I forgot one more point. Someone here said something lauding BioWare because in some interview or video they mentioned they had spent 8 years worth of man-hours on developing Frostbite for this game. We know the game was 4 years in development. Assume a minimum of 40 hours per week per employee and in 4 years that means that 8.4 employees developed the game on Frostbite with an allocation of 8 years worth of man hours.
<snip>
I believe the quote was that Bioware spent 8 man years worth of work developing Frostbite for RPG. The definition of a man year is one individual working exclusively on that component for one year full time. Presumably it did not take 4 years for them to finish that aspect of development (or maybe it did?) hence I would suspect a decent size crew was allocated to engine development alone. By the logic applied above, 2 individuals working exclusively on engine development over 4 years = 8 man years of work.
However I do suspect that design and testing of said engine is also factored into the 8 man years claim. No matter how you slice it though, the interesting part is that it did in fact take an additional 8 man years of effort to make a supposedly finished engine work for this game.
Say what you will about Frostbite, it does present pretty visuals when it works (it does work on my rig and my rig is certainly not beefy by current day standards, good but not great). I suppose that was part of the allure to go the Frostbite route as I would hope that adequate investigation into its shortcomings for this project was done prior to selection and factored into the project timeline and cost.
Depending on the state of the associated tools for Frostbite, it can take significantly longer to augment an existing engine than to create a new one or modify an existing one that the team is familiar with. Somebody else's code can be hard to read and figure out unless its adequately documented and heavily commented.
When I was developing (not engines, business apps and frameworks, significantly less complicated stuff), I often found it faster to rewrite pieces as opposed to try to make what was already there work. If not done correctly or if done in haste can lead to "spaghetti code" which is a nightmare to debug. Imagine if you will a bowl of spaghetti that has been left on the counter to dry out, now try and take out a single strand of spaghetti out of that mess without breaking it or any other piece of spaghetti (definition of spaghetti code). This type of mess leads to bug fixes being very difficult to do because no matter what you do you end up breaking something else (sound familiar?). Often the only solution is to completely rewrite sections from scratch. Not that I am claiming this is the case here but this type of situation was always my worst nightmare as an individual tasked with resolving post production issues.




Ce sujet est fermé
Retour en haut





