<img src="/images/bluebolt-logo.png" alt="BlueBolt Solutions">
After nearly 7 years of full-time employment with IFCJ (and five years prior to that of part-time consulting work), I have chosen to move to other opportunities and challenges. This presented itself in the form of an offer from <a href="https://blueboltsolutions.com">BlueBolt Solutions</a> to join their team as a Senior Solutions Architect. My primary duties will consist of being a technical lead across several projects of varying size, architecting and scaffolding solutions, work with the project managers to ensure timely delivery of the completed project, and mentor junior/mid-level developers as they work on the project.
<h2>Day One - Initialization</h2>
As I expected, the first day was all about setting up my environment - getting applications installed, accounts initialized, paperwork completed. It has been quite a long time since I have built a work setup from scratch, so I fully expect to have forgotten tools that I used regularly. A (non-exhaustive) list of what I have already put in place:
<ul>
<li>Windows Subsystem for Linux (WSL) and Ubuntu</li>
<li>Visual Studio Code, including changing the default terminal to Bash and getting the environment set-up to run Bundler and Jekyll (for writing these posts).</li>
<li>Visual Studio 2019 Professional</li>
<li>Edge (Insiders, Chromium)</li>
<li>LastPass + Browser extensions</li>
<li>LinqPad</li>
<li>Microsoft Office Products
<ul>
<li>Teams</li>
<li>Outlook</li>
<li>OneNote</li>
<li>Word</li>
<li>Excel</li>
</ul>
</li>
<li>Node.js (for future React projects)</li>
<li>Spotify</li>
</ul>
<h2>Day Two - 8:00 Is Early?</h2>
I arrived at the building today at 8:00, only to find that it was locked! Thankfully, a delivery driver and an employee of another office in the building were both walking in just before me, and so I was able to walk in behind them. I'll have to find out when the building opens up, and how I can get access before then. Additionally, when I came up to the office, the lights were on and the door was unlocked, but there was no one inside. It's definitely a big change from IFCJ, where people started rolling in at 7:00, and the office was in full business mode by 8:00. I also need to get a key to the office and find out the feasibility of me starting work at 7:00. Will there be awkward moments when I'm leaving at 4:00? I'm sure it'll all work itself out.
Later in the day, I joked about being all alone in the office. I was given a set of keys, today - which is great. I can now come whenever I need to. I think I'm going to end up really appreciating the quiet in the morning - a chance to really get some dedicated work done.
<h2>Day Three, Four, Five - Getting Rolling</h2>
It turns out that there are about two other people who roll in before 8:30 - both of which seem to really like keeping a quiet, productive first hour or two. Which works out perfectly for me.
This first project, which fairly straight-forward, has its own set of challenges just getting up to speed. BlueBolt doesn't really have a formal structure for how projects, work items, branch policies, etc., are configured in Accelo or in Azure DevOps; however, this grants me incredible freedom to template the workflow how I would like it. Especially since the developer that will be doing the work is fresh to the company, there are no preexisting biases that I have to accomodate.
I'm going to keep detailed notes on what I created that worked out well and what needs tweaking for the next project. My long-term hope is to be able to get all of the Architects to converge on a very similar style of rules and structure in Azure DevOps, making it easier for developers to transition between teams.

