Zelknolf's TA Platform and Q&A

[Zelknolf]

I do indeed have a platform, but it's rather simple: stuff is still broken; we should probably fix it.

I would pick on the performance creep from ALFA's general policy on building (that is, we make a lot of stuff that we don't use, and suffer for it), issues with interface (we, for instance, require six clicks to roll dice, and that leaves people who are proficient enough to be faster miffed by the interface), and issues with consistency (namely, our DMs very reliably run on schemes that follow PnP closely, but our actual development does not; I don't particularly care which route everyone wants to play on, but having the rules abruptly change because a DM has logged on strikes me as terribad.)

Realistically? The next few months probably won't see all of that fixed. Cutting out build (which we need to do) without cutting out functionality (which we really don't want to do) is something that merits significant quantities of profanity shouted loudly enough to wake the neighbors. But I'll work on it, and work on recruiting others to help at it.

Beyond that, I imagine it's room for Q&A. Have at ye.

[NESchampion]

Could you give us a run down on some specific issues you want to tackle with highest priority? You mention performance and UI issues, but would a UI issue like multiple clicks to get to dice rolls be priority? What about removing old, defunct aspects of UI like our old Set Description functionality as an example?

Additionally, what do you view as the biggest hurdles in terms of getting fixes implemented in timely fashion? Lack of staff, which you touched on in your opening platform?

Finally (at least for now), from the outside TA's role seems fairly opaque. I as a player have almost no idea what is being worked on at any given time, what priority new projects would have, or even in some cases what has been fixed recently. Do you have any plans for putting forward a rough outline of what is being worked on and what is down the pipeline a ways so that the community knows what is going on and where new requests are likely to be prioritized?

[Zelknolf]

Hiya NES--

The thought for priority is rather simple:
1. Upgrade infrastructure
-- Why? If we can't push out fixes, we're 110% fooked.
2. Data Integrity Issues
-- Why? Because if data of the wrong type gets saved, it makes our lives miserable for a very long time.
3. Server Crashes
-- Why? Servers going down sucks for everyone.
4. Client Crashes
-- Why? Clients going down suck a whole lot for whoever it happens to.
5. Server Performance
-- Why? Slow server makes everyone miserable.
6. Client Performance
-- Why? That stuff is something that will prevent players from playing.
7. Fixes for stuff that doesn't work at all (like uncanny dodge)
-- Why? Because it doesn't do anything. That's pretty broken.
8. Fixes for stuff that's horribly broken (like invisibility)
-- Why? Well... it works. Sort of.
9. Enhancements
-- Why? Well-- yeah, a lot of enhancements are really important for people, but if we don't handle the above stuff (i.e. the path we've been taking) we slowly grind ourselves into an unplayable game. And we can see the evidence of our having pushed ourselves to our limit-- ever notice trying to drive mode walk somewhere and, upon releasing the key, suddenly turning around? That's the server weeping softly into its pillow.


And I suppose the answer to that explains why you, as a player, have no clue what I'm working on. I've been up there in the #2-5 range for the most part, and if I'm successful, your experience (as a player) only changes in that you swear a little less at the screen. Though, if elected, I do think that I'd be reworking our methodology for pushing out fixes and haks. We currently have to bring all of ALFA down to do it, and that's already pushing resources too much; once we get an underdark server, it'll be pretty well impossible.

As for largest hurdles? I'd actually say that the issues to changing hak content (see above), our issues with bug tracking (namely, several giant threads without tracking after reporting), and probably a lack of staff-- I don't really know, because we don't track bugs.

[NESchampion]

Thanks for the quick response. As an additional question, any preferred systems for bug tracking you're looking at implementing? I've made use of Mantis for instance with regards to NWN1 bug tracking back when the HAZE project was being worked on, but I imagine there are a lot of possibilities to be weighed.

[Zelknolf]

Hello again, NES!

I see a few things that we need to accomplish with any system of bug tracking-- and something like Mantis would be good at a lot of them. Trouble is that I have no idea what sort of scale we're looking at for bugs right now, as our trio of megathreads is a relatively-poor indicator. I'm hoping that the scale is small enough that a handful of subforums (a hidden one for exploits, a hidden one for closed exploits [just in case they're not completely fixed], and five public ones for projects, high/med/low priority bugs, and closed non-exploitable issues) can accomplish the task, with the bonus of not being so opaque.

