Friday Facts #143 - Nat Punching Again

[y.petremann]

I would rather they had released an unstable version june 1 clearly marked experimental for modder only and then work with them for the month to release a general public version at the end of the month. [ ... ] But I know it's really only the fault of those whiny gamer who have no idea how development work and want to play the latest version no mater what and then go whine on forum how the game is bugged and they are getting ripped and the dev are bads people who molest children, while meanwhile they never ever submitted a bug report ever in their miserable life.

Fully agreed.

I'm a developer too and I prefer transparence instead. I would rather they had released an unstable version june 1 clearly marked experimental for modder only and then work with them for the month to release a general public version at the end of the month.
Now instead it's 0.11->0.12 all over again. They broke half the mod and it's gonna take months again to get those working, since probably half the mod developer moved on to other games.

I agree that their release date handling was really bad. This "soon, maybe soon, really soon ... oh, 2 more weeks" killed the hype and anticipation IMO (I realize this is not the critique you brought forward, though).

But, on the other hand, what would be the point of releasing a pre-release version? Modders may have a lower bar in case of crashes etc, but the factorio devs would need to work on bug reports for bugs they'll find & fix while playtesting anyway. Some people will download it and complain. They have even more work (I'm pretty sure they're so quiet now because they're working hard to release 0.13). Also, people will play either vanilla at first so it doesn't matter if mods are initially available (small mods mostly will be converted rather fast or never) or they play a full conversion, in which case the probably don't care about the new version. To be honest, the early acces/"alpha" is actually a pre-release, but people just ignore it by now so you can't just throw out a buggy beta and tell them its a beta - thanks greenlight and kickstarter.

So, concluding: Yes, their date handling was bad, but that doesn't make them bad devs.

On some point I agree and some other I don't, but I prefer to expose examples of diferent developement cycles as I see them for the debate:

• Wube software with Factorio : Each version is developed internally, then versions are released as an experimental version, and at a time somes get the stable state.
  + Weekly fixed time blog post permit to have informations about the developement of the game
  + We are more encouraged to play when a new release is done
  + Developers have a high level of caring on factorio, bugs usually gets resolved rapidly (almost all bugs I've reported were solved less than 12h after report)
  - There is almost nothing that diferentiate stable and experimental versions
  - Mod developement is slow because not having access to api before the actual release
  - When previous release get stable, we already get information about next version that we want and prevent us to play actual version.
• Keen Software with Space Engineer : They put the level very high since they released his code on github and allow the modification of the game and sugestion, they started to do it in a way as "We'll see if we can trust our player to do awesome things for our game and not stealing our code, building pirated versions ...", and as for now they continue to do it ... Also they do weekly updates with an update video.
  + Permit to fasten developement with feedback of the community
  + Permit to fasten mod developement
  + Recent change with stable and development branch allow for more stable games
  - It's longuer to see "big features" because of a fest developement cycle (based on one week)
  - loss of interest due to "there is endlessly again and again each week something new ..."
  - There is still a risk of code stealing
• Chucklefish with Starbound : They have 3 developement cycle : stable, unstable (which they release from times to times, not so much these times, it's like the factorio experimental branch, version get released before stable version) and nightly which contain the latest functionnality and asset mostly for modders or those wanting to see new functionnality, but beware, the code is compiled as it is, they don't really care the game is playable, it can be started and that the most they "guarantee", They do blog post about game developement.
  + Each blog post permit to keep the hype with various subjects about what the player could do (Wube software take a more technical path in them)
  + Fasten mod developement
  - Blog post are not scheduled to specific day
  - Very long developement cycles
• Mojang with Minecraft : The developement of a version follow theses steps : internal developement, weekly snapshot, prerelease, release, bugfixes
  + Snapshots usually gets secret features
  + Versions are correctly separated by some way to identifying them (16w24c is a snapshot, 1.10-pre3 is a pre-release, 1.10.5 if the fifth after release update)
  - Due to absence of modding api, mod developement need the implementation of unofficial APIs (mcpatcher, forge ...) mod release are far far away (we are on version 1.10 and most mods are on 1.7-1.8 which is almost a 2 years old version)
  - Developers have a low level of caring on minecraft, bugs usually gets resolved slowly (almost all bugs I've reported with a valid solution were solved more than 6 month after report with a record with 2 years, i've stopped reporting bugs since late 2013 and some of them are still not solved)

And I only put games example.

I don't say that one is better than an other, simply Judge by yourself ...

[cube]

I´m glad we have a somewhat fixed release date!

