Re: [ddlm-group] Draft EBNF for CIF2

Shall we wrap up this discussion then along the following lines:

(1) Nested save frames are removed from the CIF2 EBNF syntax specification

(2) Nested save frames will not be used in COMCIFS-approved DDLm dictionaries

(3) Documentation can point to the STAR2 specification and repeat the comments that John B has made regarding handling nested save frames.

Unless there are clear objections to (1) and (2) over the next few (working) days, I think we can consider the CIF2 EBNF with nested save frames removed as accepted by us and commence/continue work on supporting software and documentation.  At a later stage I plan to post the EBNF to our github site together with a FAQ document.

all the best,
James.

On Thu, Aug 28, 2014 at 7:08 AM, yayahjb <yayahjb@gmail.com> wrote:

We will have a lot less trouble with the new dictionaries if we don't use nested save frames in them.
We don't need nested save frames in a dictionary. We don't need any save frames, nest, or otherwise
in a data file. Let's save nesting of save frames until we have an actual use case.

That being said, I am a great fan of liberal parsers than can make sense of natural extensions of the
language they should formally read.

On 8/27/14 10:40 AM, Bollinger, John C wrote:

I'm having trouble seeing the down side of providing for save frame nesting in the CIF2 syntax specifications.  Doing so would enable *but not require* nested frames to be used in DDLm and DDLm dictionaries, but any way around they are irrelevant to DDL1 and DDL2 dictionaries (whether written in CIF1 or CIF2 syntax) and to all CIF data files currently envisioned.  To a parser that does not understand them, nested frames will look like a combination of a missing frame terminator between adjacent frames plus an extraneous frame terminator at some later point, and such a parser must be prepared to handle those errors in some way anyway (that is exactly the CIF1 situation).  A parser specialized for a domain to which nested save frames are not relevant can be such a parser, since nested frames would be erroneous in its target domain anyway.

On the other hand, allowing nested frames in the syntax would maximize our leverage from the Perth group's existing tools and recent work.


Regards,

John

   

_______________________________________________
ddlm-group mailing list
ddlm-group@iucr.org
http://mailman.iucr.org/mailman/listinfo/ddlm-group

--
T +61 (02) 9717 9907
F +61 (02) 9717 3145
M +61 (04) 0249 4148