What have I learned from the participation in hackathon RailsRumble 2015

At the middle of November our small team have participated in the RailsRumble hackathon.

Despite we haven’t took any prize we were satisfied by the process itself and had some fun during it.

And after a certain time, I want to share my thoughts about how it was generally and what lessons I have learned from that participation.

Short intro about hackathons

For those who are not familiar with a hackathon format: it is a geek event where for a limited time you as individual or as part of a team suppose to provide the working prototype of your idea. It could be everything, from a web-application or mobile app to a handcrafted robot, just depends on the topic and the rules of the specific hackathon.

Various people, time pressure, continuous concentration on a project, general excitement - just a few of many other factors that make hackathons full of fun and challenges.

What differs the RailsRumble from the others hackathons is its restrictions list. Let me just leave an quote from their website here:

The Rails Rumble is a distributed programming competition where hundreds of teams of one to four people, from all over the world, have 48 hours to build an innovative web application, with Ruby on Rails, or another Rack-based Ruby web framework, along with other web technologies. Use Ruby for your entire stack, or just the backend.

After the 48 hours are up, everyone picks their favorites and the top ten winners are revealed!

I believe that it is enough for understanding format of the event.

How it was

It is good to notice, that actually it was the first online hackathon for us, and that it is slightly different than typical local one, when each participant team stays in one place all together for the whole time period.

We have decided to stay at our working office rather than at someone private place because of a such criteria:

  • it is empty on weekends, so nobody could disturb us
  • we could stay there for the whole night long
  • such facilities as high-speed Internet, whiteboard and space for everybody are already provided, so we no need to spend a time on such things
  • I have much powerful workstation here than at home :)

Because of time-zones difference we have decided to start not at exact beginning (2 AM our time) but at morning, near 6 AM and stay until the end, 44 hours later.

Some food was bought before, the rest we have ordered online after.

Next picture shows my unhealthy “starter kit” from the first day: Hackathon starter kit

For the whole period we have coded, talked, discussed things, spent time on the kitchen, and so on.

The luckiest are even had a time to sleep, but I personally haven’t came in that list and have slept only for 3 hours only x_x.

Long story short, here is the short time-lapse recap movie, hope it give some overview:



What have we done right

Here is short list of the steps done that helped us to successfully participate up to finish line:

Idea brainstorming

Okay, that is clear. If you have no idea before event start the chances that you would get one are pretty small. Believe me or not, but it take a lot of time to get an at least something that satisfies every member and at least worth to get hands dirty.

We have met a few times week before the actual start to share our thoughts and ideas with an aim to select the main one.

Preliminary planning

I think that it is the second step that you should do.

According to the rules you can not commit any code until the event starts. But you could and should plan how you would implement your idea, who would be responsible for which part project and so on.

Of course you no need to plan every minor point, but even generic map of required actions, technological stack chosen and responsibilities would affect your participation in a positive way. You would save a lot of time by this.

We have used Trello for it.

Disturbing things and facilities

As it was mentioned before, you don’t need to spend your limited time on things that are not project related. So it always better to think about it before.

Take a rest

It always good to be prepared not only technically but physically too.

A few hours late might not be a big deal if you start full of energy in contrast the situation when you start earlier but feel sleepy from the beginning.


The sad story that it’s the end of the list already. So maybe it’s the actual reason of our fail. :)

Here is the opposite one.

What have we done wrong

The list of fails and notices to have in mind for a next time.

Check the feasibility of the main idea

The biggest fail that have caught us from the beginning.

Our idea have looked achievable, some useful libraries were found but we haven’t checked the dependency on 3rd-party services. Yay!

At the beginning we have thought about internal usage of the various search engines such as Google, Bing, Yandex and so on.

But the problem we haven’t checked have related to the limits of its usage.

And all of them are pretty small!

By small I mean the numbers like 1000 requests per month and less. And that is not enough for testing, active development and future usage by other people at least for time of judging.

We have spent some time trying to find an easy ways to overcome those limits but without success.

Of course there are some methods, like switching between precreated accounts, proxy usage and so on, but we have decided that it would just wrong to base the main functionality trying to overplay a big corporations.

Because of this we have switched to another, much more simplified idea, just to not leave the distance of the hackathon.

Moral: Check it twice before start or at least have a good plan B.

Environment setup

The same story every time.. We have discussed the technological stack and libraries versions, but doesn’t prepared setup on local machines.

