Considering close association with Atlassian, I find this write up to be disappointing and shallow. It touches on some broad generalizations of publicly available information. Except that what is public represents only mostly hobbyst development which only rarely brings any advanced insight. The article would carry same merit if it simply called all business and enterprise development a bunch of PHP script kiddies and Java code monkeys.
Console development is bound by NDAs and licenses are given only to proven developers. All of them are forbidden from discussing any details in public. Simple consequence is - there is no publicly available insight from anyone doing any kind of console development, aside from a random blurb, meaning PS3/XBox/Wii simply don't exist on the web. iPhone is
Nickleodeon. Most development is trivial and irrelevant as far as business practices or quality goes. Cheap is the name of the game. Big projects with $500/hour development rates are part of corporate development and internal branding or related efforts, so they are, again, not present on the web.
Big publishers are probably ahead of most bizdev shops. A simple google for "$(AAA publisher) + agile/continuous integration" or "MMO agile/contiuous integration" brings up articles like
this or
this. Searching around gamasutra brings articles
like this. These are companies where per-product investment is $100 million and revenue measured in billions.
The final misconception is regarding priorities. Source code is a solved problem. Not only that, but entire physical development is increasingly irrelevant as far as big picture goes. Entire development budgets can be
as low as 30%, that includes creative design, testing, coding and asset creation. Success of large publishers is almost completely dependent on marketing and franchise or licensed IP.
Then there is innovation which puts bizdev into stone age. Toolchains which allow
real-time versioned collaborative development on Xbox360 with scenes comprised out of billions of polygons - all real-time and interactive. How many businesses would consider one million database rows "huge".
The article illustrates a huge clash of cultures. But it has nothing to do with script kiddies learning in garages.
JIRA is de-facto toolchain for mostly OSS development stacks, primarily Java centric. That ecosystem is open, homogenous and fairly standardized. It's used almost exclusively for various forms of CRUD development. The toolchain it presents is well suited for both the culture and problems that need to be solved. The number of openly available frameworks and libraries facilitating all aspects of development is enviable.
But when one goes out of that industry, the ecosystems are closed and proprietary. While that does have slightly negative effect on quality it also skews the perspective on what goes on. People involved in such development simply do not engage into broad public discussions nor are they allowed to. Another industry where this is prevalent is industrial automation and CAD. Reason are similar - underlying toolchain is always proprietary and trade secrets and other legal implications of discussing such work freely pose too much of a risk. Evaluating these industries in same manner would be comparing them industrial automation by Arduino project.
The other aspect ignored is the actual needs. In aerospace engineering, deficiencies in CAD tools are irrelevant. If a tool does not have "select all", someone will be employed to click on all elements. LabVIEW is a pain, but solves actual needs. Bioinformatics is a mess (this is one field to look into, they really love the insight, but one needs a degree in medical or related field + CS) but pharma is big money, so throwing more manpower is still an option. Gamedev also sees abundance of labor - starry eyed kids, willing to work 90-hour weeks - they are cheaper than retraining top-of-the-line experts, whose time is valuable and expensive.
Article ignores the reality of business - in most industries, automation of certain tasks is too expensive. General IT, with falling profit margins need to minimize expenses on custom development and infrastructure. Minimizing labor makes sense. But elsewhere, just throwing more people at it works. Automation is only introduced when required or when the field becomes a commodity. This is why so much manufacturing is still done by people - they are simply cheaper. Agile does not necessarily mean machine automation.
Very few still have time to try to figure out the precise definition of the word "Agile" because ever since they adopted "agile" practices, software development is a solved problem.