This is good. Wikipedia says the delimiter is sequence of "characters". A minor nit is that DFDL allows delimiters to contain byte values that might not have any charset mapping, so generalizes this to delimiter is a sequence of "characters and bytes".
But honestly, I think the byte value stuff is pretty obscure and not worth confusing people right up front. So I'd go with the wikipedia definition and then only later introduce the raw-byte stuff. (Which is not even implemented in Daffodil btw.) ________________________________ From: Costello, Roger L. <[email protected]> Sent: Wednesday, September 18, 2019 8:04 AM To: [email protected] <[email protected]> Subject: DFDL supports 3 types of delimiters, correct? Hello DFDL community, First, what is a delimiter? Wikipedia says it is this: A delimiter is a sequence of one or more characters for specifying the boundary between separate, independent regions in plain text or other data streams. Is that what the DFDL specification means when it uses the term "delimiter"? Next, the universe of data formats specify "the boundary between separate, independent regions" in various ways. I believe that the DFDL Working Group has categorized them into 2 categories: 1. Delimiters placed at the beginning of a region. 2. Delimiters placed at the end of a region. DFDL provides 3 properties for expressing those categories of delimiters: - dfdl:initiator - dfdl:terminator - dfdl:separator (coupled with dfdl:separatorPosition) Do you agree? Is the above writeup expressed precisely? Any recommended changes? If it is precise (and correct) I will add it to my tutorial. /Roger
