It takes a country generations to build a supply chain from factory to storefront. It takes a VC-funded app six months to dismantle it.

The Man Who Sold the Future

The Rise and Stall of Gupta Ji

Gupta Ji wasn’t just a shopkeeper to me, he was a hustler before “hustle culture” had a hashtag.

I remember seeing him in his prime. He had one goal: to carve out a name for himself in the cutthroat cloth markets of MG Road. He did it all the kind of grit you don’t see in app interfaces. He borrowed capital on handshakes, took risks that would keep most of us awake, and spent midnights personally handing out pamphlets on street corners.

He wasn’t just selling fabric. When he tucked his kids into bed, he was dreaming of a playground for them, a solid home for his family, and a legacy he could pass down. He worked harder every year. He did everything “the right way.”

Until the day his hard work simply couldn’t pay off anymore…

The Monopoly of Convenience

We are witnessing a strategic silence. Right now, giant platforms are burning billions in investor money to make prices “too good to be true.” They aren’t just giving us a discount, they are subsidizing the destruction of the shop on the corner.

They burn money today so they can overcharge us tomorrow, once the streets are empty of options and Gupta Ji has been forced to pull down the shutters for the last time.

The Damming of the Dream

The market used to have a natural flow a ladder anyone could climb. One generation started in retail, the next moved to wholesale, and the third owned the factory. It was the “Indian Dream.”

Now, that flow has been dammed.

In this new “gig economy,” we aren’t creating businesses; we’re creating dependencies. We are told “Our New India” is growing, but I find myself asking: Growing for whom? If the rich get richer while the neighborhood shop dies, is that progress or just a consolidation of power?

The thought of a country with no independent supply chain where every link from factory to front door is owned by a single boardroom honestly scares me. It should scare us all. Because when the local retailer disappears, the ladder we all hope to climb disappears with them.

The Luxury We Never Asked For

We have to ask ourselves: Is getting a packet of chips in 600 seconds actually a luxury? Or have we just been trained to believe that waiting is a failure?

Every time we hit “order,” We are paying a Health Tax we didn’t sign up for. We live in the “Diabetes Capital of the World,” yet we’ve replaced a 10-minute walk to the Sabzi Wale Bhaiya with a 10-minute countdown on a screen. We are not just hurting his pocket, We are hurting our own heart literally and metaphorically.

Then there’s the Patience Deficit. When did we become this person that if rider takes twelve minutes instead of ten we feel a flicker of unearned annoyance.

I miss the Lost Connection. I miss the Tehl ke jaana. There was a rhythm to walking down the street, the small-talk bargaining with the vendor, the random “hello” to a neighbor. That’s the social fabric of life, and We are letting an algorithm shred it for the sake of staying in the comfort.

The Mirror Moment

We are trading long-term community stability for short-term dopamine hits.

And here’s the part that haunts me: That Sabzi Wale Bhaiya? If he goes out of business, he doesn’t just vanish. He joins the gig force. He becomes the shadow at my door at 11 PM, handing me a carton of milk because I was too “busy” to walk 500 meters. I’ve turned a business owner into my personal errand runner.

The question isn’t about economics anymore; it’s about who I want to be. Are we willing to trade the soul of our neighborhood for 10 minutes of convenience?

“Every time I visit home, I see another shutter down. Next time I’m back, I’m putting my shoes on. I’m going for that walk. Because the man who has run that corner shop since I was a kid deserves more than a ‘delivered’ notification he deserves to know he hasn’t been forgotten.”

Let’s rewind. It’s a winter evening in 17th-century Italy. Galileo Galilei the man who gave us the telescope is staring at a lantern in his hand. He’s wondering:

Does light move instantly? Or does it take time, like sound?

So he comes up with a plan. He and his assistant climb onto two distant hills, each holding a shuttered lantern. Galileo uncovers his lantern. The moment his assistant sees it, he uncovers his own. If there’s a delay, Galileo could measure the speed of light.

But every time they try, the reply comes back too fast. To Galileo, light seems instant. His tools aren’t enough. Still, he plants the seed:

Light has a speed. Someday, someone will measure it.

— — — —- *Fast — -—- forward— — -—200 — -— — years— -— — later.*— — — —

Imagine stepping into a time machine and finding yourself in the midst of 19th-century France, standing alongside the brilliant physicist Hippolyte Fizeau inside his laboratory. And unlike Galileo, Fizeau has a new idea and a machine. Picture it: a bright lamp, a toothed wheel spinning faster and faster, a mirror placed eight kilometers away, and a telescope aimed straight at it.

