No, not at all. See my comments- .NET is a proven framework, my offhand comment was about the library ecosystem around it. I’m editing the post now since so many seem to have understandably took it to mean more than it said.

Арсений, привет от бывшего коллеги по Криату =) Это Дима Астапкович. Индустрия такая маленькая =)

Regarding .NET- closed source DLL’s have given me too much grief and if they’re obfuscated you can’t really get good decompiled results, even for debugging. There were 3rd party controls in HeroEngine, and one of them would create a potentially several MB string each time you queried the text displayed in a control- which meant if you wrote lots of strings to it quickly, you’d get LOH fragmentation and crash! I understand the decisions but, we could certainly have optimized for this case. Lots of examples like that- they’re even worse than the copy-paste controls I used to find people sneak into our codebase! Mostly its the shitty middleware providers on Windows (who provide “enterprise” solutions in the form of awful dlls) that I detest.

Robert: See my statements above. Regarding not having enough time- that’s why I try to set down some hard and fast rules like “never build upon closed source frameworks” :) CAT doesn’t even become an option. Now of course we are free to break the rule but it makes the worst case the expected case so of course you have time to research now.

Ultimately, open/closed source is just another factor to take into consideration when choosing middleware. It is an important factor (in my book), but it is not something that unequivocally determines the winner. I.e. the choice between closed-source FMOD and open-source Audiere usually boils down to licensing/features, not to source availability.

Of course, source availability also affects the price (free open-source libraries / expensive source licenses for otherwise appropriately priced middleware) and the cost (support, etc.) – still, it is a factor, not the factor.

BTW I have change the studio couple oа weeks ago, and I’m thinking about your GDC talk about the culture all the time. That was and is really true.

I think open source is not so much the issue here. There’s many open source projects which, in theory, are serviceable by their users, but in practice are just too complex/advanced to do so, unless the problem is superficial or can be pinpointed easily and has few dependencies.

The issue I see here is: how much do you want to depend on 3rd parties?

Pick your frameworks, libraries and core technology carefully. Think it through for all your dependencies. Who maintains it? How long has it been around? How long will it be around? Has the feature changed significantly in the past? Is it an active project? Could you compile / fix it yourself (to what degree?), is there a community of users? Are there alternatives? Anyone else in your company has experience with it? How much functionality of the framework/library do you need? Can you write the required functions yourself? Is it portable (to older/newer versions of another framework/OS/target app)?

Downside: answering all this requires research and time and often we don’t have that. Short term time saving (well you turned to CAT to begin with ;) ) seems often more important than long consequences.
(especially when the fire is burning and they’re calling the tech art firefighters :/ ).