Recently for a project I was working on, I needed to set up a video streaming backend for an iOS app. After digging around and seeing what the potential solutions for this were, I decided to use Amazon AWS as the backend. Next, in terms of the format that I should use, it seemed that http live streaming was the best option. Since it took a bit of time to organize and layout how this system worked, I figured it would be nice to write what I found.

Steps to Building your own Streaming Backend 

1. Save an original video file to disk that you want to stream. This will be typically be some quicktime format, mp4, etc.
2. Log onto your Amazon AWS account (or make one) (http://aws.amazon.com/) and create 3 s3 buckets. One will be for the original video files, one is for a thumbnail preview of each video, and one will be for the actual streaming video files.
3. Either through code or the amazon web interface, upload that file to the s3 bucket made to store all video files.
4. Now, to convert these videos to a streaming format, we will need to do several things. First, I think it is probably a good idea to go over how http streaming actually works. so:
   1. HTTP Live Streaming
      1. Basically an extension of audio (mp3) streaming
      2. File format is M3U8.
      3. This looks something like:
         

         ---

         #EXT-X-VERSION:3 #EXTM3U #EXT-X-TARGETDURATION:10 #EXT-X-MEDIA-SEQUENCE:1 # Old-style integer duration; avoid for newer clients. #EXTINF:10, http://media.example.com/segment0.ts # New-style floating-point duration; use for modern clients. #EXTINF:10.0, http://media.example.com/segment1.ts #EXTINF:9.5, http://media.example.com/segment2.ts #EXT-X-ENDLIST

         ---

      4. Essentially, it is a list of a ton of smaller .ts files. Apple recommends these to be about 10 seconds of video each. So, we need to convert our large video file into a ton of small .ts file clips, and then generate a M3U8 file that describes all the clips we need to make up our original video and therefore let the video actually stream.
5. So, one bucket will hole these .m38u files (and all the corresponding .ts files) and then that file url is directly playable by various UI elements of iOS.
6. To convert from .mov files or any original format to playable .m38u file, you will need to use the Amazon Elastic Transcoder.
   1. Open up the Elastic Transcoder on Amazon AWS console.
   2. Create a new Pipeline and hook up input and output buckets
   3. Create a new Transcoding Job.
      1. Input key:  filename of your desire
      2. output key prefix: folder you want your output in i.e. ‘myStreamingVids/’
      3. I used a system preset of HLS 2m
      4. Use a segment duration of 10
      5. output key: whatever you want
      6. Set master playlist name to final video url stream name
      7. set playlist format to be hlsv3
      8. set ouputs in master playlist to be your output key.
      9. Create job. 
   4. This will kickoff and run this job.
7. Use Amazon AWS Lambda
   1. Set up notifications to be called whenever a new video is uploaded to s3 input videos bucket. (I have it called on all new uploads)
   2. This calls your Lambda function. (its a .js file)
   3. My JS file extracts file info and then creates an Elastic Transcoder Job and adds it to my pipeline.

Unlike 2011 and 2012, 2013 was less of a year of new things. This past year was a year of focus and learning. By now, I have settled into college life, and have already lived and worked in the Bay area for two summers. I am no longer a teenager, and have entered my twenties. Overall, I spent more time working and thinking, and became more deeply involved in what I did in my everyday life – whether it be relationships and friendships, software engineering, or schoolwork.  

I spent less time on social sites, and instead spent more time reading. I took more of an initiative to keep my things organized and my room clean. The past few years were focused on breadth – more people, more topics, more adventures, etc. This year I still had those interests, but instead had more of a focus on depth – really learning about my friends, taking time to learn computer science and what it takes to be a good software engineer, reading up on music, coffee, etc. Instead of wasting a ton of time on Facebook , constantly refreshing the page to see if any friends posted a new picture or if I have a new notification, I would spend time looking up bands I liked, or reading more technical websites on computer science and programming. I found myself wanting to spend more time with a smaller group of friends, rather than spend my time at larger parties or events with lots of people I only barely knew.

It was a good year. I was able to spend a summer at Apple in California working on iOS as an intern. It looks like I even (hopefully) got a patent out of it. I drank a lot of coffee and talked to a lot of cool people.

2014 will be interesting. I will be entering my final years of college, and be turning 21. I am excited.

Last week I finished my sophomore year at the University of Michigan for computer science. I had a ton of great experiences throughout the year and here are a few things that stood out.

1.) Computer Science and Software Engineering != Knowing to code. This new trend of “learning to code in 10 days” needs to stop. It is simply not the same as learning to build high quality software, and learning how to understand and interpret complex systems and problems. While yes, I knew how to code last year, I was nowhere near the ability to solve complex problems as I am now after taking more computer science courses and spending time learning how to approach larger, complex problems.

