User talk:Tennessee Ernie Ford/Archive 08

Maths: Runs of heads
Did you find a formula for the probability of a run ? Because it seems to me that for a large number of throws, to compute a close approximation seems not very hard when the approach is right.

For a large number of throws, the expected length of a run is 2.
 * To come up with the expected (average) length of a run E, given an infinite number of throws, consider this: A run is either of length 1 (50%), or it is longer than 1 (50%). If it is longer, then after the first throw, we're again in the initial situation, so its expected length after the first throw will be the expected length E, totalling E+1. Thus, E = 1/2 * 1 + 1/2 * (E+1). Solve for E to get E=2.
 * If we had only 6 throws, the expected length would be a lot less, because a run starting on the first throw would be limited to length 6, and a run starting on the 5th throw (which probably happens 1/4 of the time) could only have a maximum of 2, and so would have an average length of only 1.5. So in theory I'd need to come up with a formula to compute the expected length of a run as a sum of the probability that a run will start on a certain position times the expected length for that position, but for a large number of throws most of these addends would be very close to 1/n * 2, making the result so close to 2 that I won't bother.

If we have n runs of heads or tails, the probability of seeing a run of heads of length k or over is R(k) = 1 / 2^k.
 * Check this for k=1 : there's a 50% chance that the run will be of heads, and any such run will certainly be length 1 or over, or it wouldn't be a run, so 1/2^1 = 1/2 is correct. Having a run that is lonnger by one is half as likely, etc.

To find how many runs (n) we have to observe to see at least one run of length k with probability p, we simply solve (1 - R(k))^n = (1-p) for n, which gives n = log(1-p) / log(1 - R(k)).

Multiply by the expected length of a run, and you have the average number of throws required: N = 2 * log(1-p) / log(1 - 1/2^k)

For a run of k=10 heads and p=50%, this comes out to 1419 throws; and 14189 throws for p=1023/1024. For validation, this is fairly close to what the excel computation arrived at, that method is prone to suffer from propagating numeric rounding errors. --◄mendel► 08:52, 10 April 2011 (UTC)


 * I'm not sure what you are asking me. This came up when I was testing DDG vs Google by asking something like, how likely is it to get 50 heads in a row. This was only one of many results. The only reason I linked is that it includes examples of people writing clearly, people who obscure their ideas through unfortunate rhetorical choices, and that neither type of writing implied whether their analysis was accurate. Whether anyone addressed the question 100% correctly is to me, moot.


 * I wasn't particularly interested in the answer per se, I was interested in whether Google or DDG gave me a good shot at anything valuable. And, to a lesser extent, whether someone who was bad at math could get an accurate and useful answer.


 * As you can see, the (ahem) odds of getting a valuable answer by randomly selecting a result were poor even if one understands math; the chances (ahem) of being misled by a clear (but incorrect) answer were high.


 * And, for what it's worth, I couldn't find anything (without digging deeper) that was both as clear and as accurate as your analysis above. The excel version might have been accurate (within rounding areas, as you note), but was hard to parse unless you are good at Excel and math. The easier-to-read examples tended to miss something critical.  — Tennessee Ernie Ford ( TEF ) 17:11, 10 April 2011 (UTC)


 * Well, to me, it seemed like normal mathematical discussion on the 'net, with people responding to different aspects of the problem and at different levels of mathematical understanding - and some levels were probably not well suited to the person who originally asked the question. However, whether an answer sounds good and is adapted to the person who asked for it is indeed no guarantee that it is accurate - to judge that, you need to know more about the subject matter. I guess that's the difference between getting an answer and getting an education. ;-P And there are still too few people around who are educated in privacy. --◄mendel► 04:08, 11 April 2011 (UTC)


 * Yes, it was a completely typical discussion (mathematical or otherwise). — Tennessee Ernie Ford ( TEF ) 06:25, 11 April 2011 (UTC)
 * a typical discussion? (youtube link)
 * via Robot Beach, --◄mendel► 15:22, 11 April 2011 (UTC)

Ernie
/wave =) A F K When Needed 19:31, 13 April 2011 (UTC)


