Perhaps a better title for this post would be “Back from the dead”! It has been so long that I got lost in my own blog setup. So lost, that I couldn’t help but wonder if it was really a good idea to use Octopress in the first place. Finally, I got around to writing this post, but not without having to check the Octopress documentation!

Anyway, now that I am back here, I intend to write more often. In fact, that’s one of my new year resolutions! But I am getting ahead of myself. Before I go on to commit to some goals for this year (as I should have done soon after reading The Motivation Hacker in 2014), I’d like to review the last year. Before any one blames me of ‘plagiarism’, I’d like to give credit, where credit is due. This format is highly inspired by some other similar review posts I have come across over the years

1. For $10,000, your car can drive itself
2. Laser message from space
3. Google Domains (invite-only beta)
4. Google may buy video search startup Baarzo
5. Plaintext password offenders
6. Lotus: A complete webframework for Ruby
7. Coder’s high
8. We have the potential to solve the biggest problems of today Read on →

1. Using Wearable Computers for Passive Haptic Learning and Rehabilitation
2. Cgrep: a context-aware grep for source codes
3. Google backed Nest acquires Dropcam
4. B.C. court ruling orders Google to block sites worldwide
5. Google launches starter kit for a more consistent web

External links (not summarized below):

1. Discussion on possible solution of the GHash problem in Bitcoin network (would try to summarize this in a later post)
2. Quantum Experiment Shows How Time ‘Emerges’ from Entanglement

1. Facebook breaks the network switch into modules
2. Facebook Saves Developing World’s Data Plans With 65% Smaller, 50% More Efficient Android App
3. Facebook outage; down for around 30 minutes
4. Code Spaces destroyed after attacked deletes Amazon-hosted data
5. Elon Musk: Hopeful that the first people could be taken to Mars in 10 to 12 years
6. Skype for Linux 4.3 Arrives with Desktop Call Monitor Widget, UI Changes

1. Amazon announces Fire Phone : First phone with head-tracking tech
   • 4.7 inch screen, Quad core 2.2GHz, Adreno 330 graphics, 2GB RAM.
   • Runs FireOS, a fork of Android.
   • Available on contract with AT&T for two years; starting $199
   • While the hardware is not exceptional, its ability to track user’s head is what makes it different from the rest.

     This is done through four corner-mounted, front-facing infrared cameras and produces wild 3D effects. But it’s not tradition 3D in the sense that images pop out the screen. The head tracking tech adds depth to the phone. It’s like looking down a street. But into your phone’s screen.

   • Using the head-tracking capability, Amazon built Dynamic Perspective, which allows a 3D image viewing experience in the phone. [Read more below]
   • Fire and Prime

     Amazon has built this phone for the Prime subscriber. Much like with the Fire TV, the Fire Phone will ship with a person’s Prime information already entered. The phone is, after all, all about selling more wares and Amazon is ensuring there are no barriers to entry for its best customers, those who subscribe to Prime. It also includes the personalized customer support service Mayday that debuted on the Kindle Fire.

   • Fire phone comes preloaed with a feature called ‘Firefly’ (read about it below).
   • 13MP campera. The quality of the photos compares with those from Samsung Galaxy S5 and iPhone 5S (Read more about the camera below).
   • Comes loaded with popular music streaming apps : Spotify and Pandora. Not restricted to Amazon Prime Music

   • Mayday, which allows customers to interact with service agents over phone or video chat in just one-tap, is also available on Fire Phone

   • Mayday would make it easier for customers to work with unfamiliar features.
2. Amazon Fire Phone’s Dynamic Perspective offers a new angle on smartphone displays
   • Amazon introduces a new way to experience a smartphone by using 3D effects.

     Instead of creating the illusion that an object is coming right for you, you’re given a, well, dynamic perspective on things. Look at a lockscreen picture of a forest, or egyptian ruins, and you’re given a chance to peer behind or around some of the objects on screen. For a more practical application, consider the maps: during the presentation, Amazon CEO Jeff Bezos tilted the phone to pan about the map, getting a better angle on details like a particular street’s name. Tilt the phone gently, and you’ll pull up Yelp info and reviews.

