Bloggity Blog

We love our work so much, we want to tell the world about it.

Pain Points - Solve the Right Ones

Written by Nate Sowder

I often talk with entrepreneurs about the importance of having a defined project scope. In a SAAS product, it’s important to limit the features list. Early on, I struggled with this in my career because I tried to do too much, too soon. I felt that the more problems I could solve with one solution would increase its permanence or give it greater value. In practice, one small addition to scope opens the door to another. Once the new pathway is developed, it too opens the door to more ideas and additions which in turn, creates a big hairy problem.

The more features I add or change, the later my launch date. Don’t get me wrong: sometimes you have to pivot, but rarely do flashes of inspiration need to be added to an MVP (Minimum Viable Product). Remember, the number one goal is to get data points that show usage and the potential for greater profitability. I can only get data points if people are using my product.

Everything comes back to research. Good research controls a project’s scope. Asking better questions and thinking more critically about the answers is a good formula to get research that is useful. This will save time and money by creating a more solid foundation for the product. A product that will surely be built out with shiny features down the road.

Starting Out: Ask yourself these questions

  1. What will your project accomplish?
  2. What deliverables do you need to see to mark the project as ‘Complete’?
  3. What features are not going to be included in the project and are deemed ‘out of scope’ for this current version?
  4. What resource limitations does the project have including time, money, and manpower?
  5. Are there any dependencies the project assumes will always be available (API’s, prospects, niches, etc.) that you could lose access to?
  6. Most importantly, how much money/opportunity will I lose for every month my project launch date gets delayed?

Managing scope is difficult, especially as more people get involved.

The addition of co-founders, advisors, and investors make sticking to the overall passion like trying to pass a bill through congress with no additional changes. Having the following areas defined will make life much easier as more people get involved.

  1. List all requirements associated with the MVP. To get these, it’s important to understand the customer groups being served by the product. Setting up use cases are beneficial to get requirements.

  2. Define the Project’s Scope. The MVP needs to be made as fast as possible so it can be tested. Sometimes there are features that can wait until a later version to save both time and money. Let’s take the Jeep I drive to work every day as an example. Chrysler offers various Jeep models that include all different types of features. Some features come standard, like seat belts and airbags. Other models come with features like power locks and windows, heated seats, towing packages, moon roofs, and leather. Keep in mind though, that the least expensive, stripped down version can still get me to work in the morning.

  3. Verify the Project’s Scope. Don’t be afraid to show your product to people. Having beta testers lined up is great way to project the initial success or failure of the product. If no one wants to use it for free, then why would they use it when they have to pay for it?

  4. Control Your Scope. Looking back, every time I’ve had to change or modify the scope of a product resulted in me not doing enough homework. There’s no excuse for scope creep. Get it together, man!

The End Goal:

Make something so brilliantly designed and developed that the people using my product will never realize how complicated the technology behind it is.

In last week's Dayton Ruby meetup, Chris Chernesky spoke about how he uses Tmux for his everyday Rails development. It was great talk and ever since the meetup, Tmux has been all the buzz at the Littlelines office.

I've been using Tmux for a few years and it has become an essential tool for my workflow. Tmux's window and session management make it a no-brainer for those who live in the terminal. In this article, I explain a few tmux commands and tools that I use daily to help me work more effectively across rails projects.

Setting up Tmux

Let's start off on the right foot. Since we're going to spend most of our day inside tmux, we might as well configure it to best suit our needs. Here are just a few tmux configurations that are essential for me.

# Remaps tmux prefix to Control-b
unbind C-b
set -g prefix C-a

# Improve colors
set -g default-terminal 'screen-256color'

# Navigate around panes easily using vim-like keybindings
bind h select-pane -L
bind j select-pane -D
bind k select-pane -U
bind l select-pane -R

# Remove administrative debris (session name, hostname, time) in status bar
set -g status-left ''
set -g status-right ''

The biggest configuration change I made is remapping the default Tmux prefix from Control-b to Control-a instead. Since I've also remapped my caps lock key to the control key, this is much more convenient for me.