 * Hey! wb. (or are you just lurking intermittently?) What's new in the world of AFKness? — Tennessee Ernie Ford ( TEF ) 20:18, 13 April 2011 (UTC)


 * Well, he's a free kill, as always. --[[Image:El Nazgir sig.png|Talkpage]]El_Nazgir 21:15, 13 April 2011 (UTC)
 * Get outta here you strange, floating individual.
 * I tend to be lurking even when not editing; though I'm certainly not on here every day. Not much is new (yet - hopefully will have a lot to tell at the end of the summer) with me, anything exciting happening in your life?
 * (Is this the first time ever you accepted Ernie as being directed at you? That one's going on the calendar.) A_F_K_sig_2.jpg A F K When Needed 22:12, 14 April 2011 (UTC)


 * I accepted your wave; I didn't actually notice the header until just now.


 * Anyhow, there's never been anything wrong with Ernie, it's just that until I had been at this wiki for over a year, I never had anyone refer to me by that term...and I've been using this userID to post about RPGs for a very, very long time. In some ways, it's the more obvious choice. — Tennessee Ernie Ford ( TEF ) 23:42, 14 April 2011 (UTC)
 * You still have yet to object to the header.
 * Which, co-incidentally, proves you owe me €50. A_F_K_sig_2.jpg A F K When Needed 00:39, 15 April 2011 (UTC)
 * Takes the 50 and runs* weee, more alcohol!!!! (just kidding) Ariyen 04:23, 15 April 2011 (UTC)
 * Woman, you might want to get out of that sunny patch. I think it's going to your head... Arnout aka The Emperors Angel 07:35, 15 April 2011 (UTC)
 * Hai Arnout, how's u? A_F_K_sig_2.jpg A F K When Needed 10:17, 15 April 2011 (UTC)
 * Fine, thank you. Aryen stole my sunny patch and my cold beverage, thats all. Arnout aka The Emperors Angel 14:08, 15 April 2011 (UTC)
 * "beverage" ;) ∵Scythe∵ 19:29, 15 April 2011 (UTC)


 * I'd posted sooner, but here *hands Arnout back his empty cup*. Oh and here *hands him an alcoholic beverage*. Have fun. :-) Oh, before I forget... *gives Arnout the sun* now don't get burned . ;-) Ariyen 01:25, 16 April 2011 (UTC)

Explanation of move.
I've always used the in between reply because it adds a layer of indentation that separates replies. I use one more to indicate that it came after the following post, and insert it in the middle to make it clear that it's in reply to the original post and not the following post. I also don't like having multiple replies on the same indentation vertically adjacent because they get muddled together when quickly reading.--Łô√ë îğá†ħŕášħ 07:36, 16 April 2011 (UTC)


 * My concern is that Ariyen and Scythe both responded to the same post...and Scythe's came first.


 * But wiki talk pages always have trouble if sub-threads start popping up like weeds. So, when I post, I always indent and always post at the bottom and always add context...unless there's a really strong reason not to; otherwise, I find that someone always has trouble following each subthread. Sometimes adding helps when more than one post comes in response to the same prompt.


 * And speaking of edit conflict, I was about to revert it back (to avoid this digression)...but you beat me to the punch. I am sorry that I made a big deal of this in what is mostly just a thread of silliness, where wiki-etiquette should take a backseat to having fun.  — Tennessee Ernie Ford ( TEF ) 08:13, 16 April 2011 (UTC)
 * Actually, I agree to the way Gigathrash posted, because that's how I felt and why I did it the way I did in the first place. I was trying to follow what I learned in my talk, but in the bouts of having fun. See, mine was "direct" towards his comment, while scythe's was practically indirect and I tried using the best of judgement as to where I felt it should be. it is your talk and you can remove that entirely, but the majority are just out to have some cool fun. ;-) *hands you a cold drink*. I could enjoy that(fun) with the way real life has been. Ariyen 08:21, 16 April 2011 (UTC)


 * Well, as I say, I'm sorry I moved anything around because everyone was just trying to have fun. — Tennessee Ernie Ford ( TEF ) 08:26, 16 April 2011 (UTC)
 * It's alright. No harm done. :-) Ariyen 08:36, 16 April 2011 (UTC)