If we're in too deep or bugs actually show up/ get fixed often enough that interested people could reasonably expect to remember to check a separate piece of software, I'd probably be looking into that.

[Castano]

other gaming worlds having prettier servers makes me jealous. I don't like to be jealous it's a negative emotion. I'd rather pound the other servers into dust with ALFA prettiness.

So......hak & tech updates pls. in this order:

1. placeables (these are easier to do from a standards point of view) - Sigil, Adobe haks have cool buildings) ( I know these are currently being worked on)
2. Anything we need to implement better animations for PCs (PRON-free kemo)
3. PC interaction with the server (e.g. my long wished for player consignment shop for magic and crafted gear, automating the crafting rolls and rules etc.
4. A cooler overland map with moveable ships and wagons and horses (we're still in monty python mode - i know it's nwn2's fault for realeasign a game world with zero mounts)

pretty servers = happy players.

[Zelknolf]

Heya Cast--- I'm not sure if these are questions? It does seem to be soliciting an opinion, so: it looks like the sort of thing that would go in the hak suggestion thread; I can say that placeables are something that are easy to delegate, if the TA who wins here is willing to do so. Likely to ignore the priorities of people more familiar with the ACR.

I'd say something somewhat-similar of animations. I actually don't think that, long term, who wins this election will make a difference there. Sand's very much interested in the animations stuff, and will likely produce something we like regardless. I'd argue that I'd put the needed emphasis on rollout structure to get those animations to you quicker and to be able to update/ fix them more readily if it came to be needed, but looking at-- say-- a year out? Prolly the same difference.

Crafting, sadly, is 110% the domain of DMA. Charter is unambiguous in that DMA owns items and item compliance; Curm is only willing to see mundane items in, and that's an extremely-low yield relative to an extremely-labor-intensive project-- or it's low-yield, low-work, and low-quality (which I'd really really rather not do).

The OLM thing is a server content thing; you don't technically need TA approval to put one in. That said, I have three HDMs who want one, so I'd like very much to be able to say "If elected, I will make you a solid OLM!" -- but I'm not sure if I'll be able to get to it in the first term. A lot of things need to be done, and thus far, it has been extremely difficult to even get critical fixes into use, so there is a backlog that needs to be seen to, and our lack of tracking of issues (mentioned above), has me entirely unsure of how bad it is.

[hollyfant]

Hi,'lo. Good luck with the election race. Here are my questions:

The Tech Admin's domain, by its nature, overlaps with that of practically all other Admins. Currently there are "grey areas" aplenty. Do you favour peaceful coexistence, clearly defined borders or co-operation?

To me, and possibly others too, the Tech department appears "aloof". Communication is difficult and haphazard, and what does and does not show up in the HAKs a mystery in itself. How would you improve the visibility of Tech announcements and the availability of contact points?

[Zelknolf]

Hi Hollyfant,

I do think that your first question applies to all admin. Whose job is it to implement something like Vaultster? Charter says Infra; practicality says tech. Whose job is it to implement something like crafting? Charter says DMA; practicality says tech. As I see it, 'clearly-defined borders' are going to cause inefficiency. I brought up crafting; say that someone wrote it anyway and it contained magic items. Curm says "no" in line with his platform -- and then shucks, there's some wasted work. In fact, there's a lot of wasted work. Now, I'm happy to fight for the things that I think are important (clearly) and maintain that we currently make humans do obvious computer work based on insecurities in implementation, but I don't think that segmenting things is going to make us very happy campers.

To the second point, I think that your question is exactly why I am running. I think that, while AL is a great dude and a skillful tech person, he hasn't done a very good job of administrating this last term. One would note that none of the fixes he mentioned in November are in-- but that doesn't mean that none are done. Indeed, Mage Armor was fixed in October. And the trouble is that it's just as opaque to the actual tech people. There are at least three people who need talk table references (BHM for PrCs, me for DM tools, and Riotnrrd for iprp_spells) but there is no organization. I guarantee that we're going to be sorting through conflicts in those numbers if we ever get to implementing things.