First Week at BlueBolt Solutions

<img src="/images/2019-06-15-n1400.jpg" alt="N1400 for June 11th and 13th">
The first three weeks of the CARA summer marathon training are now in the books, and I am still trying to figure out how to best plan my running schedule. Because of schedule limitations (having three kids and a full-time job will do that), I can only really run on Tuesday and Thursday mornings by myself and Saturday mornings with the running group. In order to maximize the effects of these runs, I have started mentally adopting a "no easy runs" motto.
<h2>Intervals and Speedwork Twice a Week</h2>
Smart training plans include one or two days of speed work in a week, combined with three or four days of easy runs, in addition to a long run on the weekend. Getting from home to work can be a direct 7.5 mile run, an indirect route (east to the Lakefront Path, then south) of 10.5 miles, or really any distance longer than that. Since most speed workouts are shorter than this, I have to adapt my plans to accommodate the time I have. For example, a workout last week was:
<pre><code>Warmup (10 minutes or 2 miles)
2-3 sets (2x 200m + 1x 400m) @ 10k-5k pace
Equal jog recovery
Cooldown (10 minutes or 2 miles)
</code></pre>
This only adds up to 5-6 miles, which means I need to add a few more miles. Instead of 200m and 400m repeats, I just did 800m repeats.
<pre><code>Warmup
3 x 1 mile @ Tempo
1 minute rest
Cooldown
</code></pre>
This is workout is probably also only 5-6 miles, so my warm up was much longer and my cooldown was longer.
<h2>Weekly "No Easy Runs" Plan</h2>
To build these workouts, I am using the CARA <a href="https://drive.google.com/file/d/1Dhab_gIIOK0cGKpUCDZ4H9rsuP2qlZxv/view">Summer Marathon Training Program Guide</a> and taking the Monday and Wednesday speed workouts from the Advanced Training Plan. I am doing the same warmup routines, and then I am tacking on a few more miles to get.. um... really warmed up. The cooldown portion of the run I also try to maintain a slightly harder pace than I normally would.
For this upcoming week, my workouts are scheduled to be:
<pre><code>Tuesday: 4 miles @ Tempo pace (6:25/mile)
Thursday: 6-8 Yasso 800m (3:00/800m)
Saturday: 13 miles
</code></pre>
I'll report back here with an update on how it went.

Project BQ: No Easy Runs

<img src="/images/2019-06-04-n1400.jpg" alt="N1400 for June 4th">
June is here, which means the official beginning of the training season for the 2019 Chicago Marathon - only 130 days away. As I look ahead to what life is like now and will be by the time October comes around, I feel that this is my last, best hope (for a while, at least) to run a qualifying time for the Boston Marathon. Once this season is over, I will be undergoing back surgery, which the doctor estimates will take a year before I can be back to endurance running. This means I will be missing out on the 2020 season and potentially not getting back into BQ shape until the 2021 season - which puts me over 40 for the next time I can run Chicago.
Previous years, I have kept it quiet about my goal of running a Boston-qualifying time. Fellow runners would occasionally ask, and my first year thinking about it (2016) I kept coy with plenty of "we'll see what happens". I ran a 03:21:43 - only sixteen minutes away from going to Boston. During the 2017 training season, I was a little more open - "it would be nice, but there's a lot that would need to go right". 03:21:56 - I was no closer to my goal, and only getting older. Last year, I was very open - "I'm going for it this year - BQ or Bust!" I improved to 03:17:15, which is still over seven minutes short of qualifying and was almost 13 minutes short of getting to actually run.
To makes matters even worse, the organizing committee for the Boston Marathon has decide to lower the <a href="https://www.baa.org/races/boston-marathon/enter/qualify">qualifying times</a>. I now need to run faster than a 03:05:00 marathon just to qualify. Realistically, I am aiming for an even three-hours, as that will give me plenty of clearance to make the "cut-off" time. It is a lot to ask for, especially with a two-year-old and a newborn at home, a bad back in need of surgery, and an extra fifteen pounds of bodyweight that needs to be shed.
In an attempt to keep myself more accountable, I'm being very public about my goals this year. I am going to set milestones that I need to hit throughout the training season, rewards for when I get them, and documenting my journey. I'll be writing about my photo project, N1400; I will be tracking my workouts under my new motto of "No Easy Runs"; and I will be open and honest about how to manage relationships with friends, family, and coworkers throughout the season.
The next one hundred and thirty days are dedicated to my own <a href="https://twitter.com/hashtag/ProjectBQ">#ProjectBQ</a>.