Here’s how it works:

The lamp’s beam passes through a gap in the wheel. It shoots off to the distant mirror and comes back. If the wheel spins slowly, the light slips back through the same gap, and Fizeau sees it flicker. But as the wheel spins faster, the returning beam doesn’t find the gap. It slams into a tooth instead and disappears.

That’s the moment. The missing light tells him: the wheel has turned by exactly one tooth in the time it took light to travel to the mirror and back.

Fizeau scribbles the math: number of teeth, speed of rotation, distance to the mirror. His result? 315,000 kilometers per second. A little higher than the modern value, but close enough to shock the world.

The idea was amazing and genious. It could measure the speed of light which is the fastest anything could be. And did it all with just the basic setup.

Galileo tried with lanterns on hills. Fizeau tried with a spinning wheel and a mirror. “Both chasing the same mystery.”

And that mystery? It didn’t end there. Their work lit the path that led to Einstein(E=mc2), to relativity, and to the modern universe we know today.

GSoC’19 — Wikiemdia

Project Synopsis (Link)

About the App

The Wikimedia Commons Android app allows users to upload pictures from their Android phone/tablet to Wikimedia Commons. Wikimedia Commons accepts only freely licensed media files. Users can upload images and then add various tags specific to them such as category, title, description, and license. Also Wikimedia Commons has won the “Coolest Tool” award at Wikimania 2019 in the Mobile category.

About the Project

The project aims to move towards Structured Commons. The plan is to modify the current procedure of media details in Upload, Search and in Explore in accordance with the Structured Commons property table. Prime features of Structured Commons include Captions and Depictions(multilingual and localised). The feature aims to use captions and Depictions while Uploading, Searching or for showing Media. As specified here.

Goal of the Project

Wikimedia Commons’ goal is to make media about a wide array of topics available to the whole world. Categories are used to help people find content, but unfortunately category search is only in English, making Commons less useful for 80% of the world. The solution to this is Structured Commons, which allows people to seamlessly search for images about particular topics in any language, and see media information in that language too.
While the Commons website has slowly begun implementing Structured Commons, this GSoC has made the Commons app a reference of what Structured Commons is meant to be:

• Search for media using one’s native language
• List media using native captions instead of technical filenames, only reverting to English if no caption is available in the user’s native language.

Features developed in this GSoC have sparked ideas in other projects, with enhancement tickets being created to replicate these features on the Wikidata and Commons websites.

Work Summary

Project Title: Add Structured Commons support to Commons Android app

The work mainly centred around introducing core features of Structured Commons(Captions, Depictions) in Wikimedia Commons App

Captions in Wikimedia Commons:

Captions are a feature part of Structured data. They are meant to store short, multilingual descriptions about files
This is a replacement of the previously used titles for images (titles were not multilingual)
Also now captions replace the previous convention of using title for filename

• At the time of Upload, title of image has been replaced with multilingual captions (This is an improvement over title as title wasn’t multilingual and doesn’t provides enough description of the image)
• Localised Captions are shown in media details for all the images that have captions
• Captions(localised) are available for all media searches in explore, they are shown on the image thumbnail instead of previously used file name

Depictions in Wikimedia Commons

Depicts is a feature part of Structured data. Multiple Depictions can be added for an image just like categories.
However unlike categories depictions is multi-lingual

• At the time of Upload now users are asked to enter “What does the image depicts?”. Several search results are shown(with thumbnail) on the basis of what the user has entered. They can select any depiction amongst the list of available depictions
• Localised depictions are shown in media details of all the images of Wikimedia Commons
• Search via depictions is also available in Explore where depictions in any language can be used to search for the media
• Search for parent classes and child classes of depictions are also added to the search feature in Explore. While searching an image the user gets three types of result (each separated into different fragments) — Media, Parent Classes, Sub Classes(of the depiction)

Preview

• Title replaced by Multilingual Captions

• Adding Depictions before Uploading image in any language (the preview here shows depictions searched in Hindi)

• Items are available along with categories for each image in Explore

• List of Items available in Explore (this feature works on every language, and shows depictions in User’s locale)

• Clicking on any Item shows Media, Parent Classes and Sub-classes associated with the item (classes available in user’s locale)

• Media details showing Captions and Depictions (in user’s locale)

Quick Links

• Github profile: Click here
• Blog posts: Click here
• Link of PR for GSoC: Click here

Final Summary

This is the last but one Week of my GSoC’19 journey. It has been an amazing Experienced to work under such highly experienced Mentors. I have learned a lot over this summer, most important among them are — Testing, Code Quality, Unit tests, Kotlin, Retrofit, Common practices for writing code and Documentation

Thank you, mentors — Nicolas Raoul, Ashish Kumar for your patience and help throughout this journey.

