A Note About The Recent Fence Bug

[Brash_Endeavors 4,735]

As Emoo said.

It is only because we love the game so much, that it affects us so deeply.

Edited December 23, 2012 by Brash_Endeavors

[Wraithglow 787]

The issue is that the problems are not detectable sometimes until after many things come into play. I am sure Rolf would rather not have issues he needs to fix later. it is counterproductive. But unless the exact things happen not all bugs can be detected. The test server is not a proper representation of the normal player environment. Not enough people on not enough of the same building to represent the lag etc. Rolf i believe does the best he can. It is very easy to say what to do when it is not you who needs to do it. I have seen many a simple problem turn into a major catastrophe.

I think the point is that if a bug isnt noticeable until it has been tested on the live server, then your test scenarios aren't thorough enough. I work on financial software and I assure you that there is no scenario that cannot be tested on a test server beforehand. Yes not all bugs can be detected, but if you tell your community beforehand what is going to be introduced, and where to test it, at least it avoids all the dozens of threads we get where people say "why did you upload this new update with this new bug" - you can say well "HERE" is where we announced it and HERE is the window where you could have reported any problems.

In my experience, no programmer should ever be the one to test an update - we just don't have the tester frame of mind

[Elen 1,013]

Keldun's post, with one little very practical add-on: Rolf, you should not stand up after doing an update for at least 15 minutes. Things like fencebugs and other major catastrophies are mostly found within just the first ten to fifteen minutes and reported. Be ready to flick the switch right there and then in these cases and do an *immediate* rollback (to the back-up that was of course made before any update...). So no-one gets hurt badly.

Most bad bugs wouldn't have bad consequences if just this simple rule could be followed imo.

This really would help. Be there and be ready.

[spellcast 983]

A couple of notes/ideas about compensation...

Would it be possible to make a NO DROP / NO TRADE / NO WEIGHT item that has the same properties as sleep powder?

With such an item where each one gives 1 hr of SB, Rolf could simply provide for players to have 1-10 of these placed directly into inventory as compensation for bugs etc, this would give a few benefits:

- 1 - Players could use them when they need them, not losing any compensation if they allready have SB stored up.

- 2 - Rolf could give out different amounts based on how severe the bug/problem was, right now it is very all or nothing (5hrs or 0, for smaller bugs 5hrs might not be appropriate, so nothing is given instead)

- 3 - Head GM's could maybe also be given the ability to spawn these for more limited bug situations where it only affects one or two people, allowing a small compensation to be given specifically to a player that is directly affected.

Same sort of item (ND/NT/NW) to give premium time in 6hr blocks, that could be placed in the inventories of players with prem time when a major outage occurs. (i'm thinking major as in, servers are down for 5hrs or more, or otherwise completely unplayable to >90% of the playerbase)

[joedobo 1,187]

When things like this happen thier is nothing worse then the classic fan boy response "shut up and quit if you don't like it". It be nice if we could moderate that kinda of inconsiderate, imo troll talk, thing from further enraging folks.

About compensation; I didn't loose anything but If I did the only acceptable compensation would have been new 5 speed horses of different parentage.

[Keldun 335]

When things like this happen thier is nothing worse then the classic fan boy response "shut up and quit if you don't like it". It be nice if we could moderate that kinda of inconsiderate, imo troll talk, thing from further enraging folks.

Problem is, that is not Troll talk from the fanboys, that is the official statement from Rolf!

[Reylaark 1,100]

I think the point is that if a bug isnt noticeable until it has been tested on the live server, then your test scenarios aren't thorough enough. I work on financial software and I assure you that there is no scenario that cannot be tested on a test server beforehand. Yes not all bugs can be detected, but if you tell your community beforehand what is going to be introduced, and where to test it, at least it avoids all the dozens of threads we get where people say "why did you upload this new update with this new bug" - you can say well "HERE" is where we announced it and HERE is the window where you could have reported any problems.

In my experience, no programmer should ever be the one to test an update - we just don't have the tester frame of mind

weeellll... I dunno There do seem to be some scenarios in financial software that didn't always successfully test for hack*ahem*, sorry, I mean "compromised" situations (yes, Bank of America actually corrected my vocabulary on that one). Ever see a mutual fund accounting system go down on the last day of the month? It's ugly. Seen the wrong price get sent out to newspapers, Euroclear systems go down, Nasdaq gets glitches, etc.

Seen collision avoidance software in a Cessna 172s be so buggy it constantly reports ghosts and misses the giant learjet coming at you... how in the world does something like that not only get through testing, but gets installed on aircraft? In another case, an Airbus, which are exceedingly computerized aircraft, crashed during its maiden flight because a programmer felt that below a certain altitude, the landing gear should automatically be deployed. Apparently all of their testing scenarios didn't account for the increased drag should a pilot willfully fly at that altitude for... say... a maiden voyage showing off to investors. You would think they have pretty thorough testing!

Bugs do happen and in much worse places than Wurm. That's not to say I'd expect anyone to kindly take it on the chin, always better to vent than to fester, but you know, a bit of perspective and understanding that this isn't at all unique to Code Club. No real horses were injured in the making of this game.

Compensation... hmm... I didn't lose all the horses on my deed, just a vast majority of them. However, had I lost all of them, I would have to call upon neighbors and friends to see what they can do about helping me out... just as I would gladly do for them. Disasters happen, and how a community responds to it can go a long way to ensuring that community's survival.

For testing purposes, could the test server really be a copy of the main servers? Not sure. The reason I'm not sure is because when testing, you'd want to be able to limit, control, and introduce variables as needed. Bugs are like needles, they hurt, and when looking for one it helps to make the haystack as small and manageable as possible.

I am sorry for everyone's losses. I know my favorite horse, Maestroheart, who's been with me from the start, won't live forever. It's always a joy to see him kickin' around when I get back to the game. I understand we get attached no matter how virtual.

Nonetheless, there is still vastly more for me to love about Wurm than there is to hate. I hope everyone can feel the same about that, because it is a more peaceful and comfortable feeling than the alternative over which I have no control. There are certainly other games, some very similar to Wurm, that passed the threshold for me. I wouldn't go near those with a 10 foot pole even if paid handsomely. I hope none of you get to that point.

I would like to +1 the idea of having patches go on either particular days, such as a Tuesday or Wednesday... or with an organized anouncement. It would let some people make sure they are present if that is most comfortable for them. It would also let others know to keep away for a while those days as a rollback may very well be needed.

Peace, Love, and good health to you and yours this holiday season.

[Keldun 335]

Ok, when looking for a specific needle it is good when the hay stack is as small as possible.

But when checking how many needles are in the big hay stack, it's not enough to take an armfull without needles and say the entire stack is clear.

And yes, big mistakes happen in big comapanies too, but Airbus doesn't crash every maidenflight, glues it back together and see if it flies this time.

[Trevize 141]

I quote myself from another topic, because i think its relevant here too. What do you think about it?

The worst thing in this story is that the bug was caused obviously by a coding error. I have a lots of levander plantations on my deed. After 1.0 we cannot pass them (before 1.0 we could). Every time when the animals got away, we could pass the hedges again (that was the first thing i tried after the patch release in both cases). So they cannot fix the hedge passing issue without messing the wooden fences. So sad. Is it really THAT serious coding problem?

[Lordcuvious 16]

I'm really glad that I wasn't one of the many who loss horses, surely there must be some form of compensation which isn't sleep bonus!

[Macgregor 46]

I'm really glad that I wasn't one of the many who loss horses, surely there must be some form of compensation which isn't sleep bonus!

No. And don't call me Shirley.

Edited December 26, 2012 by Macgregor