About NAT punching: Will it work with IPv6 connected computers without real IPv4 adresses? (Because the shitty internet provider of these unlucky people has "due-stack lite" implemented for pseudo IPv4 )

You guys are doing great work, i wish you good luck with the race to the release date!

IPv6 works even in 0.12. We're using it in the office without any problems.

[Nova]

I agree that their release date handling was really bad. This "soon, maybe soon, really soon ... oh, 2 more weeks" killed the hype and anticipation IMO (I realize this is not the critique you brought forward, though).

But, on the other hand, what would be the point of releasing a pre-release version? Modders may have a lower bar in case of crashes etc, but the factorio devs would need to work on bug reports for bugs they'll find & fix while playtesting anyway. Some people will download it and complain. They have even more work (I'm pretty sure they're so quiet now because they're working hard to release 0.13). Also, people will play either vanilla at first so it doesn't matter if mods are initially available (small mods mostly will be converted rather fast or never) or they play a full conversion, in which case the probably don't care about the new version. To be honest, the early acces/"alpha" is actually a pre-release, but people just ignore it by now so you can't just throw out a buggy beta and tell them its a beta - thanks greenlight and kickstarter.

So, concluding: Yes, their date handling was bad, but that doesn't make them bad devs.

There are only two alternatives if Wube wants to release their game only without much bugs: Not saying anything at all, or giving the release date as guess with changes if new informations turn up. I say the first is bad, you say the last is bad. It's hard to do anything which satisfies everyone.

[ske]

Musical suggestion for developers: https://www.youtube.com/watch?v=6J3mYaAZJ68&t=1m13s

[Sebb767]

There are only two alternatives if Wube wants to release their game only without much bugs: Not saying anything at all, or giving the release date as guess with changes if new informations turn up. I say the first is bad, you say the last is bad. It's hard to do anything which satisfies everyone.

Disclaimer: This is a personal opinion based on a personal observation. I do not claim to be more intelligent or qualified than them.

Well, the way they handled it as a whole was bad. From my (maybe biased) view, it went like this:
- May: People read they're starting playtesting, hype goes up. The only mention the date could not be hold IIRC was kovarex saying "2 weeks seem uncomfortably close" in a FFF.
- 1.6.: People are awaiting the release. But nothing happens, no word from the devs. Just a spike in refreshes. Tension rises.
- 3.6.: After several days of F5-breaking I read "I guess that not so many people really expected ...". Well, I kinda did. From what I read, I was not the only one. But a few days are worth it, so lets wait :)
- 10.6.: After yet another week, they hint "its soon to be released". 10 days are a little long for my can't-await attention span. Also, that FFF wasn't even much of read.
- Today: A week of sometimes hopefully looking at the page, they push it even further away. If you read the comments here, its not just me. A fixated release is awesome, but my personal anticipation at least is too overstretched bynow.

Criticizing is always easy. But there are a few things to do better:
- If you know there's a hype and you're probably not going hit that date, you should add more time ASAP. This will not prevent, but drastically slow hype/anticipation starvation.
- Don't say soon for ages. Keeping the tension up for too long will in the long run destroy it. At least give a hint if you're not close.
- Set your release dates very conservative. An earlier release can be really awesome and on time isn't wrong, either.
- Lastly, a deadline like this might be a nice way to speed up your development, but those boosts usually don't last long and aren't really worth it if you consider disappointed fans and followers.

So, setting a fixed date now was the right thing, but at least for me and a few other disappointed around here it's a little late. Actually three weeks late, but a "heads up, not this week" would've surely prevent a few disappointed visits of the download page and I probably still couldn't await it.

[silverkitty23]

- If you know there's a hype and you're probably not going hit that date, you should add more time ASAP. This will not prevent, but drastically slow hype/anticipation starvation.
- Don't say soon for ages. Keeping the tension up for too long will in the long run destroy it. At least give a hint if you're not close.

See, I may be weird, and I wasn't there, but I think a month ago the developers themselves DID think it was going to be soon. They had the NAT punching feature scheduled for 0.14. They didn't think they were going to implement it... and they were busily playtesting all the features that they had planned for 0.13.

I think it came as a surprise to them that they had to rearrange the schedule and add a couple weeks to implement an unplanned feature at the last minute. A month ago, they didn't know they were doing NAT punching in 0.13, so their date seemed reasonable to them. Then they discover "oh, crap, we have to do this feature in 0.13 we didn't even plan for, or else drop the match-making server..."

[testasdf]

It was always said to be a goal, not a promise (nor a deadline). But words spread and that part got omitted. Usually a goal contributes 1/6 to a more properly estimated time.

