I completed my Ph.D. in 1997 on The Use of Formal Methods for Safety-Critical Systems. Happy to share the findings, I put copies of thesis online, but whilst LaTeX -> dvi -> postscript may have been routine practice for users of UNIX systems running X Windows, it was not very Web friendly for many others (much as Ghostview is good, it doesn't have such a popular appeal!) Finally I've got round to converting to PDFs and have added them to my PhD page. I tried a few years ago, but my initial attempts generated huge files and I couldn't work out why; fortunately ps2pdf.com's settings were far more reasonable and the entire thesis would almost fit on a 1.4MB floppy disk!
As to the subject matter, I was unable to progress the research as I wished; the small group at Kingston University soon petered out and there were relatively few openings elsewhere, so that's when I moved more towards I.T. from computer science (they are very different activities). In the '90s there were growing hopes that formal methods would gain a more general foothold, but when I glance at FM sites now, it seems their use remains very niche; references to LOTOS, the process algebra that I used to model a communications protocol for medical devices, point to materials that are rather old - the World-wide Environment for Learning LOTOS is indicative of this. There are still research activities, typically in compilation, but overall it's a bit surprising and disappointing. Yet given the greater computing power on tap, particularly cloud computing, perhaps this area may yet develop a lot further...?