... most of the good programmers do programming not because they expect to get paid or get adulation by the public, but because it is fun to program.
Linus Torvalds

What a wonderful way to start off this blog and the new year.

What I am doing this week

  • Continuing to complete my Appeals database App at work
  • Complete the subsite of this blog
  • Snowball migration for a CMS boundary slip-up on my end
  • And a lot of life in a short time

Appeals Database & and my love for Go.

The appeals database has been a responsibility of mine ever since the DBA that created it dumped on my plate as he was getting more and more work. We were good friends at the time and I expressed my interest in databases. The appeals database was originally crafted in Microsoft Access and over the years I have learned a lot about access (probably more than I wanted to!).

Every year we make a brand new iteration of the database for the new appeals for the year. I eventually noticed that the DBA put in a very crappy authentication system using VBA. It was so easy to figure out the passwords because whenever you open the database you have full access to everything including the code used. By going to database tools and opening the VB code, you can see the login module and some if statements that contained the usernames and passwords for all the users of the system. I eventually removed that that feature and went with Access 2007 and beyond built-in database encryption.

We were pinged in an audit because that encryption system no matter how secure still needed to be a single password shared with a number of appeal's team members. Technically they said we were sharing credentials, but what am I supposed to do? Around this time I also helped set up an instance of 2012 SQL Server for some internal reporting. My next step was to integrate the appeal's backend access database into the new and better performing SQL Server. SQL Server allows for windows authentication integrations and permissions based on AD groups which made it an ideal candidate for this audit finding. Unfortunately, trying to get Access to nicely interact with this feature is like trying to do surgery on yourself.

After many, many months struggling to get the DSN working, and implementing a feature where I didn't need a DSN and the VBA code could just connect and dynamically add a link table for me I finally came to the conclusion that I wasn't having any fun. The project was stalled because working in Access was like pulling teeth. It wasn't until I started doing more and more web development, exploring NodeJS, Rails, Laravel that I got my inspiration back. The project really didn't get full steam until I learned Golang in the weeks before last Christmas. I can say I fell in love with its philosophy.

I delved into book after book, examples, and screencasts. I found out that the GitHub developers program I have included in my student account gives a 3-month trial for PluralSight. I have already done 3 or 4 course all on Go. I like it because it has the familiarities of a language like C, but also fantastically implements modern programming techniques you would find in Python or Ruby. My favorite fact is that it has is there are elements of Object Oriented Programming (OOP) without all the hassle of forcing OOP down the user's throat like you would find in Java or C♯.

Fast forward to the week of New Year, I have a functioning app that seems to be working in a relatively stable manor. It is still very alpha and by no means implements all the features I want to have in this database that will mirror the functionality of the current Appeals database. But for the first time in a long time I am actually enjoying the programming I do. No longer is it a hassle to get stuck on something and spend so much time encountering a problem finding a possible solution and having that solution kill the life force inside of me. I want to do mail-merge? I don't want to force my clients to go and learn the awful task of connecting to a data source like SQL Server, run queries, and get a dataset back they can use to merge into notices. Let's let the server do that work and avoid Microsoft office all together. I found a nifty package that does that heavy lifting for me called Gofpdf which is a port of the PHP version fpdf. Same exact docs, lots of examples, and even saw an article on Gopher Academy with a guy generating medical reports in the same way I wanted to generate Appeal notices. I tackled the problem and was able to get a notice to pull from the database super easy. I also found an easy to use package the performs LDAP queries for authentication with Windows Active Directory. Found a really well-maintained SQL Server driver I was able to connect to a Linux SQL Server 2017 instance I am running in Docker. I can now really do Windows development natively in my MacBook.

This blog is turning out longer than I anticipated. Let's quickly go over the accomplishments of this week.

  • Added a quick and dirty search function to find Appeals by ID
  • Refactored some logic for interacting with Appeals into its own model, view, and controller files.
  • Added a Count function to appeals model to return a total appeal count.
  • Implement a flash warning on the login page to tell users when errors are found.
  • Added more structed data to the Appeal struct for individual appeals. It is becoming a little unrulely.
  • Move the notice generation from the home page to the individual appeal pages.
  • Removed some junk templating from the home page to make it sleaker.
  • Updated the home, notice, appeal, and search templates to use the new structure for the appeals data I pull from the database.