Project BQ: This Time For Real

The primary piece of learning for me with this assessment was everything related to the front-end of the application. Creating a Web API in .NET and using Entity Framework is where my core developer strengths lie. While The Fellowship doesn't currently use .NET Core anywhere, my side projects and code experiments have all been in the new libraries as I prepare the team to make the shift. This was a fun little activity as a way to quickly get the basics of how Angular works - creating components, routing requests, abstracting communication with the server, and binding with the user input.
<blockquote>
The final post in a four-part series of posts on <a href="/series/technical-assessment-series">'An Assessment of a Technical Assessment'</a>.
</blockquote>
Framing this mini-project as a look into whether this is the "best I can do", it certainly falls short. In the interest of time, there were several features that I had to cut.
<ul>
<li>There is no meaningful user input validation</li>
<li>I never found the right place to use <code>ngrx/store</code> in the application.</li>
<li>I did not implement any server-side data validation</li>
<li>I did not create/implement a SQL Server database, instead just using the <code>InMemoryDatabase</code></li>
<li>I did not roll out a SQL migration script or do any custom model building</li>
<li>There is no retry policy on persisting requests to the database</li>
<li>Error logging is non-existent - I could have easily dropped in Application Insights, but it seemed overkill for this project</li>
<li>There are no unit tests, integration tests, or other automated testing to ensure the code is working as expected</li>
<li>There is minimal dependency injection, only for passing the data context around.</li>
<li>Least importantly, the UI looks terrible. But I never claim to be a designer, so I'm ok with that.</li>
</ul>
This project serves well as a Proof of Concept. It shows that I can quickly learn the basics of a new (to me) JavaScript framework, put together a simple solution to prove I can do it, and check it all into GitHub for others to examine.

Technical Assessment - Final Notes

I am starting to make some real progress with this. I have a better idea of how components, service injection, and input binding works in Angular. These are all concepts that I am very familiar with in my extensive use of <a href="https://knockoutjs.com/">Knockout</a>, but anytime one is learning a new framework, there will be a learning curve in figuring out how this particular framework implements various features. Next on the requirements list is the ability to create a new blog post.
<blockquote>
The third in a four-part series of posts on <a href="/series/technical-assessment-series">'An Assessment of a Technical Assessment'</a>.
</blockquote>
<h2>Creating a Form</h2>
Back to the Angular tutorial, this piece took several steps to get to where I needed to be to start writing code. Unfortunately, the opening tutorial (<a href="https://angular.io/tutorial/toh-pt6#add-a-new-hero">Add a new hero</a>) only shows how to declare an input's id and then how to wire a button to pass the value of that element back to the component. This works fine for a single value, but I need the ability for a user to edit multiple fields and send them all to component. I now needed to jump into the "Fundamentals" section of the docs; specifically, the "<a href="https://angular.io/guide/forms-overview">Forms</a>" section.
I decided to do a quick shuffle of files, and move the <code>blogpost</code> model into the "blogposts" folder. This is one of those decisions that I couldn't tell if it jives with convention, but it felt a little dirty keeping component-specific files in the root "app" folder. It's decisions like these that I'm excited to learn more about as I dig into larger Angular projects.
Creating the form, adding some styling to it, and binding it to a model in the component was all pretty straight forward. This is another case where knowing how one MV* framework (Knockout) informed me on what I was looking for, and the tutorial did an adequate job of showing how to put that concept into practice. After a bit of trial and error, I assembled the necessary controls to display a half-way decent form.
<img src="/images/techass-create-new-blog-post.png" alt="Create New Blog Post">
Working through everything, now was the time for the big test - would it actually submit my data to the server?
<img src="/images/techass-post-request-body.png" alt="POST Request Body">
Huzzah! This part works - now to the API controller to see if it's receiving the data.
<img src="/images/techass-modelbound.png" alt="Controller Model Binding">
Ok - well, it wasn't really that easy. I had to do a bit of refactoring from the default template. They were all expected changes, though. This part of the project was what I was most confident that I could get working fairly quickly.
I created a <code>BloggingContext</code> to handle the ORM side of things, and to make it easier, I just mapped it to an <code>InMemoryDbContext</code>. Changing the connection to a SQL database or an Azure Cosmos Db account would be fairly trivial.
<h2>I Think I'm Done</h2>
At this point, the chief requirements are completed.
<ol>
<li>Allow the user to view blogs posts in chronological order</li>
<li>Allow the user to create a blog post</li>
<li>Allow the user to remove a blog post</li>
</ol>
I'll write one more post about where I would take the solution from here if I wanted to really put together a complete software package.