Thanks for reading this article. Here is a short Video I have created to help you guide for GSoC . Hope you’ll like the Video.

Let’s connect on LinkedIn, Twitter and GitHub.

Happy Coding!

Week 9 & 10 — Wikimedia Project

Tasks worked on in brief

• Showing localised captions in Media Details
• Showing localised depictions in Media Details
• Showing Items in Explore
• Adding Search via Depicts(localised) in Explore
• Replacing title with captions in Explore
• Replacing title with captions in Home
• Showing depicted images on clicking any item
• Adding pagination in Items and Depicted Images
• Tapping a depiction in media details should open search on that depiction
• Adding Unit tests
• Adding Documentations
• Fixing the bugs raised by mentors and other community members

Preview for the changes

1. Localised captions in German

2. Localised depictions in Hindi

3. Localised Search via Depicts in Explore

4. Replacing title with localised captions in Explore/Home (The image shows captions in German)

5. Showing depicted images on clicking any item

Link to various commits of the work

The given links refer to this PR

Commits for localised captions in Media Details: showing captions in Replacing title with captions in Exploremedia details, show captions in media details, captions working perfectly

Commits for localised depictions in Media Details: parsing response for depictions, displaying captions and depiction QID in media detail, displaying captions and depiction QID in media detail, displaying captions and depiction QID in media detail, showing depictions in media details in production flavor

Commits for showing Items in Explore and adding search via depicts: Showing items in explore, added search via depicts in explore, enabled pagination in depiction search

Commits for Replacing title with captions in Explore/Home: show captions in explore, show captions in home, showing captions in place of title in home and explore:media

Commit for Showing depicted images on clicking any item: showing depict images under items

Commit for Tapping a depiction in media details should open search on that depiction: fixed issue #3113

Commits for Unit tests: added some tests and media deatils in depiction detail activity, added depiction search test

Commits for Documentations: added documentations and worked upon review comments, added documentations and worked upon review comments, added documentations and code cleaning, added documentation and code refactoring, added documentations

Commits for fixing Bugs: fixed issue #3100, fixed issue #3098, fixed issue #3099, fixed issue #3104 and #3098, fixed issue #3108 and #3107, fix isse #3108, fixed issue #3110 and #3112, fixed issue #3113, fixed issue #3076 and #3109, fixed issue #3113, fixed bug with back button in media

Other commits: minor changes, minor changes, resolved merge conflicts, resolved merge conflicts wih backend-overhaul

Week 7 & 8 — Google Summer of Code

Tasks worked on in brief:

• Searching depictions from Depicts Activity
• Extracting Wikidata Id of Depictions during Upload to send it to Wikidata
• Fixing Bugs, Issues and Crashes
• Working on Unit Tests
• Working on Reviews by mentors on the PR
• Added Documentations

Preview of Depictions in UploadDepictions Activity

Link to various commits of the work

The given links refer to this PR

Searching depictions from Depicts Activity: searching depictions from depicts activity

Extracting Wikidata Id of Depictions during Upload to send it to Wikidata: sending wikidataentity id to upload depictions, uploading depicts

Fixing Bugs, Issues and Crashes: fixed crash due to depicts, fixed crash on selecting depictions, Fixed issue with next button diabling in media detail activity

Working on Unit Tests: added tests for depictions, added all the unit tests and fixed few more issues

Working on Reviews by mentors on the PR: added changes after review

Added Documentations: added some documentation and other changes

I haveOther Commits: modified depicts interface, Fixed issue with next button diabling in media detail activity

Challenges Faced

During this phase my prime Challenge was learning. I have started learnt basics of Kotlin. Also I learnt how to Write Unit tests. During this time I have studied a lot about Unit Tests and Instrumentation tests. And successfully written tests in Kotlin

Week 5 & 6 — Google Summer of Code

Tasks in Brief

• Storing and Retrieving captions from Local Database
• Implementing Search for Depictions Using Wikimedia APIs
• Using Data-Client Library to perform the network Calls like Uploading Captions and Depictions [WIP]
• Fixed Issues raised by Mentors and other community members

Link to Various Commits of the Work

The below links refer to this PR

Commit for Storing and retrieving captions from local database: adding captions to local database

Commit for Implementing Search for Depictions: added search for depicts

Fixed Issues: Fixed issue #3035, fixed issue #3033, fixed issue with project build, fixed issue #3044

Challenges faced in this period:

This period was primarily dedicated to the Bug fixes and transitioning from Wikimedia custom APIs to Data-Client library. Learning Retrofit and how to move from Custom APIs to Data-client library was the biggest milestone of this period.

