[2020] I datamined the Factorio Forums to get (almost) all timestamps of the Friday Facts posts.

To expand on what Posila wrote, I want to go a bit more in depth.

Before about FFF-252, the Friday facts was more of a 'reactive' dev blog. We would do some work all week, and then when Thursday/Friday came around I would ask around and see what we had done that week. This worked really well during the 'good old days' where we were adding lots of features, and work progressed at a rapid pace.

However as the development focus started to shift from adding new features to polishing and finishing things, this approach started to not work that well. Very often on a Thursday, we would have 'nothing to show', and the responsibility to 'manufacture' something to show fell to kovarex, Albert, or I. This placed a lot of pressure on us 3, and Friday felt like a weekly guillotine just waiting to drop, until we scrambled to push it back another week (a little bit hyperbole). This resulted in a lot of Fridays us being in the office till 8/9/10pm, and never being able to focus on other things without the worry of "whats going to be for FFF?".

So around half way into the year we have some meetings to recap what's been going on, and during I discussed that the way we approach the FFFs is not working, and its going to collapse sooner or later. So I made the proposal, of reversing the way FFFs as a concept work. Instead of using FFFs to look back and describe what 'just so happened' to be finished or ready in the last week, we changed it so FFFs would plan and somewhat dictate when things should be ready and finished. (Part of this was also that the topics should be ready for me by Thursday evening/First thing Friday, so that I can go home at a reasonable time.)

This proposal was accepted, and part of it, is that I took full responsibility over producing the FFFs. This new approach to FFFs took some time to get used to, and had some learnings along the way, and there was still the occasional Friday evening scramble, or 'manufactured' topic. Soon enough we got quite into the rhythm and the way it started working started to really make things work much smoother. The quality of the topics started improving, more team members were involved, and kovarex and Albert could rest easy (and focus on other work) knowing that FFFs were planned and solved quite well in advance.

Over time the experience we gained with the new FFF process started to restructure the way we approach the entire development process, and the result of that is the switch from larger releases (0.16, 0.17), to the smaller and more iterative release structure (0.18->1.0).

For instance before:

• We work on a feature until its 'done'. Sometimes it is presentable, but still needs a lot more work to be ready for release.
• We show it in FFF. It doesn't matter here if it is ready for release.
• Sometime later, maybe multiple months or longer... it is released and you can play with it.

The problem here is quite obvious looking back. You don't get to play with it for months and months, we don't get feedback for months, and when you do play with it, there is no clear place to discuss the feature.

And now:

• We start work on a feature, and have some rough estimate of when it should be ready.
• We have this in mind during the team meetings, and make a sturdy plan about when we will show the feature in the FFFs.
• When the feature is 'ready for release', only then do we show it in FFF.
• After the FFF, we try to release the feature as soon as possible (which is why we now have a somewhat predictable Tuesday release).

(By feature, I also mean polish, new graphics, sounds, etc.)

Here there are many advantages:

• You just saw the feature in FFF, and now you get to play with it!
• Players are more focused, because the release only has this 1 feature. We get really direct testing and feedback.
• The developer is still very much in the 'zone' around the feature, if there is a bug, he still has all the code in his mind so fixing it is easier.
• If you have feedback, it is obvious where to say it, in the FFF discussion, since it is still alive and very relevant.
• We can close the topic very quickly, after a week or 2 of bugfixing, we don't need to worry about it any more (keeping down technical debt).

So in short, FFFs have changed and grown in the same way we as a team has changed and grown, as well how the status of the game has changed. It is all connected and it is all related to one another. It is clear now that the approach we had during the early days, does not work now as we approach the final days of 1.0 development. We have learned a lot, made some mistakes along the way of course, but we are now in a great place to deliver 1.0 under-budget and ahead of schedule! (there is no budget I just like the phrase).