This is itself why I think myself suited to the job. Clearly, I'm not especially well-liked by all around here, but the last time I had power to make changes, they were all posted beforehand in a public forum; after the posts were up I mentioned in chat what I was working on, and what I was putting into the haks. I invited feedback (it did seem to take people a couple months to figure out that "Zelknolf is working on [x]" means "[x] will happen if no one says anything; if you object, say something where Zelknolf can hear.") and, when I released anything, posted comprehensive release notes (and still do, but it's a whole other rant when I give a "I've changed your module" note to someone and it disappears into the ether-- so I guess I'll have to maintain release notes for the modules myself if I want them to exist). Or, at least, I attempted to; I think I did a pretty poor job of explaining the implications of fixing syntax errors (things that used to not compile suddenly starting to compile can mean all sorts of things). When I went on haitus from ALFA to finish my master's research, which was run through an instance of the Aurora engine, everything I produced was transparent enough to place before a committee of doctors of English and be understood (indeed, I get to add that I am a master of rherotic and composition thanks to that little bit) and professional enough to place before a software company and to get me a new job (indeed, this is why I'm a Wisconsinite these days). I do know that I am capable of providing the needed organization. I'm not especially happy about it, as I have to do so at work already, but I know that I can make it work, and that I can make it understandable to a non-technical audience.

[Zelknolf]

There haven't been a lot of questions this election, and I get the feeling that it's because people have already made up their minds for the most part. I do think it worthwhile to offer a plea of sorts.

My platform is "things are broken, and we ought to fix them," because that hasn't been happening. I would ask you all when the last hak update was; I don't remember. Every month for the last eight months it has been "hopefully next month," and the updates we're looking for aren't questionable; they aren't looking for anyone to make an executive decision. They're things that have already been approved -- like prestige classes -- things that are obvious fixes, which we'd actually had fixes written for since 2003 -- like mage armor (though I did fix the fix before reposting it; mage armor used to play less-than-nicely with normal magic armor and magic vestment. Even then, that's been available since October) -- most recently, Skywing put some work into creating an improved compiler, which turned up data integrity errors in our ACR. I don't expect the common person to follow, but it is essentially when we write our code in such a way that it lies to itself. (We say "This little bit of programming will give you an integer." and then write it such that it never actually does. Sometimes we get away with that, but it often causes crashes, with the computer grumping vigorously and shouting "You told me this is an integer, but I can't add it to three!")

Given that one of the bugged functions (the debugging system) gets use in persistent storage, I've a hunch that it spent so long being so crashy because of this. No real blame to be put on AL or Cipher; the compilers didn't catch it (and they really should have, and this sort of thing is part of why Skywing was fitzing with the compiler). The problem is that I wrote fixes. Immediately. They got packed into haks. Immediately. They got put on the FTP with release notes explainig exactly what changed:

acr_debug_i
void ACR_CreateDebugSystem()
Previously: It was declared as an int, and implemented as a void.
Now: It is int in both places, and returns the debug number.

acr_resting_i
int ACR_GetIsFatigued()
Previously: It was declared as a void, and implemented as an int.
Now: It is int in both places.

acr_spawn_i
const string _WP_RESPAWN_DELAY_START
Previously: Declared with no value
Now: = "ACR_RESPAWN_DELAY_START"

And this has upgraded us to "Hopefully next week."

It is next week.

It isn't isolated to just the hak updates. Tech people have been looking to me for organization and direction, and I've had to effectively turn them away. I do say what the truth is-- they have to poke AL about most of it, because I lack the authority and information to tell them what they want to know or to approve / disapprove of technical fixes. Though I believe AL's own platform shows how well he's been tracking that; he says that, in our heyday, we had 4-5 active contributors. I say Zelknolf (DM GUI / performance / bugs), Blindhamsterman (prestige classes), Regalis (instanced areas), Riotnrrd (item properties), Sanderman (animations). If Acadius dubs himself active and four is, truly, the threshold for keeping track of work and changes, we needed this organization by his own marker in December. If he does not, February. I think that "four" is a pretty poor threshold, and would argue that "two" is enough to have cause to put effort into coordination. It is easy enough to prove that AL was aware of that much, as he added me to the tech team himself-- and then did not make any mention of projects dubbed dead or orphaned, projects added, or things approved-- even when I put them directly in his inbox. Toughness would be done if the needed approval was on the table, but only half of the affected admin have said anything where I can hear it.