Tasks worked upon in Week 3 and Week 4

• Adding APIs to send Multilingual Captions to WikiBase
• To fetch captions from the Modified UI and render them to the APIs
• Replacing file name everywhere in the app with the first caption entered by the user
• Fetching depictions from the UI and uploading them to WikiBase
• Adding some unit tests and documentations
• Resolved some issues raised by mentors and other community members

Link to various commits of the Work

All commits are made in this PR:

Commits for Uploading captions: send captions as labels, fixed issue with the captions, added upload for captions, uploading captions to wikibase, uploaded captions as labels to wikibase

Commits for Depictions: Merge branch ‘fix-2942’ of VitalyVPinchuk to use the Code for uploading Depictions

Commits for unit tests and documentations: added documentation, added some tests and documentation

Commits for some issues: fixed issue #3006, resolved issue #3004, fixed issue with categoryPresenterTest.kt, resolved issue with uploading captions, resolved issue with api call

Few other commits: resolved merge conflicts with refactor_uploads, optimised imports, #commit, #commit, undo formatting changes and a few more

What have I learnt in these two weeks

I have learnt a lot about Structured Commons. Also I have learnt Wikimedia Commons APIs (how to call them and how to parse their results)

Challenges Faced:

Initially I faced difficulty in calling the Structured Commons API. But bit by bit I become familiar with their implementation and parsing. Next was to use the new Data-Client library. This part was significantly challenging. However with the help and support of my Mentors Nicolas Raoul, Ashish Kumar, Vivek Maskara we were able to overcome blockers and call the new Library.

Thanks to the mentors and community members :)

Week 1 & 2 — Google Summer of Code

Goals of Wikimedia Commons’ GSoC

The goal of my current GSoC is to add Structured Commons support to Commons Android app.

• Submit captions to Structured Data
• Show caption in media details
• Show caption instead of filename in Contributions and Explore
• Structured Commons “depicts” property

Tasks in brief of Week 1 and Week 2

• Replace Title with multi-lingual Captions in First Step of Upload Activity.
• Add another step in Uploads: “Select Depictions”

A demo UI for the tasks done:

Below is the Modified UI of Step 1 of Upload Activity. Following changes can be observed:

• Captions instead of Title
• Captions are multi-lingual unlike the Single field for Title

Added a “Step 2” to the Uploads Activity for adding Depictions. These suggestions have image, label and description with them.

What I plan to do in next two weeks?

• Implementing the Structured Commons API for sending mutli-lingual captions to Wiki-base
• Fetching depictions from API using the captions/descriptions entered by the user, recent searches and by using TextWatcher

Link to me Current PR is here

Thank you

Contributing to Wikimedia is an amazing experience, it has many perks other than contributing to open source.

• Firstly, Going about the Wikimedia code-base itself gives great learning to Android developers.
• Apart from that Wikimedia has got highly experienced and cooperative mentors. For even beginners to easily start with contribute to the project.

Thus I am very proud to be selected at Wikimedia. 6 May 2019, was the beginning of my community bonding period. The tenure so far has been incredible I got to interact closely with the mentors. Both of my mentors (Nicolas Raoul and Ashish Kumar ) are extremely cooperative and interactive. We had two Video Meets and other offline communication as well.

In the first video meet, my mentors helped me to walk through the application, and understand the project from perception of other working communities of Wikimedia. I got to know about the Wikidata, how the information is shared to the Wikipedia database. And how this information is further used at Wikimedia and Commons.

In the second video meet we had discussions about “my progress so far” and “How to start with the project?” this is where I discussed upcoming tasks and other queries with the mentors.

Both the mentors are amazing and friendly to empower me in our general discussions.

Apart from the meets we regularly had offline discussions wherein I get various resources to go through or work upon. All these learning’s have made me much familiar with the working beyond the Commons mobile Application and to other Wikidata items.