     Or maybe you like shopping – in one example, Bezos pulls up a line of dresses on Amazon, and tilts the phone to scroll through the stack. Reading more to your liking? Just tilt the phone to scroll up and down a web page or e-book. You can even set things to autoscroll at a deliberate pace, essentially putting your literature on cruise control – drop a thumb to pause i

   • Field of view of a single front-facing camera is too narrow to track user’s head
   • Fire Phone uses 4 cameras mounted on the corners to track head and eyes.
   • The company collected a lot of images of faces and applied machine learning to perfect the tracking capabilities.
   • Dynamic Perspective SDK is available
3. Fire phone introduces ‘Firefly’
   • A feature that lets you identify things that you see in the real world.
   • You can also directly buy off using this.

     During the demonstration on stage, Amazon CEO Jeff Bezos showed off how Firefly can be used to scan barcodes on items, a book cover, a CD and more. There’s a Shazam-like audio recognition component as well – Firefly can identify songs, and it can even listen to a show, and then let you access that specific scene on Amazon, which is a clever trick.
     ….
     Firefly can also recognize art, and then pull up the associated Wikpedia entry, something Apple’s Siri and Google Now can’t yet do.

     At launch, Firefly is capable of recognizing over 100 million items, said Bezos. That includes 245,000 movies and TV episodes, 160 live TV channels, 35 million songs and 70 million product

   • There is a dedicated button on the phone (besides the volume rocker) for Firefly!
   • Firefly looks like an extension of Flow, Amazon’s visual-recognition technology that used to reside in its own standalone mobile app.
   • When pointed at a sign board, it can identify the number (using image-to-text) and determine if a phone number doesn’t exist, or directly call/save the number on the sign board.
   • Firefly SDK will also be available, including text, audio and image recognizers, content databases and support for custom actions allowing developers to bend Firefly to their own personal needs.
   • MyFitnessPal created an app, using the SDK, to find the nutritional value of food item seen through the camera.
   • The ability to recognize objects/entities through the camera lens of a phone would significantly benefit Amazon as identification is the first step in the purchasing process.
4. Fire phone’s camera and unlimited cloud storage, for free!
   • Amazon Fire Phone hosts a 13MP camera
   • Side-by-side shots show that the images are comparable to those from iPhone 5S and Samsung Galaxy S5. However, it’d be best to wait until more reviews pour in or try out yourself before commenting on the camera quality.
   • There’s a dedicated camera button, allowing users to jump to camera directly by clicking the button (without unlocking the phone or searching for the app)
   • One click opens the camera app, second takes a photo.
   • Unlimited storage on Amazon Cloud Drive

     Each Fire Phone comes with unlimited uploads and storage to Amazon Cloud Drive, for life, for free. There are no conditions on that offer, which makes it much more attractive than either the paid offering from Amazon (iCloud Photos) or Google’s Drive offerings, which also use paid tiers to limit how much you can upload at once.

‘Tech News’ is a series of articles where I attempt to share and summarize recent developments in technology (and the related industry) that interest, amaze and/or excite me. These would attempt to cover topics ranging from impressive apps to Google’s latest acquisition to jaw-dropping updates from SpaceX or Tesla Motors and much more.

1. Facebook’s Slingshot challenges SnapChat with ‘Reply to Unlock’
   • Facebook’s new app Slingshot is being seen as a competitor to SnapChat.
   • It lets you share photos and 15-second videos with users on Slingshot.

   • The content is meant to be ephemeral, much like SnapChat.

     As for what happens to your content, once everyone you’ve slinged something to views it, Facebook says it will delete the photo or video from its servers. There is a seven-day delay in case it has to investigate abuse reports, but the team says the content is gone after that. Whether users trust Facebook to keep its word is up to them. Past privacy missteps might scare people, but Facebook is under close scrutiny from the FTC to abide by the privacy policies it puts forward.

