We have to organize a very long list of manually running tests. Currently we use Word documents, print them out check them off etc. Ugh-ly but works, with problems:

 * Requesting and executing subsets, depending on the actual changes. (i.e. tests 153, 155-157 and 24325 don't need to run if Module X hasn't changed)
 * Branching. Different releases (which are still maintained, up to a point) have different test protocols, but share a lot of contents. Manual copy & paste.
 * Tracking "Test Template" vs. "Test executed" documents. It's a bit old-style office with lots of forms & binders. 
 * Tracking in source control together with the source changes. Ideally, the same way features are copied / moved between branches, i.e. as diffable document that is included in a commit.

Documents need images (usually, hardware setup specs and screenshots of "good" curves), "Pass / Fail" checkboxes and minor formatting. 

I am personally not wholly sold on the "use HTML documents and track the test documents in git", but I agree with the other guys that if it works it's better than Word documents. 

We are open to other solutions, but only with minimal custom implementation efforts ATM. e.g. installing dedicated software and configuring some templates would be great, but we have no resources (nor patience)for "hack something together with XML and stylesheets and a database".


Some corner stones: source control is git, 6 software developers, ~1.5KLoc, one run through full manual test suite takes 4 man-weeks (OOM), but that's rarely run. 

---

Note: *Please don't tell me to test automatically,* thank you but we already do. The manual tests involve integration, visualization and external measurement equipment. In addition, the information produced is so multifold that the discoveries of a engineer who knows the whole system discovers a whole lot more with a single look at a chart.