Factorio Release Cycle

[Stevetrov]

I am a team leader on a software dev project that is similar in size to factorio (I have 10 people working for me).

I was wondering why factorio has such a long dev cycle (0.14.0 was released in Aug 2016)? I know when 0.15 hits it will have a long list features a lot of which have been ready for a while.

Last year I moved my team to a monthly release cycle, and with the help of automated testing we have been able to provide new functionality monthly to our customers that has been received very well. Because the changes have been more localised we have introduced fewer bugs. Of course some changes take months of development, so not every dev contributes to each release.

Do you guys find that a long dev cycle makes it easier to get the work done without the distraction of releases? In our case we found regular releases kept things fresh and stopped people from getting bogged down as well as making our customers love us!

For factorio a continuous dev approach has the advantage of always giving people new toys to play with to keep them hooked.

Having said this I am very impressed with the quality of the game.

[ssilk]

May I ask what game?

And between 0.13 and 0.14 was only a couple of weeks. I think the reason for this long cycle is that they want to come out of beta state.

[Grubsnik]

Also keep in mind that we are currently playing with 0.14.22, not 0.14.0. So there has in fact been a bunch of releases since 0.14 went live.
They just have a fixed list of features they want to have implemented for 0.15.

You might ask why they didn't just release those different features as they got ready on the way, but I think that is a simple matter of generating attention. Everyone is hyped for 0.15 right now. If we had been getting small incremental additions every month, we would never get to this level. And it's when we are at this level we start trying to tell our friends to come join the fun.

[SaintFlow]

Also keep in mind that we are currently playing with 0.14.22, not 0.14.0. So there has in fact been a bunch of releases since 0.14 went live.
They just have a fixed list of features they want to have implemented for 0.15.

You might ask why they didn't just release those different features as they got ready on the way, but I think that is a simple matter of generating attention. Everyone is hyped for 0.15 right now. If we had been getting small incremental additions every month, we would never get to this level. And it's when we are at this level we start trying to tell our friends to come join the fun.

Well to be perfectly honest, I have been there. In January. Told friends that we shall get together with a bigger group this time around soon in February and this and that.

Jump to now. It's end of March, months have passed, many new games have come out, life has changed for some and to be blunt, they just dont really care anymore. They are already making fun of me even when I just mention Factorio 0.15 because it takes so long and still gets delayed. They are just over it pretty much which is kinda sad. I can't blame them, but eventually, hype passes and indifference kicks in. Not for the diehard fans, but for a bigger group of people which had interest, but just moved on at some point. Too many fun things to play around with these days. However, as the Niche Factorio is in does not really have competitors, it won't matter in the end. As you don't have any alternatives if you want to have this kind of experience. Oh and IF it would be about building hype (which it is not i am pretty sure), you'd want that for 1.0, not for 0.8

I think they bit off too much for 0.15, but that's how it's gonna be!

[Hannu]

In my opinion Factorio's update cycle has been very good. It is good that there are not major updates monthly, because it takes weeks before all mods are updated. I appreciate also high quality over rushing to unrealistic deadlines.

[OdinYggd]

Factorio's experimental releases are better quality than Bethesda's finished release games by a huge margin.

Many players have never even seen an actual bug, let alone a gamebreaking one.

Just be patient waiting for releases. These things take time. Especially graphics- coding it is one thing, graphics is a whole new set of complications and tooth-pulling grindwork to do.

[Ingress]

I too would like to know what game? I am always looking for new and exciting

[Stevetrov]

May I ask what game?

Sorry if my post was misleading I am not a games developer, the software my team develops is an internal application that is not publicly available.

[Loewchen]

Most of the features in .15 break existing mods, so releasing new features once they are ready would kill moding as we know it.

[onebit]

Most of the features in .15 break existing mods, so releasing new features once they are ready would kill moding as we know it.

I think the large releases is worse for mods. When 0.15 comes mods will break. The modder fixes them with a patch. But now the mod doesn't work on 0.14. And actually this is 0.15.0 and we have 6 months before 0.15 stable. So it leaves mods in a weird place. The mod author has to support two versions for quite a while.

If instead there were a small, stable change released every month the change would be small and less stuff would break. More people would switch and the weird interim period would be shorter.

What if it went like this? I think some of these don't really break API.

Release 0.15 (april): 0.15 science + belt optimizations
Release 0.16 (may): Tankers + HD texture support (not all graphics have hd, but the support is there)
Release 0.17 (june): New steam engines + new HD textures
Release 0.18 (july): Nuclear + new HD textures

It's probably impossible to rework 0.15 to release it in chunks now, but the dev team should consider it moving forward.

What's weird to me is that when Factorio was in the 0.14 experimental phase we got releases all the time with little things. But then things halt and there is nothing for a long time. Just stay in experimental mode all the time

The cause is probably team members working on features that won't be released for a while all on the same branch so everything gets tangled up. There's two problems here. Why work on stuff so far in advance? And #2 isolate it in its own branch if it really has to be done. This way you can scrape together all these little things and make a release out of it.

