>Hi, Dawa-
>
>>Thank you for the correction.
>>Could you explain why a fast process would affect the uniqueness of a file name, if you don't mind?
>
>Well, I double-checked the help file, and now I'm uncertain that the fast process will be a problem. Since it's based on the system date and time, I thought that more than once a milisecond would cause a problem, but the help explicitly says that the file names _will_ be unique if calls occur within a millisecond. So, I apologize for that misdirection.
>
>However, I'd still use the API for something like this since the function isn't officially a file-based function.

It's unique on the same machine - however, if two (or multiple) machines were using sys(2015) at the same time, there's a chance they'd eventually produce the same string... pretty much as in the proverbial Shakespearean monkeys paradigm. Of course, the solution to this is quite simple: prefix the filenames with a machine-specific string.