I would like to raise a question for everyone about what I am describing as fragility in LyX.
I use Lyx as a method of recording my own thinking on various subjects including mathematical analyses, a tool to annotate material (usually academic papers which can not be annotated directly) and to prepare papers for publication. I feel that the fragility I am detecting may be increasing and would then be attributibe directly to LyX. However, I am increasingly using more advanced techniques and these may be the cause of the events which am noting rather than with LyX itself. The number of "strange occurrences" has made me ask if any of you have had any similar experiences, which appear inexplicable but hardly worthy of raising as a particular concern on this forum. Most of these I am attempting to describe have arisen whilst simply editing text. With a single backup file attempts to return to an earlier unaffected version are thwarted. Examples: I was annotating a file using the PDF comments module. The pdfmarkup annotations were working correctly, visible in the pdf output. I edited some text. In subsequent version of this document no markups are visible in the pdf output. I have more than forty similar files all producing the expected annotated results. Tables: With formal tables the order of editing can produce different outcomes. I have worked round this problem by starting a completely new table as the subsequent editing of erroneous tables has produced variable, unexpected and inappropriate results. hyperref: unticking hyperref, after its having been set, produces a number of error messages and output with "hyperrefImplicit mode ON; LaTeX internals redefined" on the first page numbered 1. The rest of the document starts on the page numbered 2! biblatex with biber: Possibly it is to be expected to be more delicate. However, with a new pdf requiring more than 20 seconds to be produced, human error is likely and can lead to an inability to produce any further pdf output. Solution is close down LyX and restart. Questions: Am I alone in these experiences? Does anyone have the same or similar to recorrd? Is there enough information to help the developers determine the nature of the problems and their solutions? Finally, am I missing anything glaringly obvious to you out there? Regards Frank Salter