Andy's Fun House


Weekend Projects Gone Awry, Part 3

Things will be a little different this time. With the two previous entries in this series, I was recapping work on a new idea/project and had to relay information about the project. This time, however, I’m simply extending Part 1 into the mobile arena and trying out some “new” technologies. That disclaimer done…here we go!


A Recap: SlimSocial

SlimSocial is the fledgling social network devoted to making “being social” more difficult. The key limiter was that a person can only post once per day, in theory cutting down on the useless chatter that happens constantly.


Going Mobile

After pushing out SlimSocial, several of my good (aka misguided) friends decided to latch onto the idea and enjoyed the concept. It didn’t take very long for the first “can you make it work on my phone” request to come through, and after the fourth or fifth in a row I begrudgingly admitted that I could probably look into making a set of apps. I made sure to stress to each person that it was a lot of work and unlikely to be something I’d complete for months.

So of course, I started on it a couple days later when I’d finished working on Part 2.


Executing (aka Enter Cordova & AngularJS)

Last year at Microsoft’s //Build conference, they made a pretty big deal about their new cross-platform features in Visual Studio. I’d avoided it, as my past work with cross-platform mobile development via Marmalade proved to be tiresome and less than smooth, but I decided I had an opportunity to give the whole package a shot.

One of the other things I’ve been avoiding has been the concept of SPA (Single Page Application) frameworks (Angular, React, Riot, etc). Generally speaking, I find the whole idea to be a quick way to ask for trouble with performance, and they just seemed like trouble to learn.

So there it was. I felt I had the opportunity to learn two new things, and I took that opportunity. After finishing the game engine for Part 2, I continued streaming on Twitch while I fumbled my way through learning both AngularJS and Cordova at the same time within Visual Studio 2015. In the end, I’d say I managed to dump an obscene amount of my free time into this extension of SlimSocial, but on account of forgetting to track my time we’ll just put the estimate around 80 hours.

If you’re reading this and are expecting me to begin expounding the merits of both Cordova and Angular, you should probably stop reading now.


Tech Notes

Now…I can be pretty opinionated. I’m aware it may be a bit of a character flaw, but it’s part of me so you’ll either learn to ignore it or you’ll go away. Learning two new technologies at the same time goes against some hard-learned lessons for me, but it’s not as though I intend for SlimSocial to ever be a serious time-sink (I understand that it has become one, but we’re talking intentions here).

Cordova seems like a pretty promising bit of technology. It’s not as though it’s a new piece of technology, I remember PhoneGap and others being around nearly 4-5 years ago, but it’s not the sort of thing I like to use on account of much of my mobile work having very specific performance requirements that can’t be achieved by wrapping an app within an app (which is basically how Cordova works).

That said, it’s not perfect. Some of the workflow can be aggravating, and figuring out how to configure things on the different machines, while not bad, is far from perfect. Even so, I imagine whenever I have projects in the future that don’t need me to squeeze every ounce of clock speed I can out of a phone I’ll be turning again to Cordova and its VS 2015 integrations.

And then there’s Angular… I can’t tell you how many times I’ve been told I need to learn Angular/React/Riot/Backbone/Knockout/Whatever. It’s borderline ridiculous how many frameworks there are that achieve more or less the same thing. Angular’s learning curve is pretty fascinating as well, starting off shallow when you look at the tutorials, and then taking a curve so steep that it feels similar to running into a wall head-first.

Angular calls itself “Superheroic” on their website, which is a pretty vague thing. Does it work? Sure. Does it do so in a way that’s easy to grep? Not at all. You have to learn intricacies of the entire framework in order to do even marginally complex actions for your app. You also need to grapple with incomplete and/or insufficient documentation when trying to learn said intricacies, basically relying on Bing or Google to show you the dark corners of StackOverflow where your answers lie.

I’ll try out Angular and one of the other frameworks for another project or two, just to be sure I’m giving it a fair shot, but thus far my decision to stay away from these libraries is feeling pretty justified.


Conclusion & Reveal

Once again with SlimSocial, I had hoped to make my prototype over the course of a weekend. This estimate was built off of my experience with doing prototype mobile development natively, and thanks to the learning curves of both Cordova and Angular I was about as far from correct as could possibly be (which is the reason I have a personal rule to not learn new technologies for business critical projects).

Adding onto this, there is a problem with distribution. In order for me to properly make this app available to individuals who aren’t interested in testing, I’ll be shelling out some serious dough to Apple and Microsoft (and subjecting myself to the chaos that is the Android build process). I’m not sure how willing I am to do that for SlimSocial. However, when the day comes that I do renew my Apple and Microsoft developer accounts, I’ll probably publish this immediately. Until then, enjoy some screenshots of the app working on my Windows 10 desktop.

home feed
friends profile


- Andy