 * I just wonder why people can't grasp the underlying logic of comment threading. All replies to the same post have the same indentation, ordered by posting time.  Which means Giga's method is completely insane. @_@
 * "they get muddled together when quickly reading" That's why I've started leaving a blank line between posts, it leaves a slightly larger vertical space between the paragraphs.  &mdash;Dr Ishmael Diablo_the_chicken.gif 13:54, 16 April 2011 (UTC)

We're not educational? Since when?
I didn't clarify at the time, because I felt doing so would be unnecessary due to the context; I think the decision to "not clarify" is a mistaken made all too often. I do wonder why I never have such an enlightened thought before hitting 'Save page'...

My point was that we're a wiki documenting a game. Sure, I've learned loads from just watching others go about their daily routine. Being in a Computer Science course, my time at Wikia - and now at GuildWiki.org - has given me a not inconsiderable head-start in CSS, JS, less so in HTML, and tidbits of other stuff. Something I enjoy the benefits of immensely, as I can learn about what I find interesting while happily ignoring what's going on around me (to some extent) when in college. With some people here - primarily but not only Ishy - it would be a crime to ignore the wonders that are submitted regularly. I don't claim to understand half of what he does; merely try to educate myself on what others have done before me. I would like it to be noted as, being someone who hopes to make such things their area of expertise, I fully recognise the learning potential this wiki can be.

However, while it's all well and dandy that I get free how2wiki lessons - it should not come at the expense of the Wiki. People are free to learn here. To be inspired by the edits of others. To blatantly copy and paste such edits (within the terms of licensing, which means pointing at the real author). But they are not free to break articles in the process. My perspective, I think, is very similar to that of your own. Just perhaps a bit sterner. So regarding the cause of the confusion, I would point to how strongly I worded my thoughts originally - in conjunction with my "there's context, so I won't bother explaining" mishap.

I'll happily help anyone, my limited knowledge permitting. I will, however, take an unfavorable stance on those who leave articles in tatters and make no effort to contact others for help. Or revert themselves. Or find an Admin and tell them about it. Or join our IRC channel and inform someone. Or e-mail a user that looks vaguely helpful with the name of the article which was tarnished. Or make any noticable attempt to fix the damage that they themselves caused. We do not get paid for teaching others. I don't mind doing so (I enjoy it). But I do expect them to meet us halfway.

They broke the article and made no effort to repair it, neither did they attempt to aid others in doing so (nevermind telling us there was a problem). I can see how Ishy didn't assume good faith. I consider three days to perhaps be unnecessary. But I consider it in no way to be unwarranted. I would consider a week or anything less to be fair, if excessive. Anything more does really beg for evidence of it being a repeat performance and prior warnings being handed down.

tl;dr: Read it as "GuildWiki is not (primarily / intended to be) an educational resource". Our primary purpose must take priority.

(If you were being given free painting lessons, would you practise on the walls belonging to your instructor - or would you limit yourself to appropriate areas?) A F K When Needed 22:39, 27 April 2011 (UTC)


 * "Our primary purpose must take priority." -- If we were true to that call, you'd already be perma'd here and elsewhere; why is it that you get miffed over on UA for me not willing to educate you on wiki etiquette over and over? You've been the recipient of hours of "education".
 * If he didn't contact anyone, how is that his fault? Apparently half the other editors didn't know to contact him! Why do you ask the newbie who's panicking to meet you halfway when you have done exactly nothing to meet him?
 * "no effort to repair it" -- He did try to fix the problem (no map) by research and editing, though he did get stuck because this wiki is a tad more complicated than others he may have experience with. --◄mendel► 23:00, 27 April 2011 (UTC)
 * You ask me to AGF with you, yet you jump on any opportunity to bring such things up. See the problem?
 * If it makes a difference: No I don't want to suggest such things with you publically, whenever you feel like doing so on a whim. I'd revert you if I thought I could get away with it. Your message reads as nothing but trolling. Keep the business of one wiki, to that wiki. As said in relation to an RfA (was it Gigathrash who penned the words?) GuildWiki is, supposedly, a 'safe haven'. A_F_K_sig_2.jpg A F K When Needed 23:04, 27 April 2011 (UTC)

(Reset indent)

Obviously, our main concern is GW and documentation thereof. But I (and you and even Mendel and Ish) have made mistakes; we're not banned for that. We're expected to learn and move on...which makes this wiki, like any other community-driven tool, partially an educational resource.

