There is currently some discission in the open source world about jpg using a patented compresion routine so it has the potential to not be around years from now. Its also easier to convert from a loseless format such as tif or png to jpg that to convert a jpg back to a tif or png since you can't restore the lost data in the jpg. Also if you were ever to loose the original document you may need the scanned copy. My scanned documnents are backups for the paper and I keep copies in several places and keep them all in the same format, much simpler than having to convert from loseless on the hdd to lossy on the flash drive to save a few bytes, then if you loose the loseless all you have left is the lossy. I work in IT and have to be aware of DR type of plans so I look for the worst case. My flash drive is a last ditch backup of my absolute most important docs so if soimeone steals the laptop, the house burns down melting the backup cd's and the bank gets robbed and the backup cd's in the safe deposit box are taken the flash has the most important <img src="/images/graemlins/smile.gif" alt="" /> I was surprised to learn that tif was the most standard dating back before jpg and gif (both of which were created to compress for slow modems not) you can view it on anything back to commodore and atari systems if you really wanted to <img src="/images/graemlins/smile.gif" alt="" />
I wouldn't want to look at XML either but its not much more complicated than HTML and its still possible at least to look at the XML of an open office spreadsheet where you can't look at the internals of a Microsoft Office Spreadsheet (though Microsoft is supposed to be going to XML with office but they are also incorporating a lot of DRM stuff which coule prevent you from looking inside as well). Again the XML is the last ditch effort just like the PSK the flash drive is kept in. HTML isn't easy for things liek spreadsheets such as my gear inventory; tables in HTML just are not complex enough.