I apologize for backseat driving. I just want features! It's mildly annoying that 0.15 science, which is essentially recipe changes, is held back.

[Stevetrov]

Most of the features in .15 break existing mods, so releasing new features once they are ready would kill moding as we know it.

I understand the desire to not break mods so that makes sense. But couldn't you release the non mod breaking features when they are ready in monthly updates.

In my team have three types of release, major, minor and bug fix. A major release is something that fundamentally changes the system, a minor release is some thing that is more on the outskirts that is unlikely to break anything else and a bug fix release is self explanatory.

For factorio you could have major releases that add major new content and have the potential to break mods, and minor releases that don't break mods but add in new functionality.

Anyway just a suggestion based on my experience and as I said I am nit a games developer so the pirorities are different.

[SaintFlow]

Most of the features in .15 break existing mods, so releasing new features once they are ready would kill moding as we know it.

well, even if that was the case (i dont think it necessarily would be), I personally take new updates from the devs over a higher number of updated mods ANY DAY of the week. Modding will become more interesting for me when the games scope is fully known and in the game.

[Keeper]

Once you add mods to a game, beta or not the thought of an update is terrifying. Yes I'm looking forward to it, but it's gonna be annoying rounding up all the mods again to continue with my current game. Games like KSP made me restart several times, and now I can't be bothered starting again again. Minecraft had a phase of updates now I think you can pick your version, but still makes modding an unneccesery nightmare. Haven't bothered since. Tbh good updates are better than more updates. Well played factorio people

[Deadly-Bagel]

I don't think Factorio intends to have a long development cycle, more just that they aim for some big features (nuclear, belt optimisations, liquid wagons) for a release. In themselves they can trigger a bunch of other developments, like needing to upgrade pipes to HD, which then means all pipe-related machines ideally need updating as well.

Not everyone is working on these big developments so a lot of smaller stuff gets done too by other devs (map updates, alert beacon, misc optimisations) which to many players can be as exciting as the main features.

Then there's context. Research on the whole is being updated, not only the recipes but the research tree is receiving an overhaul with a lot of redundant stuff being removed and new researches being added. This is the intended context for the new science pack recipes, what is there to gain by changing one but not the other?

[SaintFlow]

... This is the intended context for the new science pack recipes, what is there to gain by changing one but not the other?

I think you are absolutely right, If you update something it definitely makes sense to update all the connected things as well.

However what I'm saying is that if you go down that road, maybe not shoot for so many different things at once if they don't have anything to do with each other. Rather give players one updated part or maybe a few at a time, than shooting for many different things at once, making them wait for more than half a year for anything to happen really.

There might be times where it is not avoidable, but in general, it should be feasible.

[silverkitty23]

May I ask what game?

Sorry if my post was misleading I am not a games developer, the software my team develops is an internal application that is not publicly available.

Does your software need as many art assets as Factorio? Does it need to be "moddable" by the community? How many different machine configurations do you support?

[Stevetrov]

May I ask what game?

Sorry if my post was misleading I am not a games developer, the software my team develops is an internal application that is not publicly available.

Does your software need as many art assets as Factorio? Does it need to be "moddable" by the community? How many different machine configurations do you support?

No limited artistic assets and it's not moddable. However, we do support a number of different platforms.

As I have said the moding issue is a major concern that needs to be considered.

My project is clearly different and don't get me wrong I thing wube do an amazing job, but that doesn't mean they couldn't do a better job. None of us are perfect.

[Deadly-Bagel]

Now you sound like my dad. "99/100? Well it's amazing but you could do better."

[OdinYggd]

Now you sound like my dad. "99/100? Well it's amazing but you could do better."

Even at 100% you can still do better.

During the Space Shuttle's useful life the SSME engines improved to be even more powerful than they had been at first. But rather than creating a conversion factor from original spec to current spec, they just re-scaled the throttle control to have a maximum of 102%.

Go for throttle up.

[daniel34]

During the Space Shuttle's useful life the SSME engines improved to be even more powerful than they had been at first. But rather than creating a conversion factor from original spec to current spec, they just re-scaled the throttle control to have a maximum of 102%.

From what I've read it's not unusual for rocket engines to surpass the specs, in these cases the throttle goes above 100 percent. Sometimes engines also have a power-efficiency sweet spot which is referred to as 100%, but they are able to go above, although it's not as efficient or might shorten the lifespan of the engine.

“We have ignition… and we have liftoff of the Antares Orb 3 mission to bring Cygnus on its third (resupply) mission to the ISS,” said NASA’s launch commentator as the Antares lifted off. “Main engines at 108 percent, attitude nominal.”

At roughly that moment, the exhaust plume suddenly intensified, followed an instant later by a detonation of some sort at the base of the rocket. The booster, engulfed in flames, then slowly fell back to Earth, exploding in a titanic conflagration as its remaining liquid oxygen and refined kerosene propellants ignited in a huge fireball.