Collectively, we spent less than five minutes evaluating whether their posts were vandalistic or failed attempts to make the page better; we spent even less time trying to help them to understand what they were doing wrong. Instead we just assumed bad faith.

I don't really care what you call it, that's not the philosophy of the Guild Wiki that, two years ago, attracted me to contribute here instead of at the official wiki. Surely, the wiki that has room for those banned from GWW can have more patience than that.

(As a side note for @Mendel, @AFK: please cease arguing or baiting each other on my talk page. I will move any future posts that, in my opinion, go off-topic and/or get personal. Thanks.)  — Tennessee Ernie Ford ( TEF ) 01:13, 28 April 2011 (UTC)
 * Sorry, I didn't intend for my post to be bait. --◄mendel► 10:36, 28 April 2011 (UTC)
 * My apologies: I didn't mean to suggest it was intended as such. The side-note was meant preemptively: I could smell the storm winds a'comin'. — Tennessee Ernie Ford ( TEF ) 18:39, 28 April 2011 (UTC)

YAV
The role parameter in your joke (?) block support template would break GW:YAV. Just sayin'. --◄mendel► 07:58, 18 May 2011 (UTC)


 * I suppose it would, but I was thinking (as a joke) that anything can be used (much as allows you to throw in any title). e.g. My name is Frodo and I'm the Ring Bearer here. I support this template.


 * Worse, though: the entire idea behind the template is a little too close to endorsing backhanded personal attacks. (Not even remotely suggesting anyone committed PA in the recent edits that suggested the template, but I think it's easy to see how a little (mostly innocent) joshing around could lead to ugliness.) — Tennessee Ernie Ford ( TEF ) 15:17, 18 May 2011 (UTC)
 * Yeah, good thing it's only a joke. ;-P --◄mendel► 15:35, 18 May 2011 (UTC)

un-bossing
Thank you for your clarification on Ariyen's talk. It's nice to see ourselves agreeing even on our interpretation of that quote. ;-) --◄mendel► 20:16, 29 May 2011 (UTC)


 * My apologies for having failed to clarify in the first place. (Thanks for taking the time to say so, too.) — Tennessee Ernie Ford ( TEF ) 21:27, 30 May 2011 (UTC)

Sabbatical
You know, while I'm a relative newcomer to actually EDITING GuildWiki, I've been reading it for years, since before the "official" Guild Wars Wiki. After the official wiki was rolled out, I stuck with GuildWiki, because to ME, I thought the information was presented better - clear walkthroughs, very helpful notes, and some VERY helpful screenshots. We have some very kick ass screenshots.

But also, I noticed who the regular contributors were, 1 of which was you. It seems nearly ANY talk page I visit has a comment from you, and it felt like you truly cared about making GuildWiki better, keeping it fresh & current. I never bothered with the behind the scenes stuff, which I'm guessing where all the e-drama is hiding in the open, until fairly recently, and even then I've basically skimmed over the arguments and what-not on some of the admin talk pages.

It was rather disheartening to be honest.

But it triggered something in me - I figured if people would rather argue with each other over policy about NPA and what not and not keep the wiki updated, then it was finally time for me to step up and help out where I could (and poor Ishmael, I bug him about anything I don't understand or know how to do). I'm doing my part (insert Starship Troopers poster here) wherever I can.

GuildWiki needs contributors. I do agree with you that GuildWiki needs to catch up. I'm an avid reader of GuildWiki, and while I haven't come across any stated GuildWiki "goals" I think everyone can agree that GuildWiki needs to strive to be the wiki of choice. That is a goal. It may be a never ending goal, but it is one that is easily worked towards. With the overall population of GW dwindling, it's getting harder to keep up & stay current, but it's not impossible. With GW giving users a handy in game link to the official wiki, we're looking at fighting an uphill battle. It's hard, but NOT impossible.

