Miscellaneous Tech

• Bitcoin resources → Youtube resource      Learning Bitcoin from CLI
• APK (Android Package Kit) - file format used by Android to install apps
• has all components of app and packaged in single apk → code, resources, assets
• Like .exe are executables on Windows → APKs are for Android
• Mod APK (Modified APK) - altered version of original Android app
• Sideloading - installing apps on device from sources other than official app store
• Used for apps not available on Play Store or for beta versions
  
  
• XAPK (extended Android package kit) - APK file + OBB files + graphics + media
• compressed for easy distribution of large apps/heavy games for android
• not officially supported by Android → special installers are needed
• Manually, users can extract XAPK and place OBB files in respective folder
  
  
• OBB (Opaque Binary Blob) files - extra data files used by some Android apps to store large assets like graphics, media files, and game data that don’t fit in APK
• APK works as the installer and OBB provides bulk of the content
  
  
• After we Delete File permanently → system do not immediately erase data → only index pointer (the file’s address) is removed from file system → data is still intact
• In Hard Disk Drives (HDD), data blocks remain intact until new files overwrite them → the sooner you try to recover, the better your chances are
• Sometimes, we may get corrupted images (half visible, half damaged)
• In Solid State Drives (SSD), if TRIM is enabled → OS quickly wipes free blocks
• very hard to recover data
• In Phone Storage (Flash Memory) - recovery chances are very low because memory cells are reused quickly
  
  
• Dual-boot - install multiple OS on separate partition/disk with bootloader at startup to select & load only that OS’s kernel and system files from its partition while other OS stays inactive → after installing Linux, Windows Boot Manager replaced with GRUB
• X’s (twitter) API return both id and id_str → some environments can handle 64-bit integers directly, but in JavaScript integers larger than 53 bits risks precision errors
  
• unacceptable risk when identifiers must remain exact
• id: 64-bit unsigned integer              id_str: same identifier represented as string
  
  
• Proxy hiccup - temporary, short failure in proxy layer but client and backend are fine 
• Causes - when proxy reloads its configuratio, edge-case connection or websockets can be forcefully dropped during worker process swap
• proxy has limited pool of TCP connections to backend servers → If sudden spike in traffic, then proxy may run out of ephemeral ports → reject new requests until old connections close and free up space
• Solution for client - clear cookies as error redirection may be stored in cache, retry again as network fluctuations could be there, check internet
• If client gets error, automatically sends request again → but if all 10,000 clients retry immediately at exact time, they will unintentionally cause DDoS attack and crash proxy permanently → Exponential Backoff with Jitter + Idempotency
• client waits 1s to retry → If that fails, it waits 2s, then 4s, then 8s. "Jitter" adds random few milliseconds to each wait time so that all clients don't retry at exact same synchronized moment
  
  
• CMDB (Configuration Management DB) - centralized repo that stores info about all critical components (Hardware & Software) of organization's IT environment
• each component is called Configuration Item (CI) and its detail is called attribute
• CMDB map and track complex relation and dependency between all CI whether they are hardware, software apps, cloud service, or network infrastructure
• In ITIL (Information Technology Infrastructure Library) framework, CMDB acts as "single source of truth" for IT operations
• CI can also be Logical Entities like Policies, SLAs, specific IT personnel
• CMDB data is fed via manual data entry or automated Discovery Tools that scan corporate to detect new device, monitor config changes, update dependencies
• Benefit - If router needs maintenance, consult CMDB to check all dependency to schedule downtime safely & notify right people. To find exact failed component, if any service goes down. Useful for large companies to track cost and value.
  
  
• Compiler vs Interpreter
• Compiler - translates entire source code into machine executable code at once before execution → Faster execution + more efficient
• checks for syntax errors at end during compilation → Example: C, C++
• Interpreter - translate code line by line and execute instruction by instruction
• Easy to detect errors but slower execution → Example: Python, JavaScript
• Authentication vs Authorization
• Authentication - confirms identity of users by verifying login credentials
  (e.g., username, password, biometric) → happens before any authorization check
  
• Authorization - checks if user has permission to perform certain actions/access specific data based on roles, permissions, or access levels
  
• Integrated Development Environment (IDE) - place to write, run and debug code
• Example: Visual Studio, PyCharm, Jupyter Notebook
  
• storage/emulated/0/ - root directory of user-accessible internal storage on Android 
• represents user’s primary storage ("Internal Storage”)
  
  
• To run batch file in PowerShell → add ./ before script name for security reasons
• This 100% ensures that running batch file of current directory
  