I do admit that AL has been more active since being challenged, and maybe this is the kick in the pants it will take to get our organization put together. I stood for the position because, frankly, I didn't think AL would run (one might note his platform in November)-- and I stay in because I am unconvinced. There are things that he notes as good ideas, yes, but he has had the power for three straight years to make it happen and has not.

I know that I am not well-liked by all, and many disagree with things I preach about in firey ways. But the TA isn't the face of ALFA; the TA doesn't give strikes; the TA doesn't make ruleset decisions; the TA makes implementation decisions. I do hope that I can convince people to vote with that in mind, so we can start patching up our game.

[Sandermann]

I think theres not many questions because the majority of the membership really has little indication as to what the TA does unless things go horribly and visibly wrong.

With control over the ACR the TA has a very real impact on the in game environment.

I really only have one thing to ask, where in your interpretation does content cross the line between a tech domain issue and an individual HDM issue. What should be ACR and what should be down to HDMs?

[Zelknolf]

Hiya Sand,

Well, to answer the letter of your question, I'd note what the ACR itself actually does. It's in-game infrastructure, data collection, persistency, organization, and a bit of debugging. If it is written well, it is completely invisible, because it's responsible for handling the things that nwn2server doesn't handle itself.

But, to answer the spirit of your question (wherein the charter's only mention of core-rules-ish things is indeed under the TA's domain):

I would respond by saying that the borders aren't (and won't be) my job, first. The charter's answer to "where are the boundaries" is "the admin should get along. If they don't, the LA decides whose job it is." -- so, an excellent question for Veilan!

Experience says that Veilan thinks core rules are the DMA's job. He did, as head of S&T (a DMA staff head), have no problem with discussing changes to the core rules and dubbing the DMA's ruling as binding. I doubt he would do that if he didn't think the DMA's power legitimate.

For my own opinion? I'd say that, while core rules are unabashedly the TA's responsibility, and I think that it (rightly) gives the TA de-facto power of veto over any major change to the ruleset (by simply saying "Approve all you want; we shouldn't implement this."), that power should only be used when the implementation itself is a bad idea (i.e. "Don't worry! We'll just write to this here remote database across the ocean every tenth of a second and everything gets caught in the refresh! What do you mean 'sparse data'? It can go into a table, can't it!?" = "Zelk disapproves, even if it provides muffins and backrubs to all." -- "We should add a Bladesinger PrC!" = "Zelk disapproves, but doesn't stop the implementation if that is the majority opinion.").

For the concepts of rules changes, that hits every admin. A rules change affects players (as they often must use these things) and thus is the PA's responsibility, for he/she must represent them; it affects DMs (as they must handle them in practice) and is thus the DMA's responsibility; it has a clear connection to Tech (as they must make it), Infra (as we must be able to handle it), and the HDMs (as they are responsible for the maintenance and content of their servers). I'd even argue that Lead has a fair-sized stake in it (as a PW's features are the main hook when attempting any PR work, and there are features-- or the absence of features-- that can make a PW unsellable). I'd prefer to put things to a vote among that group. I think I could make the judgments if I had to and would do so with the intent of making our game more functional/ usable/ consistent-- believing, purely and truly, that it needs enhancements to not get in the way of fun-- but I'd be happier and think the game world at large would be more consistent with votes from all of the major stakeholders.

[AcadiusLost]

I'm sorry, I do feel the need to dispute a few things here.

They're things that have already been approved -- like prestige classes --

The PrCs BHM has been working on are presently still, by his posts, "ready for testing"- that's a far cry from being held up by hak update schedules. I will make every effort to put the resources into the haks once they are finalized and ready to roll out.

- most recently, Skywing put some work into creating an improved compiler, which turned up data integrity errors in our ACR. I don't expect the common person to follow, but it is essentially when we write our code in such a way that it lies to itself. (We say "This little bit of programming will give you an integer." and then write it such that it never actually does. Sometimes we get away with that, but it often causes crashes, with the computer grumping vigorously and shouting "You told me this is an integer, but I can't add it to three!")