So at least an hour was spent only on basic installations and problem solving for three different computers.

It would be beneficial to have an environment ready for everybody.

I prefer [Vagrant] for this, so next time we definitely should start just by running vagrant up on the precreated Vagrantfile.

Reimplementing the wheel

For a custom project it might be applicable to create some specific solutions for common things. But it definitely not worth to do it when you limited by time.

Simplify as much as possible. Try to reuse existent libraries and solutions.

If it’s Rails then use an existent gems, if it’s Javascript - jquery-plugins, npm packages or any other framework libraries, dependent on what you are using.

Don’t spend a time on writing something from scratch until you really need to. It error prone and time consuming. And generally useless because almost always you will never use same codebase after even if you will decide to support a project after the competition.

Share roles and do what you are better in

My recent tasks at work are related to backends. But almost 70% of hackathon time I have spent on frontend, just because nobody else could/want to do it.

Of course sometimes it is fun to learn new/refresh forgotten skills, but if you have a chance - try to find a proper person in your team who will fill these gaps better. It all about balanced team.

I believe that if we have a frontend developer (our team have consisted on two backends and one designer) we would achieve and got done dramatically more.

At the same time I have learned the different types of web-animations, specially the CSS and Javascript based ones.

Don’t know where to use that skill though :J

Think about UI & design

The practice shows that even not rocket-science project could be warmly welcomed by audience if it looks good and haven’t leave the feeling of incompleteness. But even high-tech project with raw blank page layout and blue links on it hasn’t so many chances to be even tested by people, just because it seems broken in such case.

It’s sad but true. From over 250 projects have participated in RailsRumble, for my opinion near a half looked broken, and some of them actually was.

Usually at local hackathons only near 30 projects presented at final stage. But online format allows much wider amount of participants.

So you can’t get positive feedback if people even can’t try claimed functionality. Or if it looks so bad that they just wouldn’t to.

Concentrate on the main things

All the cool features could be added at the next iteration if you have a time.

Just take you main idea, make it work and be satisfied with it. Only when you already have something to show check what could be improved or added.

Don’t try to do everything at once, there is a big chance that you will finish with nothing.

The working prototype is better that a none-working full-feature project.

Spend time on testing

On the hackathon there is big chance that you will skip all the tests :) But you should check at least manually that main flow is working, there is no obvious bugs and so on.

We have spent a half an hour together checking a project before final countdown. During that time several bugs were found, but despite that we have fixed them, at least the one actually come to production and was found after code the freeze.

Luckily nobody have used it and we have had a time to provide a hack next day. (As commits was forbidden and [Heroku] boxes are read-only, the hack was based on a cron job that periodically makes some changes on remote server using its console add-on.)

Restrict access but leave the guest login and respect people privacy

This point depends on the type of your application. In some cases if you don’t want your application to be broken in a fist minutes of public access try to restrict it to logged in users or hide the dangerous endpoints at all.

Here by broken I mean not only totally crashed application but also the empty one, in cases when you need some precreated data being shown to users, but unfortunately being deleted by someone.

At the same time registration should be easy (you can just prepare guest account) and don’t ask more permissions than you really need to. I have seen a lot of projects that requires full access to read & write on Github, including private repositories. It is clear that only people with an empty accounts would try such service during the judgment period.

Thats all!

Additionally, you can find a few more worthy tips specially on how to improve the score judges will give your app in this blog-post by Jamie Dyer, who was both a competitor and an expert judge in the Rails Rumble some years ago. I have found it useful, but sadly only after participation :(

Conclusion

Each hackathon is different. Some looks like a competition, some aren’t. Despite of its format it’s up to you to find a reason for participation. And based on that you might not take the notices above to seriously.

In any case you would have a nice time with a various people on a same mental wave with you, discover new technologies, try yourself in a new roles, share skills, listen for interesting mentors and much more..

Networking, pizza and beer (juice) could be also the reason to participate. :)

So you will never regret despite the result. See you at hackathons! Say Hi!

P.S. Our small laggy project made for fun is still online and could be found here.

The idea is pretty simple: create a quiz, each question in which is a fighting round. Fighting generated on server and even it looks a bit similar (because of small amount of .png-sprites prepared) is different each time. The animations done by css-transitions and classes that managed by the Javascript. All the rest is the basic Rails application.

Cover image attribution: TwilioCon Hackathon.