EvilBivol-1 wrote:At the risk of abusing the subject, I am often seen posting about the limitations of developer resources and the degree to which they define the development realities of a flight sim studio versus the expectations of flight sim enthusiasts. Although my own experience generally revolves around Eagle Dynamics, the article below describes the "behind the scenes" struggle leading up to the cancellation of Jane's A-10 Warthog. It's a goldmine of insight into some of the reasons why modern air combat simulators have all but vanished from the PC games market. It is information which, I believe, every flight simmer should read and digest, because it is in light of this kind of reality that I've come to appreciate the work of developers and hope more community members can as well.

The article demonstrates the impact that individual programmers can have on the entire project, the marketing calculations that determine development course, the ever-increasing complexity and cost of production in light of flat or diminishing sales numbers. I would remind you also - it dates back to 1999-2000. Ten years on, the complexities of a flight simulation and our expectations have only skyrocketed since, making the oppositions outlined below all the more unbridgeable.

I will quote some paragraphs that I find especially relevant, particularly in terms of ED's own experiences and decisions. However, I would encourage you to read the article in full: http://www.gamespot.com/features/pcgrav ... index.html


On Human Resources:

...the idea of a "development team" can create an illusion of continuity that simply doesn't exist when tracking the personnel who worked on the Austin team's projects. Only two programmers (producer Will McBurnett and wrapper programmer Steve Muchow) actually coded the Longbow series through all its iterations. The remarkable thing about the Austin Skunkworks team was that at any given time, the team members included some of the industry's most outstanding talent. While the Austin team has been marked by turnover, the Baltimore team has remained fairly stable over that same period of time. Perhaps one of the lessons to be learned from the cancellation of A-10 is that constancy and longevity go hand in hand when working on such complex, demanding projects as military flight simulations.

...

The development of A-10 became one of unchanging goals coupled with evolving means. As a source on the development team puts it, "Besides all the people coming and going, and trying to get the graphics engine nailed down, the game we were working on the last day was the same game we pitched in our design document." Unfortunately, the same cannot be said of the development team. When dealing with a team as talented as the Austin Skunkworks, it was perhaps inevitable that individual team members would leave to pursue different or more lucrative projects. As a game, A-10 never changed focus. But as a development project, it was a completely different story. As work progressed, it became clear that the initial intention of using the Longbow 2 code would simply not fly.

...

Furthermore, most of the Longbow 2 code had been written by people "long gone" from Origin Systems, and as work progressed it became clear that this was a significant obstacle to the game's timely completion.
This problem of outdated, cryptic code permeated every aspect of the sim's development. Alex Pavloff, who was then between his junior and senior years at the University of Houston, was hired as a summer intern on the project just as it was beginning, and shared the difficulties presented by having to reuse old programs.

...

"Simply put, the mission builder was pretty lousy," says Pavloff. "The program was a Win32 GUI [graphical user interface] application written in C. If I knew then what I know now, I would have asked permission to rewrite the thing in C++ and MFC, or even in Visual Basic. The mission builder had originally been written by Tsuyoshi Kawahito, who had left Origin after Longbow 2 to go to work for Microprose on European Air War. Apparently, he was the master programmer. He'd work late, get his stuff done, but he didn't comment his code. Clark Janes, the guy who got to work on the mission builder for A-10, was constantly fixing bugs in the editor, and the real designers and I would sometimes get multiple builds in one day in an attempt to make the thing work. Sometimes, however, the missions that we had made wouldn't work anymore in new versions, leading to much repetitive work in an annoying GUI."

...

As such problems began to appear, it became apparent to the team that the game could not possibly meet its ship date of first quarter, 1999. There was simply too much uncommented and hard-to-follow code from Longbow 2 that had been written by people who had left the company. This code had to be changed.


On Marketing Calculations and the Benefits of Independent Development:

At the time that A-10 was cancelled, Origin was predicting the game would sell 40,000 copies.

...

