Reverse engineering question

[briannacd]

I'm trying to build a disk packer program to fill up every last sector for UDF (ISO+Joliet finished). In my sample set, I have at least two problem directories. When I look at what ImgBurn has created for these directories, I see that a number of files have a data area of 42 bytes before the filename, but some only have 40, and I can't figure out how this is determined. In one directory, I get an average pad length of 41.95 over 114 entries, yet I have another directory that has an average pad length of 41.3 over 202 files. Any help with this would be appreciated. I've looked at the ecma-168 and udf260 specs but they are as helpful as a screen door on a submarine.