I don't know if you even play GW any more - if you do, great! Keep spreading the GuildWiki word in game. Make a pet project that will benefit the wiki!

Is the only thing that's making you want to take a sabbatical the e-drama? If so, then ignore it until you get called in to moderate it, and yes, concentrate on what GuildWiki does best - present clear, concise content. When a new content update hits, ensure that we're updated ASAP. Do what it takes (I'm NOT implying you don't already, just to be clear).

Step back, enjoy your breather. I'm hoping you don't quit due to other peoples e-drama, as that would only serve to hurt the wiki overall. Mauirixxx 11:01, 9 June 2011 (UTC)


 * I'd like to offer some support too. Out of all the people here, you have always been the most likeable and levelheadded person around. Often have I seen myself struggling with ways to explain or express my thoughts, to see you had already done so better than I could have ever managed. I agree that the wikidrama is getting too much what the wiki is about, and not about the game, the fun, and tbh it's one of the reasons I've been less active here for the past year or so.
 * All in all, I want you to know that I support you, pretty much no matter what you (strive to) do here. --[[Image:El Nazgir sig.png|Talkpage]]El_Nazgir 12:51, 9 June 2011 (UTC)


 * Thank you both for your comments and support. (Ironically, one of the things that makes it easier to take some time away is that Mauirixxx and Kaede have indeed stepped up to make sure the most recent game changes are getting added to this site.) — Tennessee Ernie Ford ( TEF ) 16:54, 9 June 2011 (UTC)


 * "you have always been the most likeable and levelheadded person around" —— +1
 * Apologies for being late to the party; in America atm with little internet access.
 * On behalf of all of us, thanks TEF Ernie! :-) A_F_K_sig_2.jpg A F K When Needed 04:11, 10 June 2011 (UTC)
 * Awww, Teffeny gets so much love over here! There's so much love in the air I think I'm going to hurl. -- Konig / talk 01:06, 13 June 2011 (UTC)

outage rebuttal
&mdash;Dr Ishmael 17:12, 28 June 2011 (UTC)
 * They *did* run tests on their backup system, and the last test reported as "healthy". (At least, that's how I understand the "despite reporting as healthy" phrase in their explanation.)
 * The replacement parts had a newer firmware version than the existing parts that failed, and it was this firmware that caused the incompatibilities. It's not Curse's fault that the vendor didn't send them identical parts.


 * I have sent them a message to let them know that it's still not listed on other sites, etc. Ariyen 18:53, 28 June 2011 (UTC)


 * It is Curse's fault for not bothering to verify that the vendor sent them identical parts. It is Curse's fault for having a "test" that can report unhealthy systems as healthy.


 * The only true way to ensure that your backup systems and plans work is to run simulations: take down systems, restore from backup, check if everything works. Pretty much every best practice study I've seen on this says the same thing: you cannot really predict what will go wrong. You cannot be sure that your staff will know what to do. You cannot rely on any standard tests to ensure that everything will come together. You have to practice: it's not ready → aim → fire accurately, it's ready → aim → misfire → re-aim → misfire less → re-ready → re-re-aim → misfire even less.


 * Lots of companies don't have to bother doing too much testing, because the cost of testing is higher than the cost of complete FUBAR for a couple of days. But: that's not true for Curse, who depends on uptime and their reputation in order to sell their type of product (especially since most of their money is in their delivery system, not via page views or clicks from websites).


 * Put another way: if GWiki was deciding today (instead of last year) whether to move from Wikia (and if so, where), would we have made the same decision?


 * I am also sure that we aren't getting close to the true story from Curse. (Not saying we should.) Based on my experience of being involved in internal disasters like this, I am sure that there was a lot more that went wrong than we heard about (I'm also sure that there were some heroes among the tech staff and vendor relationship managers).