Now let's take a look at a couple aliases I setup to help navigate tmux on the command line. With these three aliases we can easily list the running tmux sessions, join a current session, or kill tmux sessions.

alias tml="tmux list-sessions"
alias tma="tmux -2 attach -t $1"
alias tmk="tmux kill-session -t $1"

Killer Rails Project Management with Tmuxinator

Working with a typical Rails project may involve several commands and long-running processes. For example, we may need to run the web server e.g. bundle exec rails server or we may need to run Sidekiq for our background workers. Tmux is great for this scenario because we can harness the power of Tmux's windows and panes to organize these processes in a single screen.

To further make our lives even easier, Tmuxinator makes managing Tmux sessions a snap with the use of simple Yaml files. The best part, it's a ruby gem so all we have to do is run gem install tmuxinator and we're ready to go. Here is an example of a configuration file I setup for one of my Rails projects.

name: website
root: ~/code/website
  - editor:
      layout: main-vertical
        - emacs
        - guard
  - tests: bundle exec rake test
  - console: bundle exec rails c
  - workers: bundle exec foreman start
  - logs: tail -f log/development.log
  - server: mosh

With this configuration, I just run mux website and Tmuxinator will cd into the project's root directory, create a new (or join an existing) Tmux sessions, create six windows, kick off emacs and guard panes, run tests, and start the background workers. Just one command and we're ready to go to work.

Switch Projects and Windows with Ease

So now we have our Tmux setup and Tmuxinator helping us with session management, we can start moving around our various projects. We can do this in a couple ways: from the command line or within Tmux sessions.

From the command line:

Getting around tmux sessions on the command line is pretty easy. Using our aliases we setup earlier, we can use the tml command to list the current sessions and tma to join an existing session. We can also use the mux [project] command that Tmuxinator provides to either create a new session or join a session if it is already running.

Tmux on the command line

Within Tmux sessions:

Once we're in a tmux session, getting around windows is easy too. I usually hit Command-a [Window #] to jump to a window or Command-a w to toggle a list of windows to open. But what if I need to switch to another project within Tmux?

Thankfully we can run command Command-a s to toggle a list of running sessions to open. This is a great way to navigate all your projects without leaving Tmux.

Tmux on the command line


What I really like about this setup is that not only can I navigate projects easily, but I can setup projects just one time - saving the need to create window or panes every time. And the kicker, I can hop into one project, make a change, and hop right back where I left off with just a couple keystrokes.

Do you have any Tmux tips you'd like to share? Please leave us a comment below.

Littlelines Roadtrip (Summer Edition)

Written by Matt Sears

The Littlelines crew will taking the show on the road this week. We always like to take the opportunity to stretch our legs and get out of the city for a few days to meet folks in the industry. This time round we're loading up the car and sticking to the great Midwest.

Steel City Ruby - Pittsburgh, PA

Steel City Ruby Logo

Our first stop will be Pittsburgh for the Steel City Ruby conference. Steel City is entering it's third year and the speaker line up looks like it's going to be another great one. We had the pleasure once again of designing the conference tee shirts and we think they look really great.

Weapons of Mass Creation - Cleveland, OH

WMCFest Logo

Next up, we're heading north to Cleveland for Weapons of Mass Creation. This is the premier art, design, and music event in the country. This is the fifth year running and has grown to a 2,000+ attendee event. Checkout the amazing speaker line up - can't wait to listen to all the great talks!

If you see us, stop and say hi, we'd love to meet you! And hey, you never know - we might have goodies to hand out.

Welcome Dustin Armstrong

Written by Matt Sears

Dustin Image

It is my great pleasure to welcome Dustin Armstrong to our development team! Dustin brings his clean, responsive, front-end coding skills to team with a design talent to boot. Dustin is a 2009 graduate of School of Advertising Art and has been writing solid code for the past five years. Dustin's experience in design and development will help our clients products be a glowing success. You may follow Dustin on Twitter and on Github

Welcome aboard Dustin!

A Survival Guide for Legacy Rails Apps

Written by Matt Sears

Hello, my name is Matt and I love working with legacy Rails apps. Ruby on Rails is now over 10 years old. That means there are a lot of (old) Rails applications running out there. At Littlelines, we've worked on hundreds of Rails projects. Most of them we build from the ground up, but often we have the opportunity to work with legacy Rails apps - some as old as 2006! More often than not, we discover these apps include many many common mistakes made back in the day and it's our job to fix them. Over the eight years of writing Rails apps, I've picked up a few tips and tricks that can help us get through the agony of legacy code.