Technical Assessment - Adding Create Functionality

Now that I have a static list, the next step is to make it slightly more dynamic. This means that I need to learn how to send something back to the server, instead of just retrieving data from the server. My guess is that a delete method should be easier than an insert of a new record, though I expect them both to be about equal. The first piece I recognized that I needed to add is an identifier for indicating which record to delete. I am a big fan of using a <code>Guid</code> as the datatype for a "primary key", so I quickly added a new field to the <code>BlogPost</code> and added that to my seed data. But how to I make a button in <code>Angular</code> and then make a <code>DELETE</code> call to the controller with the id of the selected post?
<blockquote>
The second in a four-part series of posts on <a href="/series/technical-assessment-series">'An Assessment of a Technical Assessment'</a>.
</blockquote>
It was time for me to turn to the <a href="https://angular.io/tutorial">tutorial</a>. The Tour of Heroes tutorial on the Angular website certainly looked to be what I needed. I tried jumping straight to <a href="https://angular.io/tutorial/toh-pt6#delete-a-hero">"Delete a hero"</a>, but it's talking about services and delegates and... whoa. Ok, time to take it back to the top.
<h2>Refactoring the Template</h2>
I zipped back to the top of the tutorial and quickly made sense of a lot of the work that the initial Visual Studio template created for me. Much of what the early parts of the tutorial teach the user were refactors that I thought would be necessary, but I didn't know how to do it yet.
<ul>
<li>Refactor the <code>BlogPost</code> interface out to its own class.</li>
<li>Refactor the data retrieval into a <code>BlogPostService</code> and inject it into the component.</li>
<li>Rename the <code>home</code> component to be <code>blogpost</code> - it's still that I'm still using the generic name.</li>
<li>Refactor the <code>blogpost.component</code> to use the new service.</li>
</ul>
What do I get after all that refactoring? Exactly the same output - which is what I had hoped for. No new functionality has yet been added, but I feel better about where things are.
<img src="/images/techass-recent-blog-posts.png" alt="Recent Blog Posts">
<h2>Delete a ~~hero~~ Blog Post</h2>
<ol>
<li>Add a delete button and have it call a delete method.</li>
<li>Create a delete method and have it call the delete service.</li>
<li>Create a delete method in the service, and have it call the API server's delete method.</li>
</ol>
<img src="/images/techass-404-delete.png" alt="HTTP404: NOT FOUND on DELETE">
Oh, and I tried all kinds of things to get it to work. I tried explicitly setting the route in the attribute. I tried explicitly setting the action in the route. I tried using different action names in the URL. I tried turning it into a <code>GET</code> method. NOTHING WORKED!!! What was I doing so very wrong???
After a much longer time than I wish to admit, the problem was that I needed to put the id parameter in the attribute. Clearly I needed more coffee this evening.
<pre><code>[HttpDelete("{id}")]
public void DeleteBlogPost(Guid id)
{
 new BlogPostService().DeleteBlogPost(id);
}
</code></pre>
From there, I was able to verify that the call made it into the <code>BlogPostService</code> that I had previously created. It properly removed the item from the original list and returned a <code>200:OK</code> response. All is well, and I am going to call this step complete. Once I get the entire front-end working, that's when I will handle persisting these changes to some storage medium. This part reminds me of my <a href="/posts/2016/user-can-create-new-patron/">User* Can Create**** New Patron*****</a> post from 2016.

