github_interview.md

Scenario 1

A product I love is BurdaStyle.

####Features To Add

• Premium email newsletter subscription that comes with a few simple patterns and articles about techniques.

  This European based company publishes a magazine multiple times a year with very limited distribution in the US. non-European customers can subscribe to the US version of the magazine but it's expensive and orders take a long time to process.

  This feature would allow them to engage more of their customers in a more timely manner with some of the great content they include in their magazine.

• Send print shop ready versions of their patterns so customers can easily have the pattern printed at a local print shop, such as FedEx stores or an architectural print shop.

  Burdastyle specializes in selling all of the patterns from their magazine in digital format. The sewer gets a pdf of the pattern broken up into a grid of 8.5x11 pieces of paper. They must print it out, trim the blank edges away and tape the grid together using numbered markers on each edge of paper. A typical pattern can be around 50 pages. What you end up with is a unruly patchwork of paper that, unless you went crazy with the tape, doesn't stay in tact very well.

  Digital patterns have helped them offer patterns a la carte without having to get the subscription but once the sewer has the pattern the experience is time consuming and frustrating.

• Online basic sloper generator which will create a digital sloper pattern based on a user's measurements.

  Every at home sewer needs to adjust commercial patterns to get the best fit. Commercial patterns are designed around a sloper based on a set of "standard" measurements. To draft a sloper, a patternmaker uses the measurements and follows a set of linear steps. An expample step might look something like "Draw a horizontal line on the bottom of your paper one quarter of your waist measurement."

  Once you've sewn a few garmets most people begin to have the desire to modify the pattern a little bit, both for fit and for self-expression. There comes a point when sewers will try their hand at drawing up their own sloper so they have a solid base to alter commercial patterns or make their own designs. While it might sound easy, the instructions out there are horrible.

  Commercial software exists for fashion designers but most are old and clunky. They are also more expensive than a home sewer wants to pay. Instead this feature would allow a user to input their own measurements and get a file they can have printed and build off of.

####To remove

• Advertisements

  They are implemented in an annoying way and completely distracting. Either reimplement them in a less intrusive way or find a better way to generate revenue.

####Their customer

The target customer is an English speaking sewer over a wide range of experience levels. Typically someone gets interested in sewing their own clothes because they get a custom tailored piece that no one else has for way less than they would pay at a store. There's great satisfaction in wearing something you made yourself.

####Threats to their market share

Currently other large pattern companies online offer their patterns in paper format. If those companies made the transition to having more of an online presence and began offering their patterns in digital format, especially done well it would put pressure on BurdaStyle to improve their product. Small designers are already utilizing digital patterns but don't have the vast library of patterns or exposure as BurdaStyle. It's possible that a small designer with an established successful digital pattern business could partner with one of the larger patternmaking companies. Burdastyle has already invested heavily in fostering vibrant online community, something no one else is doing at their scale.

Patternmaking software exists but is clunky and outdated. All available software only runs on Windows and costs around $100-$200

From a totally different angle, the company also offers online classes for sewing and patternmaking techniques. Currently Craftsy has a firm hold on that market and is doing it very well.

####Expansion opportunities

The craft sewing market beyond garments that employ very similiar skills with lots of overlap in customers.

Professional development for sewers who want to commercialize their skills.

####Pricing

The magazine is published with about 20-30 patterns per issue. Each of those patterns are for sale individually on the site at an affordable price point allowing them to continue making money off the designs. The product also makes money off of advertisements and online classes.

####Business Model

Selling to sewers looking to expand their skills and make unique modern clothes by selling digital patterns and online classes. Patterns are priced competively with physical patterns sewers can buy in a store. Classes range from $50-$150.

##Scenario 2

We live for our customers ideas. Without them feature Y would have never been born. We thank you. So for this very reason we we keep a close eye on the list. Recently you requested feature X and it quickly rose to the top. We examined it carefully and see where the root of the problem lies. Unfortunately when it's weighed against some of our other big projects in the pipeline we don't think it's something we could give our full attention to at this time.

We're looking forward to launching more of your features in the future so keep this ideas coming. In the meantime we're working hard to improve our product and deliver great things.

Thank you for everything you bring to Github.

##Question 1

• A user goes to http://github.com
• DNS looks up the destination IP
• Request is routed to the IP
• Load balancer directs to correct server for IP destination and port
• Nginx on server divies up request and sends it to a ruby web process
• Ruby web process is running a rails application
• Application looks to see if the user is authenticated using tokens sent with the request and checking in the database to seeing the credentials that were sent match the tokens
• Application then routes the request to the correct controller and action.
• If the user isn't authenticated the app will send back the main sign up/login page, most likely useing a cached version of the file rather than assemblying the html every time.
• If the user is logged in the application reroutes the request to the user's profile action
• This action will lookup the user's profile information first in the cache and if not found, assemble the needed data from the db
• The profile page html is created.
• Once the application has completed it's task, it sends the final html file in packets back to the browser
• Subsequent requests are then made for the assets on the page. If a Javascript application language is being used to construct the guts of the page it will make additional requests for data it needs to build the view.