My last post about the temp files was to point out that, while Disk Cleanup normally gets the temp files, quite a lot of Windows software does have "memory leaks" that can fill the temp space Windows is allowed to use with stuff that Disk Cleanup won't necessarily find. And the maximum temp space is only a small part of the free space on your drive. (My machine, which I overloaded, has 90 GB free space on C:, and 1024 MB RAM, but ran out of usable temp space.)

Most of the temp files that Disk Cleanup might miss will be cleared when you reboot, but occasionally a few are missed even after reboot. Usually the ones that turn into "dead" temps are the ones in the folder with the document or other file you're working on, so they can be scattered anywhere on your drive if you have several "working folders." Most of the other temps are likely to be in Windows System or in Office folders or in one of several "TEMP" folders.

I ran Disk Cleanup immediately when the icons "went fuzzy."

The 45 MB of temps that I manually deleted from my document folder, after Disk Cleanup "missed them," cleared enough space for my desktop to operate normally, as indicated in the previous post.

After posting, a full search of the drive found an additional 60 or 80 MB of temp files that should have closed when I exited Word, but that hadn't (along with quite a lot of other temps stuff that "looked normal.")

A reboot cleared most of the other stuff, but even after the reboot I found a half dozen old-dead-trash temp files from previous tasks - some several weeks old - that had been missed by multiple Disk Cleanup and reboot passes. A couple had survived at least 2 defrags.

Disk Cleanup is the "normal maintenance," and is usually sufficient.
An occasional reboot is "second echelon maintenance," but should be done occasionally just "on principle" or whenever something doesn't "look or feel right."
A manual search for "lost" temp files is sort of "minor surgery," that shouldn't be needed very often; but that is "a good thing to do" anytime there's a significant problem.

Of course, none of this apparently helps with the "Sorcha disease," but may be useful as "general info for future reference."

John