Bug fixes should be invisible to the end user unless your initial security was so poor it requires fundamental change. Forum software should have been stable 15 years ago, and require nothing more than maintenance releases. I like software that works. I have programs on my phone that were made when froyo was the new hottness. They work fine, and "material design" wouldn't make them work any better. I also use xfce on my desktop. One of the biggest reasons I like it is development is glacial. If they aren't changing it, they aren't fucking it up; like this new forum software. I'm typing this on my phone with the cramped portrait layout csuse they fucked up landscape and I can't see what I'm typing.
Sometimes major framework overhauls require some significant UI changes in order to enable features "downstream."
I'd actually be worried about using anything that maintains glacial development cycles. While some software, like xfce, may be able to get away with it, many software packages cannot. *nix is a funny beast because significant overhauls of something like the kernel can occur without necessitating significant changes downwind. But other times that is also not even true, because significant security developments in the kernel stack may require downstream developers to significantly alter their frameworks to incorporate the new modern security features.
Software that appears stale may just be a preference for the old and familiar, and some software like xfce fit that bill, making changes as they need but otherwise keeping the UI the same.
But your examples like Froyo-era applications - if they still have UI that fits the Froyo style guides, then I reckon there hasn't been any further development in the modern era. Granted, for some software that isn't necessary, but that also indicates the software has been basically abandoned and there could be glaring security holes or system-breaking bugs that become more glaringly obvious with each OS update. Sometimes these holes aren't things most end users need to worry about, but I'd never vouch for that notion.
The entire software development stack is fluid, so anything that stands still should rightly be considered suspect. There are exemptions of course, like xfce. And that helps prove that UI doesn't always need to change, but sometimes the UI change is crucial for user base growth because stagnancy is frowned upon, even if that's mostly an unconscious attitude. Newness = the hot newness. Stale = old, buggy, a security and/or privacy nightmare. And sometimes that UI change is necessary simply to be able to adopt new features, and adding features isn't something that should be frowned upon.
Now when it comes to forum software specifically, they are most definitely NOT complete/done. Anything web-facing is never done. Frameworks have holes, services have holes, etc. At the very least, development is an on-going cycle of patching old holes, and then often patching the new holes opened by either the patching of old holes or introduced via underlying framework patches. Development is also always changing, as the programming languages themselves are always (slowly) changing, adding new features, fixing buggy old features, fixing security holes, introducing more efficient algorithms and CPU architectural changes, new memory techniques, etc. All of this then necessitates that software that uses that code then needs their own updates, and sometimes significant changes in order to take advantage of newer techniques and new features. Which then means a new product version for end users. And, back to the human psyche, that's usually a good time to utilize the modern styleguides and interface frameworks, and that is partly just to keep up with whatever the style trend is at that time, and also to prettify the new package to help convince end users the upgrade is desirable. Most end users don't care about updating to be compliant with this or that, or including new framework changes now so that new features are easier to introduce later, or fixing this security hole or this bug that only those on the backend need worry about, etc. The newness needs window dressing to win over the end users.
Yes, I imagine sometimes some changes are introduced solely because developers are bored and want to sell a new version (or the corporate motivations dictate the developers), and mayhap on occasion even the new features are simply a means to sell a new version and try to convince people its a necessary purchase, but often there's a host of underlying reasons that kickstarts the development in the first place.