But yeah, not everyone knows software development enough to recognize the difference. Maybe the listed tasks are all done, but then there are the minor details and unexpected problems...

[TRauMa]

but the factorio devs would need to work on bug reports for bugs they'll find & fix while playtesting anyway.

It's called a git commit message and that can all be automated these days, if they are doing thing right.
And they have to waste that time anyway, if they want an organised repository.

And some kind of dev you seem to be. The problem is not closing bugs in a tracker, it's handling duplicates and incomplete reports and the general fallout from releasing unfinished patches, no matter how much you mark them as incomplete. People will use them, even if they are not the target group and have no business doing so, and your support team will have to handle their problems, even if they shouldn't have to. It's the golden rule of development, if you release it they will find bugs. Releasing anything is a big deal, particularly to an audience as large and full of anticipation as this, and your hand waving all of this away doesn't instil much confidence in me that you really know what all of this entails.

All in all I'm slightly taken back by the reaction to this FFF. We get weekly posts with pretty pictures and stuff, we can discuss game design questions with the devs, they talk about what they're up to in detail and don't do patch drops from orbit like so many others. Think of all the things we'll get in 0.13 - compare it to other games you play and their time between patches. And their price tag.

I get it, I'd have liked to play around with the new features three weeks ago, too. But the alternative to vague dates that get pushed back are either unfinished patches that make no one happy and are a PITA for the devs or no information about pending updates at all. Personally, I'd rather have it as it is.

One thing you could argue for is making prepatches available to modders and a closed group of testers, somewhat like KSP does it. But if they're any indication for how well game devs fare with this, I don't think the effort is well spent there.

Sorry, BTW, if any of this sounds aggressive or cranky, it's really not meant to be - I should be in bed instead of stalking factorio news, and sometimes my tone slips in this state.

[xxninjaxx25]

I can wait 10 more days just means I get to play factorio more now since another game got delayed a couple of months!

[brunzenstein]

The current alpha Factorio version is very playable - much more enjoyable then in fact all similar /competitors finished games.
So, the later 0.13 comes (e.g. full of new cool features, more polished and bug free) the better.

[malecord]

My 2 cents...

The release policy is somewhat secondary for me atm. Yeah it's annoying to have these big mod breaker releases. Yet this is an early access game so I don't feel entitled to be disappointed. I definitely would if this were a fully released game.

As for the transparency, I feel there has been a little disrespect toward the users. Imho it makes perfectly sense that the main dev goes on holidays before release: the weeks immediately after release are (notoriously) the worst so it's better to be at full capacity. Yet this schedule clearly contradicts the long standing 1st of June statement and the "good chances next week" statements that followed after. Hype is OK, but it is better built by leaking carefully chosen words and pictures. Not everybody appreciate cheating on dates.

Well... Thanks for this FFF. Waiting for the (maybe) 27th then.

[brunzenstein]

As for the transparency, I feel there has been a little disrespect toward the users.

There are horses and there are riders - and there are bystanders.
A rider (even he must respect out of pure self interest in a way his horse) as a free man can decide always whatever he wants to do, e.g. finds proper to lead his horses to. The horse itself has, even on its own powerful, a minimal saying where its going, and thats perfect with me...
Thats the way capitalism works in a nutshell.

[Koub]

People, do you prefer to play a game developped by :
1) A company that's expert in communication, hype managing, and stuff, but makes poor games
2) A company that's expert in making an awesome game, but doesn't master the complex mechanisms of communication, buzz making, hypr building and stuff ?

I'm the 2) kind of guy. Honestly, if you're more the 1), what the hell are you doing here ? there are literally thousands of dev studios that hype like pros, know how to communicate, and stuff ... and make bad to very bad games. You should definitely play theirs instead of bombarding goods devs that prefer releasing something playable for the next unstable release.

[brunzenstein]

People, do you prefer to play a game developped by :
1) A company that's expert in communication, hype managing, and stuff, but makes poor games
2) A company that's expert in making an awesome game, but doesn't master the complex mechanisms of communication, buzz making, hypr building and stuff ?

I'm the 2) kind of guy. Honestly, if you're more the 1), what the hell are you doing here ? there are literally thousands of dev studios that hype like pros, know how to communicate, and stuff ... and make bad to very bad games. You should definitely play theirs instead of bombarding goods devs that prefer releasing something playable for the next unstable release.

+1 -well said

[ChurchOrganist]

Ewww.. should have knew they would do like every shit developer and release this at the last possible day of the month they say they will release it..
Well now at least we have a date so I don't have to come here everyday to be disappointed again.

Speaking as a member of an Open Source music application development team your sentiment is bang out of order.

