I’ve been hard at work mapping out the future of Code Journal and took some time to go through user e-mails that requested new features; the idea was that I would simply implement the feature that the most users had e-mailed asking for. That feature was a resizable UI.
The problem was that the people, my customers who I of course care a great deal about, were very passionate about this particular feature and the feature seemed simple enough to implement; after all, I’ve implemented resizable UI’s on a number of Mac OS X apps for clients, so why couldn’t I apply the same the techniques to Code Journal? Right? Wrong.
Code Journal started out pretty simple; it was an app that pulled and processed JSON data from the Github API; in fact, the most complicated thing in the private demo (think alpha) version of the app was Github’s OAuth. I was thrilled. The app was nice and simple.
Overtime and as it closer to the all important 1.0, compromises had to be made. There is no need to enumerate them here, but suffice it say that one current feature (Github Enterprise support) is responsible for a disproportionate amount of code and almost all of my support requests. Does that mean I’ll be pulling enterprise support? Absolutely not. The truth is that these request have more to do with customers having non-standard settings on their Github Enterprise deployment or the server that host them, rather than anything to do with Code Journal. Enterprise support is a slam dunk and I personally love being able to user Cde Journal with clients who have Github Enterprise servers; best part is that most of the time all I have to do is fill out the enterprise field in the app with my credentials and we are good to go — no need to bother a sys-admin at all.
In short, the complexity that supporting enterprise customers caused was a good thing. Back to the feature request: resizable interfaces. I understand why some people might want this feature, but, to be honest, I don’t think it’s a good idea. Even on my 27 inch Cinema Display, I often find myself pressed for screen real estate when tacking some hairy memory management problem or am “in the zone.” Like many of you, I am not willing to compromise by putting my monitoring tools into a separate space and, again like most of you I imagine, I only have the one screen. The other issue of course is that Code Journal is a Mac app and a lot developers that work on Macs work on 13” laptops.
Still, it’s hard to ignore your paying customers, so I went against my instincts and began coding. I won’t bore you with the details, but it was doomed from the jump. I spent two weeks coding and agonizing over resizing table views and experimenting with Auto Layout. What resulted wasn’t terrible but had a number of rough edges and, more importantly, was not a feature that I could see myself using and believe would have been a major source of maintenance. The feature is not going to be released. At this point in the product’s life, it is more important to keep complexity out of the equation than to add every suggested feature; this particularly true of features that don’t provide any additional functionality. The best feature is the one you never have to maintain.
Comments? Questions? Share them with me on Google+ or Twitter. This post was made possible by Code Journal and Fingertip Tech, INC. If you are Github user please check out Code Journal and if you are interested in having an Android, iOS, or web app developed please contact me. Also, check out the new free version of Code Journal for iOS.