   • An important catch in the app is that you can’t view the content shared by the user until you reply to it. (‘Reply to Unlock’)
   • This aims to increase reciprocation and make everyone a content generator.
   • Slingshot was born out of a hackathon. The motivation can be summarized by

     “My brothers are really not technical in any way”, Slingshot inventor and four-year Facebook Product Designer Joey Flynn tells me. “I’d take a photo, upload it in iMessage, and send it to them. And I would just see “Seen”, or in a month I’d see them and they‘d say ‘that was a cool photo’. But hey, I want to see what you’re up to! There was this missing reciprocation” he explains. That was the spark for Slingshot.

   • This app was earlier released on app stored by mistake
   • Facebook had tried to buy off SnapChat, but SnapChat had rejected the $3 billion buyout offer

2. Browser based javascript cryptography considered harmful

   Some of the things pointed out in the article. [Not being an expert in cryptography, I would recommend readers to read the original article to avoid any miscommunication]

   • Code could be injected when in transit. Can use SSL/TLS to avoid that, but then, in most cases, you could simply avoid cryptography at user end.
   • There’s no certainty of information being removed from memory.
   • It is not possible to verify the code/environment to be exactly as it is required to be.
   • Javascript lacks a Cryptographically Secure Pseudo-Random Number Generator (CSPRNG).
   • Stanford Javascript Crypto library (SJCL) developed at Stanford is a great work, but still insecure.

     The authors of SJCL themselves say, “Unfortunately, this is not as great as in desktop applications because it is not feasible to completely protect against code injection, malicious servers and side-channel attacks.” That last example is a killer: what they’re really saying is, “we don’t know enough about Javascript runtimes to know whether we can securely host cryptography on them”. Again, that’s painful-but-tolerable in a server-side application, where you can always call out to native code as a workaround. It’s death to a browser.

3. UK intelligence forced to reveal secret policy for mass surveillance of residents’ Facebook and Google use
   • Secret policy define UK residents’ communication on Facebook, Google, Twitter, etc. as ‘external communication’ because they use platforms based in the US, even if the communication is between two people in the UK.
   • This grants the secret services to indiscriminately intercept communication on these platforms.
   • For an ‘internal communication’, a warrant would be required, however that is not the case with ‘external communication’.

     The Government believes that, even when privacy violations happen, it is not an “active intrusion” because the analyst reading or listening to an individual’s communication will inevitably forget about it anyway.

4. Dropbox acquired Parastructure, a Big Data Startup -

   So what exactly is Parastructure? According to the company’s LinkedIn profile, and its (now-offline) home page, it builds “beautiful data analysis software powered by cutting-edge open source infrastructure.” Parastructure’s GitHub pages give some clues as to what that data-analysis software was tackling, with areas covering Spark for cluster computing; Phoenix, a SQL skin over HBase; CrunchBase; and so on

   • The acquisition was done quite stealthily.
   • Probably, Dropbox requires the software and the team to analyze its data
   • Parastructure was an enterprise oriented startup and its acquisition indicates Dropbox’s intention on strengthening their image as an enterprise product.
   • Dropbox’s increased effort in Dropbox for Business brings it in direct competition with Box, which is currently leading as an enterprise solution.
   • Both the companies are quite aggressively acquiring companies.
   • In fact, Parastructure was acquired on the same day as Box acquired Streem
5. Apple adds a new entry level iMac for ~$1,100
   • 1.4GHz dual-core Intel Core i5 processor, 8GB memory, 500GB HDD, Intel HD Graphics 5000 on board    

‘Tech News’ is a series of articles where I attempt to share and summarize recent developments in technology (and the related industry) that interest, amaze and/or excite me. These would attempt to cover topics ranging from impressive apps to Google’s latest acquisition to jaw-dropping updates from SpaceX or Tesla Motors and much more.

