A long time ago, I used Quicken and Quickbooks.... but Intuit became evil, in various ways, the least of which may be how they make it really difficult to avoid regular (costly) upgrades.
Harena used to use MS Money, and liked it -- but then each "improvement" took away good features and it eventually became unusable.
Meanwhile, I wrote myself a checkbook-balancing program in MS Access 97, which (though limited because of Access's limitations) worked pretty well. If it weren't for the limitations, I'd be tempted to try to use that again: Access 97 started working ok under WINE several years ago (for awhile, it was the only part of Microsoft Office 97 Professional that didn't), and I have backups of the .mdb file with the code in it -- if its limitations didn't prevent me from adding some features I will need.
I've also tried several open-source programs, but they all have irritations that make them fall short of what I wrote in Access.
I started porting the Access program to Ferreteria, and got it partly working, but it got to the point where I realized it really needs those features that the Access-imposed limitations in its design don't make possible. I've got a better design sketched out and am hoping to start on that soonish.
The key tool I'll be using for the rewrite is Wikcess. I originally came up with that as a way of supporting wiki operations but with semantic data built in, and not dependent on markup -- but it also seems it really ought to work well for financial data, where you really want some flexibility in how the data is structured and what pieces of information each transaction includes. Some transactions, all you have is one piece of the story (a date, an amount, a local account, a remote account) -- while for others, you may have a lot more (e.g. a "split" with each piece being assigned to different categories, maybe a conformation number, notes, scans of receipts or work orders, and so on).