As someone whose only experience is in corporate service IT development, from the outside the game dev industry looks incredibly cowboy-ish; it would be nice to read a story where a team built a game where they met their deadlines, had no crunch times, followed excellent coding standards (CI, unit tests, peer dev etc.) and maybe even used a modern language (perhaps, heretically, a non-native language like C#!)
- There are tons of positive "post mortem" [http://www.gamasutra.com/features/postmortem/] write ups on gamasutra, but the ones with problems are usually more valuable. One recent example of "doing it right" I remember was the XCOM:Enemy Unknown remake. Check their GDC talks!
- One thing that makes games special is pushing the envelope. Yes, you can make a game on a very solid stack. But it won't stand out.
- Game coders think service people are mad also: How can you use a database that you do not fully understand? You don't even have the source? Using a framework or library or cloud that you do not 100% understand seems "cowboy" to them.
- Games are "fire and forget". That means technical debt is very different. It is bad at the start of a 3 year cycle, but at the end it is totally ok.
- Not mentioned by you but important to point out: While working on games is great for learning and unlike anything else it is also a terrible industry for coders. Insane crunch time. Low pay.
- Games are "fire and forget". That means technical debt
is very different. It is bad at the start of a 3 year
cycle, but at the end it is totally ok.
Clash of Clans was released 2.5 years ago and will likely continue to get active support for many more years. FarmVille was released almost 5 years ago. WoW more than 10, for a different kind of game that also isn't fire and forget.
They get a lot of flack for other reasons but engineering wise companies like zynga get a lot of stuff right.
Personally as an engineer I like the coding approach of game dev a lot more.
I know what you're saying, but I wonder how true that is, and my analogy is the movie and music industries. You can make a very successful movie or album without pushing any technical envelopes, infact there are people who consciously and deliberately use "obsolete" tools and techniques.
Not to mention widely used on-premise enterprise software with weird upgrade-paths and hidden costs. If you look behind the curtain there is a lot of legacy cruft in COTS. Customizing such installations seems cowboy-ish to others. SaaS/cloud is better and hide ugly details but may off limits for high confidential data.
Game programmers often reinvent the wheel but learn from others through post mortems (former Game Developer Magazin). Enterprise devs often cook their own customized solution based on COTS. You can still install and play almost all PC games on modern hardware with Win 7+ 64-bit. Try that with old SharePoint, Exchange, SAP, Oracle/Sun, Novell, etc. software and not to mention the troubles to restore an old backup.
As someone who saw both camps, I would say both sides can learn from the other side.
Is that really true? When a team uses an engine licensed from another company, do they really take the time to fully understand its source?
Game companies don't "copy and paste" old versions of games to jump start the development on new versions of games? cp doom1.c doom2.c
We often crank out several great games a year on time and under budget. There's a lot of cool stuff going on in this industry right now.
There are still a lot of fucked up companies but you generally just don't hear about the development of >99% of projects.