What the heck do you think is a big development budget meant for, paying excellent programmers perhaps, or gasp quality management?!
And who would still hold the medieval superstition that arbitrary deadlines might be less important than consistent coding conventions?
Addendum 2017-01-02 07:49:
Btw, according to personal experience it goes like this far too often:
design your algorithm
notice it doesn't run as intended
start reviewing your code
stop reviewing your code - boss says so
notice that some third-party tool works different from its specs (or a built-in library tool)
test a dozen workarounds
as soon al one workaround works under certain circumstances half of the time, stop testing workarounds (boss says so)
nobody is ever, ever, ever again to touch this workaround - it might break; and if we need some extension, we write an extension around the workaround without really understanding it anymore, and then a workaround for that extension
nobody is ever, ever, ever again to touch this workaround for the extension for the workaround because point 9
Regarding point 3 of the second list, iterators either issuing a compiler error or just cloning the collection to make it immutable is normal in everything that isn't C++. Unreal is reasonably modern, therefore its iterators are immutable, like most other newer languages.
Funny thing about the Torque engine -- it has a longer history than this article suggests. Torque is not just some indie engine, it used to be the engine that powered Starsiege: Tribes and its successor, Tribes II. Both games were very popular in their day and had a lot of features that were far ahead of their time, especially Tribes II. They were very performant -- especially in regards to network latency. Tribes is a game known for its speed. The most-played game mode is Capture the Flag and flag runs are typically made at over 250 km/h in estimated real speed by jetpack-laden armored men skiing down mountainsides shooting each other with exploding blue discs. Somehow, this was playable on a 56k dail-up connection -- a miracle which today I still fail to understand. Yet another technical aspect of the game which always fascinated me was the sheer scale of its maps. In-game maps were actually sets of "props" (in-game prefabrications like base buildings, walkways, bunkers, etc) which were populated when the game loaded. The only static asset in the map itself was the terrain, although I have even seen Tribes 1 maps where terrains were copy-pasted and rotated to make areas that look like caverns or other unconventional ideas. This allowed for virtually limitless custom map making without requiring the client to download any new asset files -- after all, why bother making a whole new terrain when you can just arrange props to make a level that you like instead? Don't like the terrain at all? Make your level a base 500 feet in the air, because Tribes also has ceiling-less maps. And this is in 1998.
Suffice to say, when I was younger I was quite enthralled with the Torque engine for being able to do these impressive things so early in the development of "real" 3D games. Unfortunately it seems that the engine hasn't kept up its reputation.
I'm going to grognard just a little here and mention that this is kind of the reason that older games are way more impressive than modern games in terms of raw technical achievement -- namely games that came around at the dawn of 3D games, the late 90s. During this time, the best games were made with relatively little money by highly dedicated individuals. The Unreal engine was originally developed in Tim Sweeney's garage. The Quake engine was made by John Carmack and John Romero sitting around eating pizza and drinking Diet Coke for days on end in the same apartment/office. These were games that had purpose-built engines made from scratch, and in both cases the creators went above and beyond because they took pride in their work. Hiring 50 developers who don't know each other and expecting them to exhibit the same kind of dedication and passion, especially when that passion is usually shat on by your boss and the budget, is silly. We need highly dedicated and talented individuals and the people that are able to find them if we want to see less hacky, lazy work in game development, not tons of eager but inexperienced people managed by movie-producer-like people that only see a bottom line.
I actually have a copy of Torque2D from nearly a decade ago, thinking an engine would give me a head start over bootstrapping everything with Allegro or SDL, but I never really did figure out how to get it to do anything interesting...