Tuesday, March 15, 2016

Using JIRA for Movie Making

Tackling the unknown known.

Tomorrow marks the 3 year anniversary since I started filming my movie Synchronicity. (see post from 3 years ago). While the filming only lasted 7 days, the footage has sat in editing for over 2 years. It's been an epic hobby, one that I tend to in occasional bursts, but starting this week my leisurely pace is about to ramp up into full blown efficiency.

Part of what makes a project drag on is the difficulty of quantifying unknowns. While many of the tools involved with edit are things I've never done before, the overall project is divided into sections (number of scenes, number of lines of dialog, number of shots with special effects, etc.). The reality being, once the initial filming was completed, there weren't that many unknowns left in the project. All this to say, I have no excuse not to pin a deadline on the calendar and get my butt into gear.

Of all the tools I embraced for film making, I never thought JIRA would be one of them. JIRA is a time-tracking tool designed for companies making software, to facilitate the industry standard practice of 'agile software development'. It's a tool I've been using every day at work for the past 3 years, and I've come to find it as valuable for business collaboration as Word or Excel. And, as my engineering role has become more and more focused on project management, I eat, sleep, and breath JIRA.

Sprint 1. In general, work is grouped into JIRA 'epics' according to the tools/expertise involved.


I can already hear you laughing. Why would a project so rooted in creative flow use a tool as cold and dry as JIRA? The longer I stared at footage in Adobe Premiere watching replay after replay, I realized a scary amount of commonality between movie cuts and software releases.

My plan is to regularly cut fully length versions of the movie until the release date; v0.1, v0.2, etc. until the big day! The movie premiere! Synchronicity 1.0! Where as the similarity brakes down when you consider that software has the wiggle room to patch future releases; Movie 1.0 will remain the same for all time. (Though there is the occasional director's cut: Blade Runner 5.0... but I digress)...

It's never gonna be perfect

How many software releases have been 100% feature complete? ZERO. A thousand good ideas reduces to a hundred good ideas and further down to a dozen core features (without which the product would be dead on arrival). Imagine software designers and architects as equivalent to a movie's director or visionary. These people are talented and creative enough that given 5 years they could make the most beautiful masterpiece ever.

Nobody's got 5 years. We ain't all James Cameron. Like it or not this titanic is gonna have a little rust on it. I could spend an eternity meticulously crafting every frame, but at some point someone's gotta see it. It's learning to compromise. No. It's being forced to compromise. Because without that force it's just "one more fix here", "ooo, and another small fix here" and "maybe if we rearrange this is could turn out better"...

Imagine your movie producer: The film is scheduled to hit theaters! We need to see progress on the film! Does it work as an overall story? Until then, ignore the bells and whistles!

James Mangold, the director of 3:10 to Yuma (2007) had planned to build an entire wild-west village for his flick. As they ran up against their budget, set building had to cease, and some houses were left exposed as little more than a collection of 2x4s. But at the end of it all he managed to dazzle the audience with a several minute chase sequence from one end of the village to the other. He achieved all the emotion and excitement needed for the climax of the film, despite Russel Crow and Christian Bale running through half built sets. Two extra houses wouldn't have made it a better experience.

Imagine your software CEO: We gotta ship it! We need to do regular releases! Do the core use-cases work? Until then, ignore the bells and whistles!

Bill Gates, the CEO of yadda, yadda, yadda... insert your own inspiration reference to how DOS 1.0 shipped as a piece of crap. :P

Reality is a B****

Before putting my remaining editing work into JIRA, I made an intuitive guess that I had 8 weeks of work left. Afterwords I spent several days cataloging all my desired fixes into JIRA (everything from single frame tweaks to major plot point reworks). Using my existing editing experience I was able to breakdown and justify each task into detail. 

Example details for a ticket. A specific set of high level requirements, a proposed implementation, and an itemized list of time.
When the dust settled I had created 135 tickets and 30 weeks worth of work (with my measure of 12 hours per week). This is the reality check. I'm dreaming too big. Gotta get back to the core.

The next step will be the hardest. To pick a calendar date sometime this summer to premiere my movie. A date that will likely end up 10 weeks earlier than my proposed work. I'll commit to it by organizing the premiere event with all that my friends involved in the film. I'll be forced to drop a 1/3 of my originally planned edits, all with a workable final produce in mind. Priorities will trickle down quickly.

Another great exercise in intuition vs. measuring. Keep an eye out for the burndown charts. :)

No comments:

Post a Comment