April 27, 2018

On Managing a Group Project

  1. Stay Silent
  2. Do Everything Yourself
  3. Work Only When You Have To

The semester ended, and I just finished a reasonably large group project on Virtual Reality (link here; teaser here). Let's talk about what I learned.

Stay Silent

Group projects require good communication with members to succeed. One of our difficulties was getting everyone on a common platform to make communication less of an issue. One member did not own a Facebook account, our defacto method of communication, so we couldn't just create a new chat group. He did, however, prefer texting, which would in theory make communicating much more instantaneous, but also make it complicated for us when we use a computer (very few messaging apps also have a computer-friendly interface). Thus, I opened a Slack group. This would've been great, except that everyone was new to Slack.

One thing about effective communication is that the medium must always be convenient. The members should all be very familiar with the intricacies of whatever medium you use to communicate, or at least, have been using it for a while such that it becomes a natural habit to check (see email). Since none of us have used Slack before, we wouldn't check it often, which kills the platform. I tried to get around this by asking everyone to pin a Slack tab on their browser, but to no avail. The group still suffered from painfully slow reply times, often spanning hours to days. Later in the project, I decided to scrap Slack and just message each member individually. This worked slightly better, with less-than-an-hour reply time. The downside is that it promotes information asymmetry.

Sometimes, even though the medium used to communicate is new to the members, everything still works out. Our group used Trello to organize the project, which was new to us. We would assign parts of the project to others; everything would be submitted via Trello. Trello is an online pin board, so members don't have to check it too often. It also, by default, emails you about cards you are assigned, so there is often no escape.

Projects need a constant stream of communication. A haphazard update post just wouldn't do. Every member needs to be aware of the direction the project is going in, what everyone is actively working on, and how everyone is contributing to the final product. If you have to ask the question: "Did I message/update/notify them enough?", then you may have to reconsider your communication system.

Do Everything Yourself

Group projects should be done as a group. Or at the very least, the entire group should be fairly involved in it's processes (see above on communication). Nobody likes to be the one who has to do a disproportionately large/small amount of work, especially if they don't know what they are contributing towards. One of our issues include resource management and planning.

Our group had almost zero project experience, and only around a year of programming experience. They were very quick to learn the new things taught in this class, but only because they had to in order to survive. Out of the group, I was the one who had the most of both, if only in theory. I helped come up with the idea of the game, which meant that I created content for the game itself. The only thing I could delegate as a task was the model and animation creation, which I happily assigned to my group.

I cannot fault my group - they were going through school too. I couldn't expect them to put many hours into a project that they (I think) cared less about (they were neutral about it - not too passionate) than me (who's enthusiasm fluctuated violently).

This left me with writing all the content (dialog, storyline, plot), programming the game, implementing mechanics (1 of which was scrapped), and fiddling with other bits and pieces that nobody else did.

I would blame only myself on this, since I based the entire idea of the project on my own skill sets - what I was capable of doing. I had a general idea of how the project was going to be finished while I had the project idea in my head. I heard that good project managers must have a vision for the project, and this is no exception. I had to figure out how the entire project would go down before almost any work went into it, and have confidence knowing that everyone would be able to finish it in the allocated time. Let me tell you: It's hard to see into the future.

Work Only When You Have To

Don't procrastinate. Have a good, solid, fixed timeline so that your team can follow it and be on schedule... is what I would do if I were smarter, more motivated and less distracted. Unfortunately for me and my group, I wasn't. For the first month, I basically put the entire project on the back-burner and did other things. This isn't really a project-specific idea, but a life tip in general: don't procrastinate; don't wait till the last moment. The way our teacher got to know our progress was through stages - we submitted reports, files and models every few weeks for our professor to gauge our progress; those counted for marks, and were basically assignments. The week before they were due, we would just work our asses off at school (or perhaps, I would). There were a couple of instances (last 2 months) where I practically lived in the school computer lab (more than I already did - I normally would never leave that place unless I needed to go to classes). I took classes at 8 AM, and left when security kicks us all out, at 9 PM, a full 13 hours of work time, or 15 hours away from home.

line graph on week number vs hours worked

The above graph shows the number of hours I worked every week. This project was big enough that I decided to start doing time tracking, just for fun (the spreadsheet took me 2 hours to tweak). You can see the days where I do next to nothing. Also notice the spike due to a "check-up" assignment. In total, I worked 85 hours, or 2 weeks, full time.

pie chart on different categories I've worked

I used most of my time on the programming portion, where I tried to implement 2 game mechanics - the dialog system (reads a specially formatted json file that I designed) and object interaction system (scrapped; originally highlighted the object you were looking at with a yellow glow, but with the side effect of removing the original texture on the object, rendering it useless). The rest of the time was spent on testing and bug fixing my code.

I needed to delegate more of game programming to my group. But here's the thing: Unity is such a git-hostile environment that, even if my group was experienced in git, we would still be having trouble with it. I need to learn to assign tasks to the group better, and learn to trust them more.

Tags: group project unity programming