Hack Day, Summer 2018

MicrosoftTeams-image

This week I participated in one of the first combined Hack Days at my workplace, which brought together IT teams from across campus that normally don’t interact on a daily basis. I’ve always loved the concept of trusting people and letting them work on whatever they find most valuable for a period of time (it is baked into my team’s sprint schedule three times a year), so I wanted to support the event and encourage my peers to do so in the future.

I’ve been acting as a Scrum Product Owner for the past year and am in the process of transitioning to a team wide ScrumMaster for the foreseeable future, so my hack day projects were centered around building two small tools to assist in the calculation of business value and effort estimation (exciting stuff!):

  • Protocalc: Not the best name I’ve ever come up with, but it works! This simple tool sucks in a JSON object of categories, questions, and answers (and weights to those categories and questions) and dynamically generates a web form out of them. Filling out this form will give you a business value estimate for any item or feature, both out of 100 and out of 20 (depending on how you’d rather round). It’s a simple way to share an understanding of what drives value on our team and removes a lot of the overhead and ambiguity in calculating some usable value indicator. The question and answer set can be easily swapped out for different teams and different value streams making it reusable.
  • Plokker: Most Agile teams regularly estimate the relative effort of items on the backlog as this scoring is useful for determining how much a team can and should commit to in sprints. A pretty popular way to do that is through “planning poker” in which team members determine their own estimates and then try to reach a consensus.

Don’t judge the JavaScript too harshly (it was a hack DAY), but it was really reinvigorating to sit down and just create with no real risk in doing so. I’m still feeling the positive effects from the experience days later. If these small tools have any real impact in our day-to-day work I’ll be sure to report back!

FSL Database / Website

“Sorry I had to run out to meet with another student. The website you built looks amazing and is very user friendly. Thanks so much for working with Cristina and IKM to move this process forward. Much appreciated. Belinda” (Assistant Vice President)

Nice to hear after a good demo!

EDIT: Another!

“Good day all, I must co-sign on Mrs. Belinda’s words. Jordan has been awesome
to work with and has brought to reality a system that will make our productivity more efficient with less human error. Those extended phone calls have been worth it, Jordan =) I am very happy we made it here together! Thank you.”

Campus Webcams (C#)

A simple site that reconstitutes single JPEG images from within the UCF firewall for outside consumption via Motion-JPEG

Purpose

We got an urgent request to recreate UCF’s Campus Webcams site, as UCF Marketing was replacing the main UCF site with a WordPress install (thus removing all of the various sub sites of that legacy implementation). We had no contact with the original developers nor any glimpse of the code that powered the previous Webcams site. In addition to reproducing the original site, it was requested that the new site also double the number of cameras viewable.

Method

I got to work quickly on an .NET C# MVC3 application, building out a default Area for the public and an admin Area for a front-end to a database that stored webcam information. I gathered data manuals on all the various camera manufacturers represented throughout our departments and discovered ways to pull single-frame data at specific resolutions from each. I wrote an Action that composes constant snapshot data into a Motion-JPG stream. Finally, after some research, I pulled down a Nuget package for an image processor to do Gaussian blurs on the images (for legal reasons).

Highlights

There is nothing quite as bizarrely satisfying as writing the raw response output of a Motion-JPG stream. It’s too bad that no version of Internet Explorer (through 10) supports Motion-JPG; any IE visitors get redirected to the raw Frame output (and a basic JavaScript refresh every eight or so seconds). It was also fairly satisfying to get into the groove of making all of my applications responsive for browsers and devices smaller than our minimum window size. Finally, just because I could, it was a blast to develop a KML feed that could display the cameras on Google Maps.

SDES Directory (C#)

  • Type: Website
  • Language: C#
  • Framework: MVC 3 and Entity Framework 4
  • Authentication: None
  • Data Source: SQL Database

Purpose

In collaboration with UCF Marketing, I developed a JSON model to store basic directory information on all our division’s departments and offices, including hours of operation, location, phone, fax, etc. JSON, however, was not designed to be stored in a flat file, so I got clearance to recreate the model as a SQL database and .NET MVC application.

Method

I designed a SQL schema to match the existing JSON model and migrated the data to the database. I then developed an MVC3 application (with a default and admin Area) to allow edits to the data, allow the public to view the data in an ordered fashion (and submit changes via an email form), and allow any public developer to read a generated JSON feed of the data.

Highlights

The app itself is pretty simple, but the definite takeaway from this application was the implementation of OutputCache on the controllers. Knowing that the JSON would be consumed by UCF Marketing and most of our sites, I was able to easily implement an application-level cache of the feed so that it loaded instantly and structured updates to the data at a set interval.

Takeaway

Small, good ideas work! Sometimes a full-on application seems like overkill for a simple idea, but if that simple idea saves hours a week, then it’s clearly worth it.

SDES Directory (C#)

SDES Template Engine

  • Type: Class
  • Language: PHP

Purpose

In our division, there are various entities we consider customers. Most of those customers cede support of their web environment to us, but some retain the right to have a full-time “developer” on staff to make edits. Unfortunately, there is crossover between sites we’ve developed and sites that these developers feel belong to them. This has been a problem for years; only now am I taking action programmatically. I got clearance to develop a simple PHP class/framework/engine to structure the basic data each of our sites require (title, subtitle, navigation, social networks, etc.).

Method

I wrote everything by hand, top to bottom, imagining our basic PHP sites as objects. These objects needed many small pieces, like the title, the associated department, and the contact phone number, but I considered larger items, like database connections, the page requested, and the layout/design of the site. I wrapped most of this data in a straightforward class that accepted changes to its properties exclusively via setter methods, exposed data exclusively via getter methods, and farmed settings out to a config.ini file.

Highlights

As every site was converted to this simple engine, it gave me the ability to wrap simple HTML blocks into helper methods. Now, every site has a standard way to print contact information, social networking URIs and buttons, footer information, and so on. It has been an easy bridge between design and development to save time and preserve a standard.

It also gave me a chance to develop a database class that preserves a single database connection via the singleton method (only allowing a single instance of the database object).

Takeaway

It feels like I’m creeping a little too far into custom-PHP land; I know we should be using something like Symphony or CodeIgniter and a templating framework like Twig or Smarty, but for this environment, where access to production is treated like a right and not something to be protected and given out sparingly, I needed something a bit more customized and esoteric. But rest assured this custom implementation will be trashed and swapped for some open-source framework as soon as our environment makes sense.