Technical Assessment - Adding Delete Functionality

Yesterday, I did a phone interview with a company offering the possibility for a rather exciting position. The person I spoke with seemed very intelligent, articulate, passionate, and sold the organization well. It's a small organization that is growing, and is looking to consolidate all of their development work into their Chicago office instead of relying on developers in foreign countries. This feels very similar to <a href="/about">what I did</a> with Falkor Group and again with The Fellowship — start with a small team, map out a strategy for the next several sprints, then mentor and grow the team from there. Of course, the interview process is a two-way sales process, and they need to make sure that I can live up to expectations. This leads to why I am doing a Technical Assessment.
<blockquote>
The first in a four-part series of posts on <a href="/series/technical-assessment-series">'An Assessment of a Technical Assessment'</a>.
</blockquote>
The requirements are fairly straightforward.
<blockquote>
Technical Assessment
This assessment is intended to give us a better idea of your technical ability and code style as a developer. It will also give you an opportunity to learn about and utilize the tools we work with at the company.
Instructions
Create a dynamic blogging application that meets the following requirements:
<ul>
<li>Allow the user to view blog posts in chronological order</li>
<li>Allow the user to create a blog post</li>
<li>Allow the user to remove a blog post</li>
</ul>
Technologies
You must use the following technologies to complete this assessment:
<ul>
<li>.NET 4.6+ / .NET Core</li>
<li>Angular 5+</li>
<li>ngrx/store</li>
<li>Entity Framework</li>
<li>SQL Server</li>
</ul>
Notes
<ul>
<li>You do not have to implement user authentication into your solution</li>
<li>Your code should follow SOLID principles and make use of modern design patterns</li>
<li>Be prepared to explain your solution in depth</li>
</ul>
</blockquote>
<h2>Visual Studio -> New Solution</h2>
I set off right away to get working - this should be easy, right? I've never used <code>Angular</code> before (Knockout is the JavaScript framework of choice at The Fellowship), and I've never used <code>ngrx/store</code> (client state isn't that complicated in KesherNet). But Visual Studio has a template for creating a new ASP.NET Web Application using <code>.NET Core</code> and <code>Angular</code> - this should be super easy!
<img src="/images/techass-new-core-web-application.png" alt="New ASP.NET Core Web Application">
<ol>
<li>By the power vested in me, I command thee to build!</li>
<li>Oh, this project requires <code>Node.js</code> to build; so let's install that.</li>
<li>... lots of progress bars.</li>
<li>Great - now let's build and run this barebones project!</li>
<li>Internal Server Error? huh - well let's do a search for this error message "The NPM script 'start' exited without indicating that the Angular CLI was listening for requests."</li>
<li><a href="https://stackoverflow.com/a/53162998">This Stack Overflow answer looks promising</a>, I'll just do that.</li>
<li>... more progress bars. "MSBUILD : error MSB3428: Could not load the Visual C++ component "VCBuild.exe"." That can't be good.</li>
<li><a href="https://stackoverflow.com/a/39235952">Another Stack Overflow answer</a></li>
<li>... more progress bars. But it's all green text - so that should be good.</li>
<li>New let's build and run this barebones project!</li>
</ol>
<h2>Hello, World</h2>
<img src="/images/techass-hello-world.png" alt="Hello, World">
Well that worked out well enough - time to break this template down and make it do what I want.
The sample comes with three components - Home (display static content), Counter (show ability to modify non-persistent state on the page), and Fetch Data (demonstrate calling a Web API endpoint). I deleted the Home and Counter components, then renamed everything "Fetch Data" to "Home".
The initial object is a <code>WeatherForecast</code>, which I repurposed to be my <code>BlogPost</code> interface. The <code>component.html</code> is easy enough to repurpose to spit out the data coming from the controller. The new <code>HomeController</code> was refactored to return a collection of <code>BlogPost</code>. I added a second project to hold a <code>BlogPostService</code> which will handle the data retrieval. Later, this can be refactored to use EF and SQL Server to store and retrieve the data.
After throwing together some quick static seed data, the project was ready to go. I had a few typos and case sensitivities that needed correction. Once I got all of that straightened out - I finally had the initial pass at a "Recent Blog Posts" application.
<img src="/images/techass-recent-blog-posts.png" alt="Recent Blog Posts">

Rovani's Sandbox

Blog Posts - Page 4