Hi people! Apologies for yet another super late post - here's the March Fossil Sweeper "mid-month" status update!

Over the period that this post covers, mostly spend my time on more non-development housekeeping, this time focusing on adding annotations to all progress update videos on Spectra.video. I also made a start on getting back into globe asset production.

Map

Back when I last updated the globe in test builds, I'd mentioned how my plan was to do terrain deformation shape keys for the entire globe back to the beginning of the Cretaceous, but I eventually decided to just ship shape keys for Australia an Antarctica since Blender was wigging out and giving unpredictable results while working as I added more and more shape keys on top of each other. Contract work and other commitments in the meantime guided me towards focusing on smaller work items in the intervening time. Past few months aside, I feel like I've still been able to maintain a reasonable level of productive output on worthwhile stuff that the game will benefit from, but the globe really does feel like this blocker that's preventing me from making meaningful strides towards release, and that's brought a degree of frustration on top of an already difficult year.

I think at some point if things don't go well, I will need to address the difficult question of whether the animated glove is essential for the project (obviously it's not, but it is a key feature in some respects, and cutting it isn't something I want to do unless absolutely necessary), but I'm not there yet, and as I try to wrestle get back on top of things after the interruptions that have arisen over the past several months, I find myself eager to attempt to push through and establish *some* kind of workflow that will allow me to move ahead.

When I gave up and shipped the globe with partial animations, I neglected to keep notes on what I'd tried and where I'd left things, so I spent time in March reviewing old files and working out which version of Blender would be best to continue with. I'd done experiments in Blender 3, 4, and 5 last year to see whether the shape key stability issues I'd experienced, but since they didn't and since I wrote all my existing Blender addons for Fossil Sweeper production for Blender 2.79, it makes sense at the moment to stick with that.

I've made a start on getting the continental drift animations in, and I've been thinking about alternative workflows and what I can do to make things more robust/stable while I work. Ultimately it's going to boil down to a slower, more labour intensive process, which is annoying, but if that's the cost for avoiding the wasted effort I was seeing last year, then it's a net gain.

Museum

While recording footage for the previous month's video, I did a grep across my code for the line where I was duplicating light data so that I could record the "re-enactment," and spotted that there was another place where I was doing a "shallow" duplicate of template light data when research ends and lights are added to the player's inventory. I don't think this would have surfaced as an issue during museum editing at the moment, but things might look different in the future, and fixing it now pre-empts possible problems later.

Misc

Once again, most of my time has gone again to doing something other than working directly on the project. This time, I worked through all of the historic progress update videos, adding annotations. PeerTube annotations are more or less equivalent to YouTube's "cards" - that's what I'd point to when saying "If you haven't heard of the project before, you can find an overview here." When I first migrated across to Spectra.video, the annotations plugin wasn't behaving. With that fixed (thanks to Spectra.video's admins for sorting that out), I was able go back and add them all in, along with some corrections and additions for oversights and mistakes that I'd made when originally posting videos on YouTube.

With that, the devlog videos on Spectra are no longer lacking a feature that was present on YouTube. Hooray.

While writing the script for this month's video, I realised that I had forgotten to write about refactoring the chat integration back end in the previous post. I haven't done anything that exposes new functionality yet, but I've separated out the IRCHandler class into an ChatHandler class and a ChatBackendIRC class. There's also a ChatBackend class that ChatBackendIRC extends (in another language, I'd have implemented this as an interface) with the idea that I might be able to support other protocols in the future, such as XMPP, which is what PeerTube's livestream chat plugin uses.

And that's about it for this post. Thanks as always for reading and for your patience, if you check out test builds and have any feedback, I'd love to hear your thoughts!