Trích dẫn:
There are going to be a few blogs this week (depending on whether some of them get permission to be posted or not!), although they aren’t new feature blogs as such. These will start mid next week.
This weeks blogs will be based around some behind the scenes things at SI.
This blog will focus on how we decided upon new features for FM2010, and the process that we go through. I’ve been asked a few times on the forums, and in interviews, what process we go through and how we’ve changed the process for this year.
As I mentioned in a previous blog, we used to store feature ideas and bugs all in one database, which was inefficient, as it meant certain things didn’t get looked at, so we split our database in two, and now have one for feature suggestions, and one for bugs.
We also used to be quite haphazard about what feature suggestions went into the database. Our feature meetings were split between looking at suggestions in the database, suggestions on our internal wiki, emails sent to us, and ideas that came to us in the feature meetings.
The meetings were also a bit odd, as we used to have one meeting that went on for a couple of days that everyone would attend.
What we did this year was split the meetings into the same modules that the game is split into – match, media, database, misc (finance, database and other stuff), GUI and platform (the platform layer is the code you won’t ever notice, but without it the game plain won’t work!).
The features database was split in the same way so that we knew what to discuss in which meetings. Any new feature requests had to be added into the main database (and still are) and assigned to me to have a look at before I approved or denied it to go forward to the features databases – the reason for this stage of the process is that some ideas just aren’t possible legally, or are not appropriate for what we’re trying to do with the game, so it makes sense to get rid of these suggestions before the discussion stage as there’s no point talking about something that isn’t going to be possible.
Each feature meeting in each area took at least 2 days. Some we had to re-convene after other meetings if there was a cross over on areas. Every feature suggestion in the features database was discussed – all 800 of them!
Some of these things are not what you might call a feature. Changing the text on a button in the game wouldn’t be classed as many as a feature, but we have to have everything in the database separately, so things like that do count. It’s one of the reasons why we don’t have the normal “this is how many new features are in the game” in the initial press release as we would have had in previous years if I was to give a raw number from the features database, we’d be looking at well over 200, many of which would not have counted in previous years.
Some of these features might take an hour to do. Others might take months, or in some cases, years. Others have to be done in phases, so some of it makes it into this years game, and we already know how to expand it in future years. But by having it in one database, it makes it a lot easier to track these things, and make sure nothing important is missed.
The meetings themselves were also done differently this year too. We switched to a voting system where everyone who attended a meeting (and the meetings were open, so anyone in the studio could attend a meeting even if it wasn’t in their own area) had an equal vote out of 5, and could also play one “joker” per area of the game, which would double their vote. This lead to some features having a 150% score, but that just meant the features were very high priority. It also lead to some features having less than a 50% score, which means they are unlikely to ever be done, as there will always be things that we deem more important, although each of these features will be re-visited each year when we have the meetings.
After all of these meetings, the production team at SI sat down with the senior coders in each area to get expected timelines for coding the feature, and once that was completed, features were assigned to the different members of the dev team based on priority (the score the feature got) and the time on their schedules.
As you can imagine, we had far too many features we wanted to add into the game that we had time for, so this scoring system was essential. It means that the highest priority features could get done, but also that if there was time left over, others could do too. Those that didn’t make it this year will likely make it for FM2011, and certainly some things that we really wanted to do this year just weren’t possible timeline wise, alongside some new ideas.
As for new suggestions, throughout the year, more feature suggestions are entered into the database, from the forums, our research team, test team, internal ideas and external ideas. Some of these are “fast-tracked” and get discussed immediately, particularly if it’s something small or something that is directly related to one of the new features already in the game. Others will be there ready to discuss in the FM2011 feature meetings and, with the amount of ideas currently in the database (well over 1,200), likely the FM2012, FM2013, FM2014 and FM2015 meetings too!