1. Take Stock of the Situation

It is impossible to understand the present or prepare for the future unless we have some knowledge of the past. So first things first, let's take an assessment of where we're at in terms of test coverage and code quality. If we're lucky enough to take on a project that has tests, the first thing I do is install SimpleCov to measure how well our app is tested. Once I have the SimpleCov report, I'll take a screenshot and save the report. Then we can use CodeClimate to get an overall grade on quality and security. CodeClimate will analyze our app and report on all the hotspots and security violations in the code. Finally, I'll write this number down and take a screenshot.

Now that we have metrics to work with, we can do some Opportunistic Refactoring. In other words, always leave code behind in a better state than you found it. The fun part is seeing how far we can improve the code's test coverage and quality and we can challenge ourselves to take the score from an F to an A.

2. Sharpen the Saw

If I had eight hours to chop down a tree, I'd spend six sharpening my axe. – Abraham Lincoln

As developers, we spend most of our day buried in our text editor. Luckily for us, most text editors allow us to customize and fine tune it to make it work more effectively for us. Even better, some editors like Vim and Emacs allow you to create custom functions to help automate repetitive and complex tasks. These can be incredibly useful when working with legacy code. One of my favorite functions converts old Ruby 1.8's hash syntax to 1.9's new syntax.

Ruby Hash Converter

Some text editors like Vim and Emacs ship with built-in Macros support. Macros allow us to record keystrokes and play them back. These can be great tools to automate simpler tasks with less effort. The nice things about Macros is we don't need to write a function, we just need to hit the record button and play it back.

Let's take look at an example. Let's say we are upgrading a Rails application and we discover that it's using the old style of validating ActiveRecord fields. We can create an ad hoc macro to convert the first line of the validation code to the new validation syntax and play it back for the next three lines.

Emacs Macro and Playback

3. Learn Something New

As soon as we stop try new things, we stop learning. We can always be better Rubyist and so as a general rule, I try to learn something new when starting a new project. It doesn't have to be anything huge. In fact, it's usually a small thing and something that fits with the project. It can be anything from replacing our view templates with a new template engine like Slim or something as small like using Ruby 2.0's new keyword arguments. Most likely our legacy app is using some old and unmaintained gems. This is a great opportunity to see what we can replace them with. Head on over to the The Ruby Toolbox or Awesome Ruby and see the latest and greatest libraries available.

To give you an example, A few years ago I was tasked with upgrading an old Rail 1.2 app to the latest and greatest version. During the upgrade process, I discovered a lot of embedded SQL and the project owner had complained about how slow the searching has been on the site. So, I thought it would be a good opportunity to learn more about how to do full-text searching in Rails. And this lead me to discover the great Sunspot gem. With Sunspot, I was able to eliminate all the embedded SQL and make the search perform much much faster at the same time.

4. Have a Plan

As a rule for every new project, I make a list of things I'd like to accomplished by the end of the project. It's usually a small list that contains some very high level goals. In most cases, the goals coincide with making a better Rails app and me a better a developer in the end. For example, here is a list I made on my last project.

  1. Upgrade application to Rails 4 and Ruby 2.
  2. Raise Code Climate GPA to 4.0.
  3. Increase test code coverage by 20%.
  4. Watch RubyTapas episode on Null Object and apply it.
  5. Try out Byebug gem and see how it stacks up to Pry.
  6. Write a new Emacs Lisp function that converts erb to haml across multiple buffers.

5. Keep Pushing

As many of you may know, working with legacy projects can be boring and frustrating. But, we can take steps to make it a little more fun and learn something new in process. Even when the code is horrendous and we're cursing who ever wrote this pile of #@$%, we can still learn something new and challenge ourselves. And it isn't always new tricks or tools we learn, it's also the past mistakes that teach us. If you commit to taking these steps, you'll improve your skills at a much faster rate and you'll find yourself stepping out of your normal routine and applying new solutions that will ultimately lead to you becoming a better developer.

How about you? If you have any tips for working on legacy Rails apps, please add a comment below.