  
• Software bloat - when program becomes slow and oversized due to inefficient code or unnecessary features → apps consume more memory, storage, CPU
• coder-decoder (codec) - compress and decompress digital data (audio + video files)
• Audio Codecs - MP3, AAC (used by iTunes and YouTube)
• Video Codecs - H.264/AVC (widely used for streaming, Blu-ray, and web video), HEVC/H.265 (better successor to H.264), AV1 (royalty-free codec for streaming)
• AV1 → built by alliance of Google, Microsoft, Netflix, Amazon, etc. bcz H.264 charges royalty → give efficient bitrates compared to other codecs (means better video quality at low internet speed)
• Issues - consume more CPU/GPU + very slow at encoding + compatibility issues with old devices since AV1 is new
  
  
• Date Based Versioning - Stripe does not use traditional ways to name API versions like v1/payment or v3/users → uses "Stripe Version : 2015-07-18" passed via header
• BENEFIT - Users on older version have no tension to switch every time new version arrives (Maintains trust) + All versions available till date can be used
• How is it possible to maintain thousands of version ?
• Request Compatibility Layer → normalise any old version of API calls to latest version → then send to core business logic (written wrt latest version)
• Gate System → Feature flags based on version date and feature
• Let's say "amount" is deprecated in new version but old one still use it
• new version API → "amount" will be off but for old ones it will be on
• No messy if-else statements + All gates are at one place + High code readability + Easy to manage
• Transformation modules → used when sends response back to user
• small, isolated modules to downgrade latest version to user's version
• latest version → downgrade to 2025 → downgrade to 2024 → so on...
• In 2027, only make for 2025 and 2026 (bcz others already exist)
  
  
• Digital Object Identifier (DOI) - unique alphanumeric string assigned to a document providing persistent link to location on internet → attached to object (not location)
• In past, researchers cited sources using URLs → if website was reorganized or publisher changed servers → links break (called "link rot") → can't find source
• Even if journal moves article to new web page → DOI remains same and will redirect you to correct place → like permanent "digital fingerprint"
• Parquet - optimized for fast processing & high compression in big data ecosystems → group data by column reducing read time, storage space when query specific fields
• closed protocol - company don't want to publicly document/share how their system works internally → like WhatsApp never publish how their app talks to their servers

***********************************************************************************

Image Formats

• JPEG (Joint photographic experts group) - lossy compression
• JPEG = JPG due to new windows update of "3 characters limit on file extension"
• convert RGB to YCbCr → Y=brightness, Cb=blue difference, Cr=red difference
  → used in compression bcz eyes detect variation in light and shadow better than fine details in color → discard Cb and Cr color channels + keep Y channel sharp
  
• Discrete Cosine Transform (DCT) → image is divided into 8x8 pixel blocks 
• DCT transform these blocks into frequency domain → isolate gradual color changes (low frequency) from sharp edges or noise (high frequency)
• Loss - High-frequency details are divided by quantization matrix and rounded off
  → permanently discard less visible data before using lossless Huffman coding
• EXIF (Exchangeable Image File Format) - standard to store metadata directly within image → automatically record GPS coordinate, camera model, date, time, exposure settings (like ISO and aperture) → take screenshot to remove EXIF
• social media platforms automatically remove EXIF to protect privacy
• Formats like TIFF, JPG, WebP, and modern PNGs can retain EXIF data
• PNG (Portable network graphics) - lossless compression [bigger size than JPG]
• reduce file size without discarding visual data → preserve image quality
• supports alpha channel which allows smooth, variable background transparency
• Used for flat illustrations → where colors are limited and can be counted
• TIFF (Tagged Image File Format) - industry-standard format used in professional photography & print publishing → preserves absolute, maximum image quality
• use no compression → preserve all layers, depth, sharp details → large size
• SVG (Scalable Vector Graphics) - XML-based format that use math equations (lines, curves) to draw image instead of grid of pixels → since rely on math rather than fixed pixels, it can be scaled infinitely to any size without losing quality or becoming blurry
• stay sharp across screen density + small in size
• WebP - modern image format developed by Google → provide both lossless and lossy compression (optimized for web image) → highly versatile, supports both transparent backgrounds of PNG and animated frames of GIF + significantly small size than JPEG
• HEIC (High-Efficiency Image Container) - modern format popularized by Apple uses advanced video compression techniques to store photos at roughly half size of standard JPEG + better visual quality → highly versatile, support up to 16-bit color depth, transparent backgrounds, can store multiple images or audio within single file
• Compatibility issue → only supported by Mac, IPhone
• AVIF (AV1 Image File Format) - open-source, royalty-free image format derived from AV1 video codec → outperforms formats like JPG, WebP and HEIC
• offers highly small sizes + preserve superior image quality
• considered next-generation standard as supports advanced features like High Dynamic Range (HDR), wide color gamut, transparency and animations
• Issue - old browsers don't support it yet

***********************************************************************************