At the same time, the developer clearly doesn't want people to be encrypting new archives with it, so have removed that functionality.
Removing the software in total would have lead to many mirror sites springing up, most with unproved providence (and a great opportunity for exploits). This method allows an official version to still exist (unmaintained, but still compatible with the current archives), whilst severely restricting new usages (by strongly warning against it and requiring using the dubious mirrors to obtain the software).
That still doesn't seem to answer the point. The differences between 7.1* and the questionable 7.2 are exclusively limited to what was ripped out. In 5 years, 7.1 and 7.2 will likely work precisely the same in terms of decryption. If a glibc update or similar breaks 7.1, it'll likely break 7.2.
Perhaps I wasn't clear enough in my first comment, but fundamentally it seems like a silly argument to make to suggest that distributing a new release for some measure of future proofing readability is necessary. Outside "don't use this, it's broken," I can't really see "this will still work in 5 years' time" as a valid reason.
No mention of why it's supposed to be not secure - it's an open source project so it would be easy to point to a specific vulnerability. All of this shortly after passing audit. There are detailed steps towards switching to supposedly secure closed-source solutions by companies known to be working closely with the NSA.
Also, since when do open source projects suddenly decide they are not as good enough as a closed-source alternative and then stop the project? Are we in danger of seeing a similar message on the homepage of LibreOffice and OpenOffice, declaring that you should switch to Microsoft Office?
I can't even begin to imagine a valid scenario in which something like this would be put up by the developers, with no pressure other than some fatal security flaw about which they just really don't want to talk.