The tasks that are assigned to me are always well structures and well explained.

My very first task has been to familiarise myself with the following three communities at Wikimedia:

People who develop the Android Commons app
People who upload and edit pictures on Wikimedia Commons
People who create and edit items on Wikidata.

I had been given various tasks to learn by actually contributing to each of the above three. That helped me learn in a quicker and an efficient way.

Secondly I was assigned the task of Getting familiar with the Structures Commons Properties table. (the core part of the GSoC project). The article provides all the necessary information about Structured Commons, need, what do we plan for future and how is the team working on the transformation.

Third, task in hand was to be aware about the APIs that are to be used in the project. I had gone through various APIs available here. I have tried some of these APIs using curl also picked up the APIs that are to be used for the Structured Commons project.

Alongside, I made myself familiar with app’s architecture. The architecture of the app is well understandable. Thus prepared I myself for the upcoming tasks, of starting the actual coding stuff

Overall, in the community bonding period I had amazing experience of interacting with my mentors and exploring the Wikimedia community. I am also prepared for the upcoming coding tasks in hand, thus looking forward for even more working and learning experiences at Wikimedia.

What is metadata?

Image metadata is text information pertaining to an image file that is embedded into the file or contained in a separate file that is associated with it. Image metadata includes details relevant to the image itself as well as information about its production.

Metadata also contains copyright information from the provider. These specialised instructions can be used to detect the source of the information like in this post we are going to detect whether a particular downloaded image was downloaded from Facebook.

(There is no documentation available for detecting Facebook special information. This is purely a research work. Based on study of data (metadata) available along with the images downloaded from Facebook)

What are we going to do in this Tutorial?

• Print all metadata of an image
• Check if a particular image is downloaded from Facebook.

What is FBMD?

On 2014–11–13, Facebook introduced a new tracking code(metadata in it’s files). It appears as an IPTC Special Instructions metadata field. They look like:

Special Instructions: FBMD01000ac60300004a1d00002d4b000067580000c9650000d5fc000054350100953a0100d3420100e84b01005f8f0100
Special Instructions: FBMD23000969010000b1590000cb7700000a8600000c07010046820100b8c0010052590200e5c902006e440300

More about FBMD:

In this special instruction each string starts with “FBMD”, two hex characters (1 bytes), and a 16-byte (4-character) length. Then comes length+1 sets of 32-bit (8-byte) values.

How to extract metadata using Java

The library ICAFE is used to extract the metadata from the image.

• Download the library and import the classes in your code
• In Android Development:

1. Download the jar file from here.

2. Add it into the project module in the app->libs package

3. Right click then select Add as Library

4. After successful gradle build we are ready to use the library

Let’s Code

We create a function which can extract metadata from any image

Let’s understand the above code :)

• Metadata.readMetadata reads the metadata from any image and returns a Map.
• We extract IPTC metadata from the map using the key MetadataType.IPTC
• Using the iterator we iterator we iterate over the IPTC data which is of the type MetadataEntry
• The printMetadata function prints the value of the metadata recursively.

Here is the code for printMetadata

How to check for Facebook Metadata

• Create a java class ReadFBMD.java and pass path of the image into the constructor
• Create a function processMetadata in the class ReadFBMD.java

Let’s understand the above code :)

• First we read the metadata using the image path
• Using getDataSets() on the IPTC metadata we obtain a map
• This map has a key called SpecialInstructions which contains an array, performing getData() on the first element gives us the Facebook data which is a byte array.

Where is Facebook metadata?

The byte array so obtained can be tested to contain FBMD data.

Note: Any image downloaded from Facebook contains the ASCII code of the letters ‘FBMD’ in this bytecode extracted from it’s IPTC metadata

Checking for FBMD:

if (dataInBytes[0] == 70 && dataInBytes[1] == 66 && dataInBytes[2] == 77 && dataInBytes[3] == 68) 
{    
        System.out.println("Contains FBMD");  
 }

Towards the end

Try it yourself….

Download an image from Facebook or from any other source and go through the data of that image. You can find many other information about the image.

Give it a try!!

Hope you have enjoyed reading the article. If you did please don’t forget to give claps on the article.

Let’s connect on LinkedIn and Twitter. For any queries feel free to drop a mail here.

Reading metadata from images was originally published in MindOrks on Medium, where people are continuing the conversation by highlighting and responding to this story.