From What I understand Bellard's code is fairly "dirty". FFmpeg, as I understand it, was hard to modify, maintain, or integrate with.

Don't get me wrong, he's a fucking wizard. He's Mozart when the most talented of us can only ever hope to be Salieri. But part of that is, he dashes off brilliant code without much thought to its maintainability, then leaves for the next project.

Yea sorry, I guess I could have picked my "heros" with more care maybe add in a John Carmack ? :)

>Don't get me wrong, he's a fucking wizard. He's Mozart when the most talented of us can only ever hope to be Salieri. But part of that is, he dashes off brilliant code without much thought to its maintainability, then leaves for the next project.

Absolutely agree's - Well that is "maybe" one unpopular-metric we need to consider as well ? - How much brilliant code can you ship that has a big impact ? I could spent 1 year to "correctly-code some system" but if no one cares or uses it.. does the code or the "better quality" even matter ?? #BigPictureKindOfWay ?

To be clear, I have no idea what is the correct answer here :)

That's the thing about Bellard. I don't think he cares really whether his code has an impact or not. And yet it does.