The success of Ultima Online had made Origin eager to capitalize on its success...
...In the words of Andy Hollis, "With the change of leadership at Origin, everything was looked at again in a new light." This change in priorities did not go unnoticed by the development team, as a source from the team reveals, "The first inkling I got was in September '98 when it became obvious that A-10 and Origin's online strategy did not mix. It was obvious that we had the best team in the building, and we should be doing online stuff! When we approached Lucas for the Star Wars license, that was the first step. That fell through, so [we] started to design Wing Commander Online. This was after the entire Wing Commander Prophecy team left for Bootprint. We asked for a decision to be made about A-10's future - with expectations that it would be supported, marketed, and sold well, or dropped right then - so we could work on the more important, more strategic, and much cooler Wing Commander Online. For some reason, Electronic Arts (not Origin) management said 'no.' They wanted the short-term A-10 money."

...

"[The] Jane's [brand] was the darling for a while, because it went from $0 to several million, and went to number one in market share. When the Jane's stuff stopped growing, it stopped being the darling. Then Jane's World War II fighters came along, spent some serious cash, and did not sell. Microsoft stole all World War II combat sales that year [with Combat Flight Simulator]. That was the death knell for Jane's products. No one wanted to be a part of it. All
the people who started it - Paul Grace, Andy Hollis, and Frank Gibeau - either left or went to different things. There was no champion at the corporate level, and the marketing guys hated it. Bye-bye."

...

When the success of Ultima Online was added to the mix, there was less and less of a reason, from a business standpoint, to continue with A-10, especially at a public company like Electronic Arts whose shareholders expected continued large sales growth. Flight sims were not a growth business, it seemed.

...

The new code, when compounded by the personnel changes, had conspired to delay A-10 by at least six months. In the end, time simply ran out. The game had missed several deadlines, and it was clearly not going to ship in 1999. In July of 1999, Origin pulled the plug.

...

According to Andy Hollis, these talks [to revive A-10 as a product of Third Wire Productions, EB] never proceeded past the negotiation stages. As Hollis puts it, "Had it actually gone forward, it probably would have made more sense for it to happen as Longbow 3, but the reality is that combat sims are really just nonstarters as far as the sales and marketing people go."

...

Simulations no longer provided a return on investment that justified their production. This didn't mean they didn't sell: Private Electronic Arts' sales numbers show that the original Longbow shipped more than 600,000 units worldwide. And, together, the five Longbow releases (Longbow, Flashpoint: Korea, Longbow Gold, Longbow 2, and Longbow Anthology) shipped more that 1.2 million copies. But the cost of production had gotten inordinately high, and returns were steadily diminishing.


On Flight Simmers' Expectations and Feature-Creep:

In the words of Andy Hollis, "Everybody was feature-creeping where the bar was," and expectations for new sims had simply become too high. "Building a product that would satisfy all these expectations would cost a tremendous amount of money," and the result is that "no one wants to go there." Are flight sims dead forever, then? "I'm sure sims will make a comeback," says Hollis optimistically. "If everyone leaves the field, then someone is bound to step into the void eventually." But how long flight-sim fans are going to have to wait remains an open question.

We now know... [EB].


On Dynamic Campaigns

While the team was working on the missions, the issue of the campaign structure came up. Because they were supposed to be based on the Longbow 2 code, the campaigns would be structured in the same way. This meant that they would have the "apparently dynamic" structure that Andy Hollis referred to repeatedly in the newsgroups as "smoke and mirrors." Pavloff explains: "The campaigns in Longbow 2, while seeming dynamic to the user, in fact, really weren't. The Longbow 2 campaign editor basically involved the designers creating multiple locations for the enemy troops on each phase line, and creating multiple paths and targets, and letting a random number generator create the missions. It became obvious that the mission builder had enough problems, and the programmers were slowly getting backed up to the point where the 'dynamic' campaign was going to have to be cut." So the A-10 programmers were faced with the possibility of having a game based on the Longbow 2 engine that actually had a less variable campaign than Longbow 2. This was an example of how, as code advanced from the Longbow 2 base, certain features either had to be left behind or made to work with the changing code. And this meant investing time. The "Longbow 2 shortcut" was already proving to be a false one.