 * (PS to Ariyen: thanks for letting Curse know about the inconsistent and outdated footers.) — Tennessee Ernie Ford ( TEF ) 20:45, 28 June 2011 (UTC)
 * GuildWiki wasn't last, not that I recall. I could reach GuildWiki before I could reach STO Wiki, and they didn't even make it on the footer here.  When it comes down to it, I wonder who really was last.  Someone had to be, after all.  Nwash  [[Image:User-Nwash-Eyes.png|link=User talk:Nwash]] 10:09, 29 June 2011 (UTC)


 * As far as the footer goes, not all of their sites can be listed from every other site. I think they're hosting something like 30 wikis now, and listing all of those would make the footer quite unwieldy.
 * On the other hand, there is no excuse for them to not have a complete list of sites on their main page. &mdash;Dr Ishmael Diablo_the_chicken.gif 12:37, 29 June 2011 (UTC)
 * There's also no excuse not to list GWiki from Guru 1 or Guru 2 (and it probably should also be listed from WoW-affiliated sites). In fact, why isn't GWiki automatically listed when Guru is in the footer? — Tennessee Ernie Ford ( TEF ) 17:45, 29 June 2011 (UTC)
 * I think I can probably get this taken care of. [[Image:Felix_Omni_Signature.png|link=User:Felix Omni]] 23:01, 29 June 2011 (UTC)


 * Cool. (I'm pretty sure they had GWiki on more of the footers before the outage, but perhaps I was just hoping that they had fixed that ages ago.)


 * BTW: did you (or the other GWiki contacts) happen to get an email from anyone at Curse since/during the outage? Did they apologize personally? Or just left us with the generic post that applied to anyone? (I tend to be willing to overlook a lot of outrage in favor of sincere apologies and credible promises to set things right or at least make things better.) — Tennessee Ernie Ford ( TEF ) 23:20, 29 June 2011 (UTC)
 * Yes, Wynthyst kept me up to date throughout the downtime and apologized fervently. She even took some time to make sure everything corresponded to Curse's explanation. [[Image:Felix_Omni_Signature.png|link=User:Felix Omni]] 01:34, 30 June 2011 (UTC)


 * I've been keeping up with everyone's opinions here, and took a look at my own disaster planning at work (I'm the sole guy responsible for *anything* tech in our little company). I have real time backups. I have weekly backups. I have offsite backups. Depending on the type of outage, or however bad the "disaster" is, I've tried to cover all my bases. If we lose all our data, me and 10 other people are straight out of a job. Period. But even with all that planning & preparation, I'm sure when the shit hits the fan, I'll find something I missed, and there WILL be downtime. It's just the nature of the game. There WILL be downtime.


 * Now, if who ever runs (owns??) Guild Wiki paid anything to Curse for hosting, then those people would be entitled to demanding better support, etc etc. From my understanding, we have a "free ride" with Curse, so much so that they haven't pushed any kind of advertising on Guild Wiki save for the footer - and that is awesome. The downside? Yeah, I'm pretty sure we're last or near last on their list of priorities of sites to get up first.


 * Again, it's just the nature of the game. I for one, am just happy they brought Guild Wiki back up. Don't be so critical of our free ride TEF - just enjoy it while it lasts. Anyways, things could be worse. We could be back on ... Wikia Mauirixxx 02:08, 30 June 2011 (UTC)
 * Guildwiki has no owner or solitary leader. We learned our lesson with Gravewit and Wikia.  No gods, no owners.  Only men.  And Women.  And the occasionally spacial temporal anomaly that has gained self awareness.--Łô√ë [[Image:Gigathrash_sig_G.jpg]]îğá†ħŕášħ  02:19, 30 June 2011 (UTC)
 * Gigs is so wise. 75.139.202.100 07:55, 30 June 2011 (UTC)

(Reset indent) Our ride isn't free; we provide Curse with eyeballs, off of which they can make money. Also: I'm not arguing for better support for Guild Wiki per se (except for a public apology: it takes at most 5 minutes for a staff member to post a note here). (I will note that I'm glad that Wynthyst took the time to apology to Felix, but I wish someone who didn't use to work here had bothered to acknowledge us.)

