship, so doin it here.

Devlog 3— Completing ProblemFirst

It’s been a while since I wrote a devlog because I’ve been spending most of my time actually building instead of documenting.

Over the last several development sessions, I put in many hours refining nearly every part of ProblemFirst, and today I’m happy to say that the first complete version is finally done!

The biggest focus was redesigning the user experience. I moved away from the generic AI-generated look and rebuilt the interface with a cleaner, more minimal design inspired by modern product websites.

I also added subtle motion graphics and smoother interactions throughout the platform.
I reworked the navigation, improved the search experience, redesigned the solution pages, simplified the overall information architecture, refined the business portal too.
A lot of time also went into making the UI feel more intentional.

After countless refinements,
ProblemFirst has evolved into the vision I originally had: a platform where people start with their problems instead of product names, while giving startups and small businesses a fair way to connect with the people who genuinely need their solutions.

But before shipping this project, I will look into the project a bit more, and make a last few changes, and look out for any mistakes as well.

Sooo excited to ship this project!!!

#Devlog 2
Today I continued building ProblemFirst, changed the landing page UI as well, wanted to keep it a bit simpler.

Including yesterday’s time a bit as well, I had also completed making the basic pages for Trending, Explore, How it Works, and for businesses page. Now I have completed making them polished and look premium.

What improved:
-Search now accepts any problem
-UI transitions into a solution workspace
-Results show in structured cards:

1. Situation
2. Why it happens
3. Steps
4. When to seek help
5. product suggestions.
   Earlier it was static, so yh now its better.

Also, I Connected the Gemini API, Set up Firebase project, Enabled Google Sign-in, built a showcase version for the businesses page.

Today I actively worked with:
-Next.js (React framework)
-TypeScript
-Firebase (Auth)
-Gemini API (AI)
-Tailwind CSS

CURRENT ISSUES (TO FIX) are Gemini responses still slightly template-like and Background animations not fully active. Also, Card navigation still inconsistent.

Today was the day where major changes were made to this website!!!!

#Devlog 1

What I’ve Started Building:
Today, I officially started building ProblemFirst — a platform that flips the usual way people find solutions. Where instead of searching for the product the people tell their problem and My website guides you to the right solution.

Who this helps-

1. People facing real-life problems
   Someone struggling with focus, health, money, daily issues. They don’t know what to search or what to trust.

2. Small businesses (VERY IMPORTANT)

Many small brands have amazing products
But they don’t have big marketing budgets and hence don’t get visibility.

Things i completed doing today:

1. Homepage (Major Focus)
   -A clean hero section
   -Search-first approach:
   -User directly types what’s bothering them
   -Supporting UI

2. Categories Page
   -Created category cards:
   Home & Living
   Work & Study
   Health & Wellbeing
   Pets & Family

• Improvement to be made - Typography can be more premium, Cards need stronger visual hierarchy.

1. Search Results Page
   (This is the brain of the product)

-It takes users inputs
-Basic matching system in place
-Results show the solution direction

• But right now only limited solutions are there for limited problems.
  More problems with solutions have to be created.

Languages I’m Using -
(Out of these using a few for the very first time)
Next.js, React, Tailwind CSS.

Gonna do these next -
Finalize homepage UI (make it more premium)
Improve the search intelligence
implementing authentication (creating the sign in page too)

That’s it for today!!
I love the UI designed for the page, and there are animations in the background too! so it looks pretty cooll.

# Devlog 6 — Today’s Progress

Today was a major development day for Blink! I worked on expanding the language features and making Blink easier to install and use.

-Added New Language Features
-Added support for logical operators: and or -Implemented while loops, allowing Blink programs to perform repeated execution based on conditions.

Added Installation Support.
Users can now install Blink directly from GitHub:
pip install git+https://github.com/arpitaa31/Blink.git

After building the core features, improving the structure, and testing the workflow, I am now shipping Blink v0.1 — the first official version of my programming language. I will still keep working over it and add more features to this programming language.

#Devlog 5

Today was another huge milestone for Blink! The language is starting to feel much more like an actual programming language now.
Added support for numbers in Blink. Improved the interpreter to handle numerical values instead of only strings.
Made arithmetic operations work successfully, including expressions using operators like +, -, *, and /. Improved BinaryOperationNode so Blink can create and process proper expression trees.

Spent a lot of time testing and debugging how these new features interact with the existing lexer, parser, AST, and interpreter, yesterday as well. Understanding how every feature needs support from multiple parts of the language made the process much clearer.

A few challenges I faced:

1. Making sure different data types (strings and numbers) are handled correctly.
2. Connecting new syntax with AST nodes and interpreter logic.
3. Fixing bugs where small parser changes affected execution.

Blink can now:
-Print outputs using say
-Store and use values
-Perform calculations
-Take user input
-Make decisions using conditions
-Repeat instructions using loops

Next, I want to focus on improving error handling, adding more features, and eventually creating a playground where anyone can write and run Blink code.

#Devlog 4
Today’s session was probably one of the biggest milestones so far!
-Added support for the say statement in Blink.
-Built a SayNode in the AST.
-Updated the interpreter so it can actually execute say statements.

Spent some time understanding how the whole project flows, which made everything much clearer.
Started working on arithmetic expressions.
Created a new BinaryOperationNode for expression trees. Added a parse_expression() function that can understand expressions like 10 + 20.

The parser can now understand arithmetic expressions, and next session I’ll make the interpreter actually calculate them.
In short i worked on the interpreter today!

Successfully made Blink print - “Arpitaa”:

Devlog #3
Today’s session was mostly about moving Blink from just reading code to actually understanding it.
What I worked on:
-Finished setting up the Parser for Blink.
-Created the basic AST (Abstract Syntax Tree) nodes:
-Connected the Lexer and Parser together.

So,
Made Blink recognize simple assignment statements like:
name = “Arpitaa”

And successfully generated Blink’s first AST (Abstract Syntax Tree which is a tree-like structure created by the parser that represents the meaning and structure of a program.) , which means the language can now understand the structure of code instead of just reading characters.

Challenges faced:
Spent some time debugging Python indentation errors.
Accidentally mixed up code between parser.py and ast_nodes.py, so I reorganized the project files properly.

Now next, I have to build the Interpreter, so Blink can actually execute code. The code i ran and the results i got on executing the code today :

#Devlog 2 Hi! So, basically i mainly continued working on the lexer. The best part of today was finally running Blink. Blink was able to read it and print tokens for variables, strings, operators and keywords one by one. Seein my own programming lang successfully read a source file! Today’s work made me understand that the lexer doesn’t execute the program. Its only job is to read the code and break it into meaningful pieces that the next stage can understand. Anddd now i have will start working on building the Parser. After running a code, i was able to see this:

Devlog 1 Today I finally started working on Blink, my own programming language. I had no idea how programming languages actually worked, so I first learned the basic flow. I worked on the Token system. Basically, Blink now knows the different types of things it can see, like strings, variables, keywords etc. I also made a Token class that stores those values. After that, I started working on the lexer. Rn, it can move through the code one character at a time, skip spaces, read strings, and identify blink’s keywords.