Programming Tales

Into the unknown

2020/09/04

I have a confesion to make, I love watching "Making Of" documentaries, about movies, NASA launches, anything. I love seeing how these complex projects get done. I compare them with my own experience in very complex projects and, in the case of the NASA projects, live precariously through them.

Lately I have been obsesing over the "Frozen 2" making of in Disney+, "Into the Unknown". It is a great example of a "Making of" documentary. It has drama, failure, success, redemption and there are tons of technical aspects to boot. One aspect that really picked my interest is the similarities between the moviemakers processes and what we do in the software industry.

As the series start the moviemakers are about a year away from the movie premiere, what we would call GA or RTM in software terms. At this time you would think that the story would be well known and that everything is a matter of execution, after all they have been working on the movie by this point for about 4 years. But this is not the case, the story is not complete and they are still deciding what songs are to be included.

It is incredibly similar to a big software project. By the last milestone you would think that all of the features are well decided and that it is a matter of writting code but that is not the case. There are changes coming all of the time that need to be addresses, feedback from users, from VPs, etc... that needs to be incorporated into the final product.

Another thing that striked me as interesting is the fact that there is not way to update a movie once it is out. Just like software back in the 1990s, once a movie is out, it is out for good or bad. You might be able to do some minor changes, just as the movie Cats did, but that would be extremely rare.

When I first started working on Microsoft Office, in 1998, similarly we were extremely careful about what we were doing because we new that it was near impossible to distribute fixes to our customers. Only corporate customers could get fixes in a semi-regular manner. Normal users would have to wait for the "Service Packs", and those were quite difficult to come by anyway. There were ways to download them but the internet was still a nascent thing. Most of our users would only get the "Service Pack" if they bought Office after the SP patch came out, as they were included in the updated media that was distributed.

This sense of "finality" is what made us follow very strict processes. The final release of the product was called "Release to Manufacturing" or RTM because literally the final bits were sent to a factory to be burnt into CDs or (and I'm not kidding) floppy disks. It took quite a while to manufacture the CDs with the boxes and all that so if you missed the RTM date you probably have to wait until you can get another window with the manufacturer. In short that was not good.

The documentary also shows how the movie morphed and changed to the (almost) last possible minute. As they got feedback from fellow directors and the public the movie was changed to try to be the best it could. It is almost clear from the show that the filmakers run out of time before they had to lock down the story to be able to animate everything in time for the release date.

One of the most interesting episodes for me is episode 4, appropirately called "Big Changes". In this episode the filmakers have a secret preview of the movie with families and kids in San Diego so they can get feedback about the movie. This happens sometime in June 2019, which is about 5 months before the release date. This preview doesn't go very well so the filmakers know that they have to do major changes to the movie. They have a meeting the next morning but they ask the cameras to leave. This actions shows that the meeting was probably going to get nasty and shows the underlying tensions that happen during any kind of big stressfull project.

As a result of the meeting there were a lot of scenes that were simply cut, massive amounts of shots that had to redone and new scenes written that had to be storyboarded. This resulted in incredible amounts of work just thrown out the window.

I have seen the same thing in big software projects when entire features, sometimes entire apps, are cut very close to the end because they simply don't work, or because the priorities for the company have shifted since the start of the project. All of the work that was done up to that point in the project is now lost.

In the case of software this can have financial repercusions. People's work is used to judge their performance and thus their salaries and bonuses for the next year. Your entire careers can be affected if your feature is cut, at the end of the year you would have nothing to show for it. I am not sure if the same applies to the Disney animators, it seems that everybody benefits more or less the same by the movie coming out and having good ratings anyway. The animators in the episode didn't seem to faced by the news other than lamenting the work being one.

The other thing that really resonated with me was the "death march" that they had to endure in the last months of production. As the time was running out they needed to get everything done on time. This meant working 6 day weeks first then very quickly go into 7 day weeks until everything was done. This is clearly something else that is in common with big software projects.

In my time in Office I saw a lot of this in the last months leading to RTM. It was specially gruesom because it was all about fixing bugs, no new features and thus no creative work at all. In the case of the animators while they were doing brutal work at least it was creative work. Fixing bugs can get very monotonous and it always seems like never ending work because there are always more bugs to fix than there's time. The decission that you are "done" is always arbitrary, usually because of the RTM date.

By the end of the series you can see the effect that the process had in the final product. While I enjoyed the movie and the music it lacked some of the magic that the original had. Looking at how it was made it is clear that it was the best product given the circumstances. And the funny thing is that that the very same thing can be said about software.

Throught the process it seemed that the communication was excellent, everybody knew where they stood in the project and that they knew (to the percentage) the progress that they were making. The project management seemed excellent, something that is not always the case with software management. I think that there's something to be learnt from them. The main obstacle was not the organization it was themselves and the creative choices that they had to make to meet their deadlines.

In closing, I like to see these documentaries because I can see myself represented on the screen. I can see parallels on the kind of things they do every day with what I do, meetings, collaboration, decissions, writting code/animating, etc... They work on a project with lots of people that do very different things and everything connects at the end.