I am arguing that Curse did a poor job in ensuring reliable backup for themselves.

I've stated my opinion, based on years of tech experience, about what it takes to backup critical systems. As I noted in my rant, small companies should only do so much — the cost of the most reliable backup procedures, kit, and testing can easily exceed the risk to the company of simpler and cheaper systems that leave coverage gaps. But Curse is neither small nor underfunded. Its very business model depends both on trust and reliability. This type of outage cost them heavily in both.

There's all kinds of Ycch that can hit the fan for which no one can really prepare (if we learned nothing else from Japan, we should have learned that). But I don't think this really qualified. I've been at huge companies that had backup plans without running simulations; they were not able to recover quickly after multiple-event situations. I've been at smaller companies that did run simulations; they could recover even after the loss of critical vendor-dependent systems.

So, yes, there will be downtime. But: it needn't be so long and, based on what Curse has publicly disclosed, I believe that much of it was unnecessary.

(Finally, as a side note, if we were still at Wikia, we wouldn't have been down for a week. We also wouldn't have gone through the huge amount of work to move here...and the loss of readers. Based on lurking at various Wikia sites, their uptime and performance has been better than GWW or GWiki@Curse since we moved. So: are we better off here? It depends how much we value having more control over style (as well as content) compared to how much we value reliability. I think that control is very, very imporant — however, this wasn't the first time we've been hit with non-performance that affected people's ability to use our site.) — Tennessee Ernie Ford ( TEF ) 18:54, 30 June 2011 (UTC)
 * Wikia is several times larger than curse (or at least used to be, I've no idea how the sizes compare now that so many wikis jumped ship on wikia) it would make sense for them to be able to get replacement hardware much faster.
 * At Wikia we were stored in a London datacenter, at curse we're stored in an Atlanta datacenter. I'm assuming Atlanta is much more on the bleeding edge of technology than London is, being the primary location for Microsoft, Adobe, Apple, and many other massive computer tech companies on the east coast.  I don't know how London compares as I've little knowledge of the area, but it is pretty difficult to compete with Atlanta when it comes to being in the 'now'.  I would guess older parts with the same firmware revision would still be around in london, which would have taken 1-3 days off the downtime.  Wikia is also much better funded, they likely would have sent someone out with a bunch of external drives and copied all the files, physically getting copies to a new datacenter.  They've shown this by taking executive vacations while oasis was being pushed to wikis.  ∵Scythe∵ 18:42, 2 July 2011 (UTC)
 * Curse is not a huge company. I have been by their place a few years ago in Huntsville (went to another interview/job location around there back, before I got back on where I worked at (lay off) and I remember the building of the cool looking flame). It was actually smaller than a couple of the other business buildings. Now, I can't say for San Francisco, but I can believe it to not be big. Atlanta is 3 hours away and to think that Curse would "know" what went down, or have "extra" is too much assumption. Now, we have had things at the job I use to work at to go down and it take about that long for things to get back up, because of similar issues. Things happen. I wouldn't hold this against them and I don't. Having a dad and learning from one of the co-workers about their networking and servers, controllers, etc. I can understand. It took this company I was at - about as long to get things back on their feet with some things and they had to do things manually, until things got back into place. I do hope curse would learn to have a secondary backup spot to start up and use as a "mirror", until they got things back on track. Ariyen 19:09, 2 July 2011 (UTC)


 * Curse is neither small nor underfunded → that's obviously different from being huge. Unlike other firms of its size, Curse depends on being online and reliable. It's entirely fair to criticize them for failing to run simulated backups, which would have caught many of the issues they themselves claim to have been at the route of their troubles.


 * It's also fair to compare our situation here to our situation at Wikia (I didn't, in fact, compare the companies in any other way). Had we remained at Wikia, we would have avoided three long-term situations affecting our ability to function (initial lack of performance, the ongoing suspension of anonymous editing, and the week of downtime causes by a 48-hour company-wide outage). Does that mean we should not have switched? Not at all. As I wrote above, ability to control our destiny means a lot to a lot of us (including me).


