There is a particular kind of optimism that I believe only software developers possess. It is the optimism that says, \u201cThis should be a fairly straightforward release,\u201d usually followed up by, \u201cIt\u2019ll probably take a week or two to build.\u201d<\/p>\n\n\n\n
This optimism is always wrong, and I never learn. But it is kind of funny in hindsight. The same kind of funny how one\u2019s old Facebook posts are funny.<\/p>\n\n\n\n
It is not merely wrong in the way a weather forecast is wrong, where you packed the wrong jacket and now your elbows are slightly damp. It is wrong in the way an ancient map is wrong, where the area marked \u201ca great shortcut\u201d turns out to contain dragons, a swamp, and a surprisingly well-maintained legacy installer from 2009 that stores its data in a location no civilized program should ever visit.<\/p>\n\n\n\n
Uninstalr version 3.1 started with what looked like a modest set of goals: Make the app start faster. Use less memory. Detect apps that start automatically with Windows. Improve overall accuracy.<\/p>\n\n\n\n
Simple, right?<\/p>\n\n\n\n
Yes. In the exact same sense that \u201cjust build a small bridge\u201d is simple, provided one ignores rivers, gravity, corrosion, permits, budgets, physics, and the fact that somebody once buried a pipe exactly where the foundation needs to go.<\/p>\n\n\n\n
Uninstalr 3.1 took me three months to develop. During that time, I made 139 test builds and 14 release candidate versions. A release candidate, for those lucky enough to live outside this particular circle of software-development purgatory, is a version that is supposed to be almost ready to ship. It is called a candidate because it is auditioning for the role of \u201cactual release.\u201d In my case, thirteen of them did not get the part.<\/p>\n\n\n\n
This release was not easy. It was educational, though, in the same way falling down stairs teaches you about architecture.<\/p>\n\n\n\n
Speaking of architecture, there is a slightly related topic I want to address: AI.<\/p>\n\n\n\n
As you have perhaps noticed, there has been a little bit of noise about those two letters lately. It seems a lot of people are releasing software these days that is built entirely by AI, pushing new versions out at breakneck speed.<\/p>\n\n\n\n
I am not one of those people. In fact, I just published the official AI Policy<\/a> on the main company website. It is very simple: Humans made this software, not AI<\/strong>.<\/p>\n\n\n\n This does mean it can take me three months to build an update for Uninstalr. But it also means there is a human level of care and quality control in that new version.<\/p>\n\n\n\n Declaring that I don\u2019t rely on AI agents or whatever the latest buzz words are, doesn\u2019t mean I think my software is perfect. If you look at the Changelog<\/a>, you will see a few known bugs even in the new version. Clearly human built software doesn\u2019t mean it\u2019s perfect and bug-free. However, rigorous testing means those bugs are detected and documented and maybe even fixed in the future.<\/p>\n\n\n\n As you probably know, Uninstalr is free software. It has no nag screens, and everyone can use it as they wish. To support development, I run a Patreon<\/a> page where I offer special perks like early access to beta and preview versions.<\/p>\n\n\n\n If you are one of those amazing people supporting me on Patreon, you might want to double-check that your Uninstalr 3.1 is showing version number 3.1.0.616<\/strong>. Because there were many preview versions and release candidates, all called \u201c3.1\u201d, which makes things a bit confusing.<\/p>\n\n\n\n If you are unsure whether you have the official version 3.1, open the About window of Uninstalr and ensure it says you are running build 616. Or verify from the executable file\u2019s properties that it is version 3.1.0.616. That is the actual, final final version 3.1.<\/p>\n\n\n\nUninstalr 3.1 “Final Final”<\/h3>\n\n\n\n