I am of course interested in straighting out the screwy declarations, and appreciative of the compiler and the effort in tracking them down. However, the declarations all appear to be associated with unused/never completed systems, so they don't strike me as top priority for addressing. I'd certainly put them at the top of the list when it came to ACR fixes, but this current hak update is focused on builder resources. The p-storage system is (very occasionally) buggy due to the lousy OEI campaign database system, which isn't well coded regardless- it's not an ACR issue, and certainly has nothing to do with unused incomplete debugging system declarations.

And this has upgraded us to "Hopefully next week."

It is next week.

This election cycle happened to hit during 2 major grant deadlines for me, with my career and job on the line, followed by major facilities work at my building causing me to work extremely irregular hours, plus a visit from my mother in law, who stays in the only room I can effectively use for tech work. And now, in some of the limited free time I can carve out of all of that, I have to answer posts like this, lest people think I'm just asleep at the wheel or something. A hak update WILL be rolled out tonight, even if it means I miss another night's consecutive sleep over it.

Though I believe AL's own platform shows how well he's been tracking that; he says that, in our heyday, we had 4-5 active contributors.

In that case I was speaking of the times when we were doing most frequent hak updates, earlier in the NWN2 development cycle. The names and faces changed depending who was burnt out / having hardware difficulties / etc, but there was definitely a large lull following BG-Live, which has only more recently turned more highly active. I think one could certainly make the argument that ALFA is more active now development-wise than it was back when we were running frequent updates; but it's also a lot harder to roll out hak updates now that we use the autodownloader functionality. I am certainly amenable to getting back into an update every month or two if we have content demand and fixes to go in.

As I've said, I'm coming out of a very busy and stressful period of RL, which has limited my ability to respond to / inspect / test useful content such as the fixes Zelknolf mentions above. Where there have been critical problems that interfere with playing/DMing ALFA, I would argue that my response time has been quite fast.

[Zelknolf]

Of course, I'll not let that settle alone--

To the points addressed:
-- Mystic Theurge, Mage Armor, Overland Maps, Vaultster.

Though I acknowledge that Overland Map fixes (also reducing the incidence of verifiable client crashes) did eventually go in. But from "I've tested everything I can with one person logging in to a little test server." to "Loading this on a test server so that we can get multiple people on." was from December until March.

By the by, what ever happened to the release notes I gave you and Curm with that? I've got danielmn trying to say that I've created a magic portal from Auvandell to High Hold on account of their absence. Unhappy times.


-- "Never completed" doesn't mean "never called" or "causes no problems." You're, of course, correct that they look very unfinished, but one might look at where the calls are.


-- Grant deadlines do suck, I will admit, and I have worked in Academia before, knowing the work that must go into a grant. But if the result of your life is that updates come every eight months, it's time to delegate. Being a bottleneck who can't work on things because there are more important things to do is not good for anyone. Letting the people who try to contribute to the community also have to do so without the information that they need in order to contribute is not good for anyone. At best, everyone ends up angry, the contributors end up overworked and stressed -- or just not contributing -- and the players end up playing in a world that can't compete with the other persistent worlds we're being compared to.

And let there be no doubt: there are more important things to do. I would like you to get enough sleep, eat meals, and work at your job. Those are good things. Those are more important than ALFA. Any one of those is more important than ALFA. If a hak update is the difference between six and eight hours of sleep one night, I'd even say that then you should be getting eight hours of sleep, and asking someone who has the time and knowledge to handle the hak rollout-- and I can verify that at least one person (hello!) approached you with a very direct statement of being willing to take some of that burden, who was experienced in handling hak rollouts, managing servers, testing content, and all of that charming stuff. That person's inbasket remains quite empty, and that person most uninformed of what needs to happen.

And that offer is on the table still, regardless of elections, positions, tensions, or blah de blah. I have, and will continue to, assert that something is horribly wrong when a game is injuring a real person. Missing multiple consecutive nights of sleep for a game is injury. Share the workload-- please.