 * People keep saying that shit happens. Of course it does. But good backup plans are meant to work when shit happens; there's no point to them otherwise. And the only way to know if those plans (and systems) will work is to practice, practice, and practice some more. — Tennessee Ernie Ford ( TEF ) 01:43, 3 July 2011 (UTC)
 * You compared this to, (or at least made note of) the Fukashima disaster. There were exactly three critical incidents that happened at fukashima.  1) the plant lost power, forcing them to use backup generators.  2) the generators ran out of gas, causing the plant to go without power to pump water to cool the reactors.  3) the tsunami that followed the quake "finished off" the plant after sustaining damage from a quake larger than the plant was built to survive.
 * Curse ran into a similar situation. The live system died, failure #1.  The backup system failed, failure #2.  The replacement was using a newer firmware version which was not backwards compatible, failure #3.  The checkdisk when the NAS drive booted took a long time, failure #4.
 * Curse had a longer chain of failures to deal with than fukashima, it would make sense to be sympathetic to curse during this issue, lots went wrong and it's probably the first time they've had to deal with this.
 * "It's entirely fair to criticize them for failing to run simulated backups" who's to say they didn't run backup simulations? They didn't have a spare copy of an expensive piece of equipment in house, you can't blame them for not wanting that kind of overhead.  The main problem lied in the new SAN hardware having an incompatible firmware upgrade to the older version curse was running, that's more of a problem with the vender than with curse.
 * "Had we remained at Wikia, we would have avoided three long-term situations affecting our ability to function" at wiki we were forced to work around Oasis (and still would be). That would have required hundreds of hours of work to reformat every page, and check them individually for display errors and undesired appearance.  In other words, Oasis is ten times the disaster this outage has been.
 * Think of it this way, if you saw a site suddenly change appearance and have everything be visually broken for a extended period of time, would you go back? I'd be betting on you finding an alternative with better looks before going back.  However, if you saw a site had a hardware failure and would be back online in a couple days, would you go back after those few days?  I know I would. ∵Scythe∵ 01:39, 4 July 2011 (UTC)

Clarifications
Soz, Scythe, I should have been more clear: I wasn't comparing to Fukashima. I was noting that Japan requires people to actually practice evacuation plans, whereas LA and San Francisco do not. As a result, loss of human life in Japan is probably far less than it's going to be in California when the Big One eventually hits. What I meant was: watching people able to follow the evac plans should have taught us how important it is to practice repeatedly.

I suppose the short story is: I'm harder on Curse than others probably because I've been in the middle of technical disaster recovery (at firms with similar capitalization and tech-dependence to Curse) — this smells like insufficient practice to me. I doubt that rhetoric alone is going to convince anyone otherwise.

Similarly, while I think we made the right choice at the time in leaving Wikia and coming to Curse, it has not been the unfettered success we had hoped. One way or another, everyone in my guild that used to use GWiki exclusively (contributor or browser) has switched to GWW (usually, b/c GWW was available when GWiki was not...and they just never felt the need to return). It wasn't about choosing between ugly and down for a week; it was about: what's available to me right now. — Tennessee Ernie Ford ( TEF ) 04:06, 4 July 2011 (UTC)


 * Another clarification: we were not "down for a week", we were down for about half a week, less than 4 days. The last edit before the outage was [ 08:42, 22 June], and the first edit after the outage was [ 06:05, 26 June] (US Central time).  &mdash;Dr Ishmael Diablo_the_chicken.gif 04:38, 4 July 2011 (UTC)


 * Fair enough. (It was longer for me — I wasn't able to view the site until the 28th, but that now seems to have been user or client error, rather than network. I admit I stopped troubleshooting after the 2nd day.) Doesn't change my essential points: (a) downtimes hurt us and (b) they hurt Curse even more. — Tennessee Ernie Ford ( TEF ) 18:25, 4 July 2011 (UTC)

Thank you,
for the commendation. Although I don't feel I deserve it (yet), as the only edit I made when you posted that, was pointing Giga towards GWW. I hope to change that in the future, though. Still, letting GWW do all the work and we copying it, sounds good to me. Arnout aka The Emperors Angel 19:53, 10 July 2011 (UTC)