Wynna wrote:Finally, perhaps we could hear from TDawg about just how many servers Standards can approve, and in what time period? 3-5 servers up front and two servers every three months after was based on advice from TDawg, as Standards head, as to what was exactly feasible to release as fit to go Live. It was not randomly chosen. Whether or not #1 or #5 or the synchronized swimming option win the popular vote, the point will be moot if 'Once approved, and standards have been passed, servers should be released when they are ready for live with no further stipulations' means 3-5 up front and two every three months thereafter.
Realistically speaking, I am not going to be involved in serious NWN2 Testing. I'm writing up a guideline on how that would work for any who follow me, but my general indication is that there'll have to be a group that follows.
The nature of how specific and critical the testing is will dictate some aspects of the crunch on that step. There are a number of things folks need to consider moving forward:
1. Will Testing include bug checking like doors that lock/unlock appropriately, merchants that spawn, no double spawns on combat, and how serious will that be? Will mob blueprints be examined to gauge drops, adherence to AI standards, and appropriate CR ratings?
Right now, all Standards /has/ to test is that merchants price correctly, and that statics don't overreward. I try to do brief runthroughs, and NickD helps, but simply put we can't cover even a significant portion of the server in a reasonable time, unless we wanted to drop everything else including actually playing ourselves. Using the community for massive Beta testing purposes will cut out a fair chunk of this, but as more servers go Live, your capacity for this dwindles - and there's awlays the unspoken "meta" risk.
2. Will servers be required to send expansion material to Testing, or will Testing have a periodic - maybe quarterly or biannually - auditing requirement on them?
Right now, a server could go Live with 100 zones, add 100 more with inappropriate content, and Testing doesn't have the resources to handle it. Some of it is me overcommitted, but some of it is just lack of personnel. If you combine that with a serious build cycle /while/ requiring new material to be approved or auditing, you're talking serious work.
3. How much "base" material can or will be provided by the Basemod, and how well will builders stick to it? There's such an infinite multitude of magic items that can be created that the best we can do is hit the major highlights, but as soon as people start getting creative (such as the magic merchants on SD) then pricing fidelity becomes an issue.
Right now, every single item in every merchant is examined. The more items in shops and the more errors, the more time it takes to run through a server. The question of standardized mob blueprints running standardized scripts becomes another question - any humanoids with variable class levels are now things that /should/ in principle be scrutinized.
4. Will there ever be an "aesthetic" component to Testing? Or put another way, approving a server's proposal on paper does not necessarily equate with something that looks well-designed or professional, and what extent is the "back end" of that important?
Right now, we've sent servers to Live with massive bugs - poor typos, missing conversations, busted ATs or AT systems, mobs without conversations or any reactions, doors that go nowhere, and so on. If they're gamebreaking - such as mobs spawning on top of ATs, they are clearly squashed, though. We've done so with the understanding that the DMA and the HDMs "sign off" on it, and on the basis that amidst all the other servers, things will improve as the players motivate them to improve. Is that an acceptable mode of operation for NWN2 - initial batch or even a year down the road? If it's not, what kind of system are you going to put in place for Testing to make that call?
On top of all of that, you have the question of what the timelines are, and how many people you have in Testing. Right now, given back and forths and multiple revisions between Testing and builders, servers typically go Live about 10 - 30 days after submission. And that's without serious in-game testing, roughly 5 - 10 manhours - maybe about 20 manhours total of Testing work per server with a lot of things left untested. How much time the testers can dedicate per day, their attention to detail measured against their attention span, how their schedules mesh with the builders, how the work is distributed - there's lots of room for slop time in there that eats up more days than you'd expect. The size you'll want Testing to be is absolutely a function of the answer to all of those questions.
If I were to guess today, I'd say you want five to eight people in a thankless, boring-as-hell job where those people are always gonna be up for attack for their meta knowledge. Tough prospect to start.
As to how you want to structure it, I was somewhere between #1 and #2 so I voted none of the above. I can't see the concept of a group server working unless it bears no real relation to ALFA canon or to Live servers in general; would just be too chaotic. I guess I hold the same opinion of releasing a single server (as soon as available) for Live. You can give ALFAns a home on Beta servers - and even harness them to help the Testing curve - in a reasonably short period of time, say Beta 2 days. And being non-Live-canon, you can kind of let DMs go nuts on them running ad hocs or random things and keep people thoroughly entertained without the seriousness of the "ALFA pillars."
The way I saw it, you'd want to release a minimum of three, hopefully more for "Live" even if it meant holding a couple back for two or three weeks. If for no other reason than to spread players out a little bit at a time that counts as "official." Ronan's concept of redundancy is spot on in that to get three finished in a reasonably brief period of time, you want to shoot scattershot by approving seven or eight for that initial bunch, but I've always believed that if you approve a server and they can build it, you should never tell them "oops, too many servers, sorry about those hundreds of manhours of work but we don't need it." In that sense, any you approve off the bat have the capacity to be ready for the "Live" push, except the server selection process would prioritize which ones get Testing priority.
So I believe in releasing as they come available with a
single hold point for ALFA "Live" that makes sure we start with more than just one. That's close to #1, #2, and #3, but not any of them, so I voted none of the above. I believe in a hold point (#1's out), I don't believe that we should hold back approved servers on density numbers (#2's out), and I don't believe we should insert any sort of "stablizing" part into adding more servers (#3's out).