You obviously have very little idea about the effort that goes into producing a stable version release - not to mention that always last minute bugs emerge which delay release.

Wube could follow our practice, however, and release nightly builds of the new version with the caveat that they are not guaranteed to be stable. That way mod developers can begin to make sure their mods are compatible before release date.

[binbinhfr]

Wube could follow our practice, however, and release nightly builds of the new version with the caveat that they are not guaranteed to be stable. That way mod developers can begin to make sure their mods are compatible before release date.

That would be nice. I already asked devs for no-stability-warranty beta release for modders preparation (and also additionnal bug hunting), but they said it was not their politic. That's their choice and I respect it even if I do not fully understand it.

[Rseding91]

Wube could follow our practice, however, and release nightly builds of the new version with the caveat that they are not guaranteed to be stable. That way mod developers can begin to make sure their mods are compatible before release date.

That would be nice. I already asked devs for no-stability-warranty beta release for modders preparation (and also additionnal bug hunting), but they said it was not their politic. That's their choice and I respect it even if I do not fully understand it.

0.13, 0.12, 0.11, 0.10... they're all early access versions of the game and everyone playing it is testing it before the official 1.0 release.

[slindenau]

Wube could follow our practice, however, and release nightly builds of the new version with the caveat that they are not guaranteed to be stable. That way mod developers can begin to make sure their mods are compatible before release date.

That would be nice. I already asked devs for no-stability-warranty beta release for modders preparation (and also additionnal bug hunting), but they said it was not their politic. That's their choice and I respect it even if I do not fully understand it.

0.13, 0.12, 0.11, 0.10... they're all early access versions of the game and everyone playing it is testing it before the official 1.0 release.

You know that is not the point, it's about making sure mods work ON RELEASE instead of months afterwards (yes, even early access releases).

You're selling a game in the current state, working mods on each update is the least people may expect.

[tekbot]

on topic for this FF: personally I know how to forward a port and dont gain anything from nat punching. but then again there are still people who use hamachi and other #$☠*@! so evidently there is a need for it.

on topic of the off-topic aka 0.13: #$☠*@!

People, do you prefer to play a game developped by :
1) A company that's expert in communication, hype managing, and stuff, but makes poor games
2) A company that's expert in making an awesome game, but doesn't master the complex mechanisms of communication, buzz making, hypr building and stuff ?

I'm the 2) kind of guy. Honestly, if you're more the 1), what the hell are you doing here ? there are literally thousands of dev studios that hype like pros, know how to communicate, and stuff ... and make bad to very bad games. You should definitely play theirs instead of bombarding goods devs that prefer releasing something playable for the next unstable release.

+1 -well said

I cannot argue with the choice given those options, but find the example lacking. why should good marketing/PR cause bad games ? of course it stands out when the two correlate and I wont discuss if factorio is a good game in a forum of hyped /disappointed fans, but we seem to have enjoyed it so far. myself included.

I also think "mastering" PR is entirely different from what was asked here.(ie the things can seem off / not add up, and it sounded like 0.13 was so much closer than it now is).
Yes we prefer a decent game to a timely release, but no i dont want to be put off 2 or 3 times when it now seems there were always quite some things left to do. as always hindsight is 20/20 and makes it easy to complain.

What i can say however is that to me it seems very evident that the frustration is at an all time high, as evidenced by the number of useless and/or complaining posts (again - this one included). I consider this to be mostly, if not almost entirely, a product of the date shuffling and hopes staying high.

As far as i know nobdy is intentionally decieving anyone. We (horses, riders and maybe even bystanders) are in the same boat on the recieving end of overestimated capabillities/reserves. Everyone believed 0.13 could be finished by first first week first half end? of june.
That sounds like most software Projects I (as a developer) have heard of or taken part In. I take a "realistic" estimate, add several impossiblecases for things to go terribly wrong and double it. The product manager sighs and doubles it again. The CEO says thats too long but he really wants the feature, and its actually really finished some time weeks into the next feature.

TL DR: what malecord said + rambling and anectodal evidence on software guesstimates.

[brunzenstein]

We (horses, riders and maybe even bystanders) are in the same boat on the recieving end of overestimated capabillities/reserves.

Fair enough - but it depends if your calm strolling on the ships bridge in a white shirt deciding the route or sweating in the coal bunker with a dirty shovel in hand taking what you get;-)
Anyhow - the deep frustrations and mangling of poor souls for not getting a game (!) earlier you mentioned becomes absolutely pointless the moment one discovers there is a real life (life is also philosophical a game with a sad end) outside of computer games.
And believe me - there is - have a look!