2.) Find friends that push your limits. (in different areas) Find a friend that is a genius at (or at least passionate about)  math, programming, art, english, etc. By having friends who are experts in fields different from your area of expertise, your conversations will be a blend of what you are good at and what they are good at. This will naturally lead to ideas that are cross discipline, both making you more creative and open minded.

3.)Explore. By far, some of my best times this year were those not on campus. I was lucky enough to take several trips with Michigan students. First, I was able to go to UPenn for a huge student hackathon. Being on a huge bus full of 50ish computer science majors and exploring the city of Philadelphia was a ton of fun – and lead to some great ideas. It was also fun building an iPhone app with friends for two days with little sleep. Next, I was able to go to San Francisco a few times – once for interviews, once for a conference – with fellow students. It was such a powerful feeling to have 20-30+ students travel 2000 miles and take over a city for a week. During the second trip, Michigan students also assisted at the conference making a huge impression on all of the recruiters, developers, and other people at the conference. They were amazed so many passionate students would use their spring break to come to a tech conference and hackathons to build products (or show off previously built projects)

4.) Never become comfortable with your current skills or abilities. It is critical that you never become too comfortable with your current level of abilities or knowledge. Constantly push yourself to either know something even better, or learn something new.

So, this will be a fairly short post, but I frequently get asked about storing data in iOS.

Essentially, when it comes to storing data  in iOS you have several options:

1. NSUserDefaults. If the amount of information you want to save isn’t terribly large, and consists of NSData, NSString, NSNumber, NSDate, NSArray, or NSDictionary, you can easily store it in NSUserDefaults. NSUserDefaults can be read/written to in much the same way you would work with a NSDictionary. For example:                                                

   NSString *test = [NSString alloc]init];
   [test stringByAppendingString:@"test!"];
   //creates string object
   
   NSUserDefaults *myDefaults = [NSUserDefaults standardUserDefaults];
   [myDefaults setObject:test forKey:@"string"];
   [myDefaults synchronize];
   
   //and this data can be accessed by:
   NSString *newTest = [NSString alloc]initWithString:[myDefaults objectForKey:@"string"]];                               

2. If you REALLY insist on writing to a local text file of some sort, do not put it in your application bundle. The bundle is read only, and you will spend an absurd amount of time debugging this.
3. You could use sqlite. Code can get a bit tedious and ugly, but it still works fine.
4. Core Data. http://www.raywenderlich.com/934/core-data-on-ios-5-tutorial-getting-started

That is all!

I love entrepreneurship, but the term is losing its meaning and is becoming a word to defend a group of lazy, arrogant people who are too “intelligent” to work and instead go to meetups and talk about each other. This must stop.

The reason I fell in love in with entrepreneurship was the core idea of following something that personally excites you and transforming it into something that can help other people. You work incredibly hard building something because you really enjoy making actual things.

Unfortunately, it is extremely easy to use adjectives and meet with other people who like these same ideals, but do absolutely nothing towards actually making things.

To build great products you need a few things:
1.) A drive to learn, and be willing to give up everything else to spend hours learning marketing, finance, programming, or design.

2.)Be able to see through bloat. You will have many, many people approach you with their ideas, opinions, etc. It is critical you do not waste your time with people who are ignorant, and just like talking about big ideas and do nothing. Find people that are willing to work their ass off, and really understand the industry. For example, one thing that NEEDS to end is this new trend of completely inexperienced business students or non-programmers starting startups that help people learn to code or learn to run a startup. This creates bloat, and is a complete waste of time. If the person running this startup for startups has never actually run a company or written a line of code themselves, what makes them qualified to be a teacher? They are not qualified. Now, if they hire people to teach at this new code or startup school that have experience, this is very valuable and a great move.

3.) Learn from mentors. I have learned pretty much everything I know from experienced developers, business leaders, etc. School is incredibly valuable, but getting solid mentors gives you real world experience. For me, I was lucky enough to get a job at Detroit Labs right out of high school for the summer. I cannot even begin to explain how much I learned just by sitting in a room with the four cofounders and seeing how they ran the company. Additionally, being a Kleiner Perkins fellow this summer was incredible. Being able to talk informally with some of the top VC’s was so valuable.

4.) Be willing to work. Really freaking hard. Don’t trick yourself into thinking sending emails is hard work. Actually work. One very valuable thing going to a very competitive school has instilled in me is learning to work hard, and more intelligently . School makes you to be evaluated by someone other than yourself. This can expose critical flaws. School is not the only way to get this type of feedback, but it is a very great means to do so.