1. Qeexo, a startup, is trying to redefine how we use touchscreens
   • Since the unveiling of iPhone, 7 years ago, not much has changed in how we interact with touchscreens.

   • Qeexo has developed a touchscreen technology capable of detecting difference between a fingertip, a knuckle, fingernail and a stylus.

     Today, FingerSense uses the standard accelerometer in a mobile device to pick up on the vibration patterns different parts of the hand produce when they come in contact with a touchscreen. The FingerSense machine-learning engine understands what part of the finger is touching the screen and triggers a correlated action. For instance, tapping a block of text with your knuckle could trigger the copy and paste menu to appear. To select text, you’d simply drag your knuckle down the text,much like you would a desktop mouse.

     Read on →

Due to the lack of an official Evernote client for Linux and dissatisfied with solutions like Neverpad and NixNote (formerly Nevernote), I have started working on my own Linux client.

It was while developing this that I learned that Evernote SDK for C++ doesn’t really provide any integration with OAuth. I couldn’t even find a well formed solution for the same on the Internet, except some forum posts suggesting using any available OAuth 1.0 client. Searching for OAuth clients for C++, I found O2. Using this, I was able to create a simple OAuth integration for Evernote.

Here are the steps :-

In an earlier post, I had discussed how I integrated Docspad API with Hasjob. As there was no Docspad SDK at that time, its absence made me realize their importance. Having to deal with raw API calls, these were the primary problems that I faced, which I thought could be resolved by a Software Development Kit (SDK) :-

1. API lets me talk to your tool, but not in my language

   Say I am developing in Python and I have a dictionary of parameters explaining all that I need the API to do, but the API doesn’t understand the dictionary! I would need to convert all those parameters and variables in GET/POST variables that before I can get the task done. Moreover, the conventions in your API might be extremely different from those of my project, making things worse.

   Certainly, I could work around that by wrapping around your API, which brings me to the next point.

2. Managing network calls and errors is annoying

   There are API errors and then there are network/connection errors, timeouts and what not. The last thing I need when dealing with a project idea is to get stuck understanding what an error code means. SDKs not only help translate the content in the developer’s language, but also the errors. An exception that describes the error makes it easy for the developer to handle certain edge cases without the program blowing up in the users’ face.

3. SDKs provide code samples

   Even if a developer doesn’t need the entire SDK, either because he only requires to make a few calls to the API or he/she does not think the SDK interacts well with his tool, an openly available SDK still provides code samples to the developer to help him/her manage basic things. Be it managing errors or creating request URLs, SDKs can act as a good demonstration of how the APIs must be used.

Niketh had already been planning to release SDKs himself, so when I suggested it to him, he was eager to get started! Soon Read on →

About two months ago, Niketh came up with an idea to augment the job application process on Hasjob. Having made the quite awesome product, Docspad for handling documents in web applications, he certainly noticed the lack of the option to let a job applicant upload a CV while applying through Hasjob. While the user could add details about himself/herself in a text field, nothing tells as much about an applicant as a personalized CV. As the code for the website is open sourced and written in Python, I took this opportunity to check out the Docspad APIs and spend a weekend contributing to a Python based open source project!

But before I dive into the task and the experience, let’s first discuss more about Hasjob and Docspad.

Hasjob is HasGeek’s job board. Hasgeek, based out of Bangalore, is known for organizing technology events for geeks and developers in India. Some of these events include Hackathons and Hacknights, which are loved among the geeks. HasGeek lives up to its name by providing the source code to their website and its components under open source licenses on Github.

With the goal of making digital content more accessible, Niketh and Aditya founded Bookpad (@bookpadinc). Docspad is their first product, which provides an easy solution for developers to make applications with document handling capabilities. The set of APIs make it easy for a developer to provide users with the option to upload and view documents through the web browser on any device ranging from mobile to desktop. Read on →