Blog Site

I already had a webpage created (You can find it here) and wanted to use a separate page to host this blog site. The web page is using the static file webserver called Jekyll which is recommended by GitHub pages when you want to make a website hosted by GitHub. I have used it in the past with other websites I professionally made. Jekyll makes it easy to add a page to a site that isn't on the default page, but the way it was constructed it was heavily influenced as a single page app (SPA). I had to go into the templates and remove a lot of the embedded code to easily switch from one template and the other. The blog site and the main page are vastly different in comparison.

The problem is Jekyll doesn't let you make any root level layouts and just gives you one, the default layout. I extracted all that embedded code from the default layout and using the conditionals & variables provided by Jekyll I was able to load separate CSS and JS for the main page and the blog pages. It took a lot of research and stringed up a bunch of solution into what you are seeing now. Apparently, there is variable you can pass from the branch templates called custom_css. I was able to use that feature to tell the default layout template to load a separate set of CSS/JS for any website that wanted to use Twitters BootStrap. See this example here: 

    
{% if page.custom_css %}
    {% include blog_head.html %}
{% else %}
    {% include head.html %}
{% endif %}

That code would either load the blog header which includes its own description and CSS files if the custom_css: true was set in the page's Front Matter.

Jekyll also has fantastic documentation for building out a blog site. The whole framework is blog-aware. I just created the for loop to iterate over all my posts and display them on the blogs home page. I built out the template from a bootstrap example for a blog site (link here). and now this post should be published and you are reading it now. I think it has been a success.


AWS Snowball

Trying to expand the capabilities of our AWS instance, I tried whitelisting our elastic IPs against the Optum daily data extract (DDE) process one of my co-workers use for getting fresh data into the SQL server for reporting. While I had the ticket open we discovered AWS wasn't in the boundary we provided to CMS. Therefore, I have a lot of work to make our AWS instance CMS compliant. This meant that all the work I had done so far needs to be removed from AWS while this change request is implemented. I have already started moving the archived data my company needs to keep to AWS for disaster recovery purposes. That 4 TB of data needed to be brought back on-prem.

This week I finally received the external drive I am going to use to house that data temporarily until CMS approves us. I also finally received the order for the Snowball extract. I thought I purchased an hardware encrypted drive but it wasn't even though the order said AES 256 Encryption, but whatever I wasn't too bummed. I installed VeraCrypt which is a TrueCrypt fork and still being maintained today. I installed a virtual drive with 6 TB of space on it and then formatted it. That took a couple days using USB 3.0. I also moved my work desktop into the server room to connect to a gigabit connection so transfers between the snowball device and the encrypted hard drive was as fast as it could possibly go. Took 2 days to copy the 4 TB of data stored on the snowball. I sent it back this week so I don't incur any more costs.


Life

On top of workin 40 hours a week and attending school full time with a 16-credit work load. I spend most of my remaining time with my new family. I have a very beautiful wife and we are expecting our little girl at any moment. For the New Year, my wife celebrates New Year's more since she is an Orthodox Russian, in light of the Year of the dog, I got her a Miniature Schnauzer and we named him Teddy. It has been a hard week. I am in Boston for essentially 12 hours a day and dealing with a pregnant wife and a new puppy is not easy remotely. But I have to admit I am holding in there and making it work. If I appear sleepy and distracted, please know that it is not on purpose.

Also, it is my birthday on Friday and probably spend most of it at work or at school. So, if you are reading it then, Happy Birthday to me!


Things for next week

  • Understanding Go Object-Relational Mapping (GORM) and using it in the Appeals Database
  • redirect to a 404 page when an Appeal is not found. Or do some fancy warning message.

P.S.

Sorry for the long post, but I really need to provide some context for the current work I am doing now.

Bibliography

Singer, M. (2017, December 22) Building medical reports in Go - Gopher Academy Blog
  https://blog.gopheracademy.com/advent-2017/building-medical-reports-in-go/

Single-page application. (n.d.). In Wikipedia. Retrieved January 11, 2018, from
  https://en.wikipedia.org/wiki/Single-page_application