Not saying it is what you want or hope but Link Collector is pretty simple and decent http://www.portablebookmarkmanager.com/
As for ideas, the thing I need most these days is related to text, csv, mysql and conversions in such formats. If by chance anybody cares I will explain further but its probably impossible, impractical and/or solves my unique need only.
This will be a bit left field of a post, but it's hopefully a tip that could advise you sometime. Have you ever taken your desired software/goal to an outsourcing shop like oDesk? Sure there's a range on there of bad to good devs, but to my mind it at least gets some of the doubt out. Being purely ethical, you post your desired project, and when the replies come back, after you weed out a couple of "yes men", the first wave tells you if it is impractical, (little is impossible, only impractical to high degrees!!), and then they don't care if it solves your unique need only. Everyone has unique needs.
An idea I have vaguely pondered is tag-teaming DC guys with the outsource guys. I've talked a bit here and there about "coding lunches". NANY seems to be the once-a-year coding lunch here, bigger than the quick coding snacks. It's a bit of a risk, to commit to a NANY coding lunch.
But let's say you spin it once through a couple of those freelancers. You'll get info back if it's doable or not. Maybe only for a couple hundred bucks.
And here is a bit of where the koala (not as bad as an elephant and far cuter) in the room is. There's a weird part of Internet Culture that likes stuff for "free". (For varying interpretations of free.) I coined the metaphor about coding snacks vs lunches to help separate where someone's genius creating a script in x hours drifts over into a beautiful but time-chewing app.
And I think when people need custom tools do do whatever saves themselves tons of headache but is low reach, be ready to spend a "few" bucks. The way you phrase it is something like "give me feedback and a three hour mock-up". Across X devs, that should be enough to spot some of the worst pitfalls, then you go to your secret DC enclave armed with 3rd level questions, rather than the opener.
You get to say stuff like (made up gibberish coming) "So I posted my idea, two wanted to do it in C++, one wanted reg C, one wanted Python. Three of them said X, Y, and H are a problem. The fourth agreed H was a problem but he had A, B, and Q concerns. Here's three mockups, two in C++, the C guy bowed out, and one in Python. So what do you guys think?"