[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [ddlm-group] Use of elides in strings

James Hester wrote:
> Thanks for the quick reply over Thanksgiving, John.  I take from your
> message that the PDB does not need any elide mechanism to be defined
> in the CIF2 syntax.  Would you therefore be prepared to vote in favour
> of not defining any elides, or would you prefer to abstain?
> 
> Votes so far:
> 
> No elides: James, Nick, Herbert if the IUCr + PDB say it is OK
> Elides:?
> 
> Unknown: John, Joe, David B., Brian, Simon
> 
I am not part of PDB or IUCr, so I don't know the software and content 
issues, so my vote should not hold much weight.

However, my opinion is that an elide mechanism is an essential feature 
of any text data format. It also does not have to be as complicated as 
it has become in defining CIF2. CSV files are commonly used for database 
dumps, and have no string restrictions, including multiple lines. CSV 
uses the repeated-quote elide mechanism as in Fortran, which avoids the 
complication of also requiring rules for the reverse-solidus escape 
character. This simple single-character elide allows CSV strings to 
contain any string with only one quoting mechanism.

If it were up to me, I would switch CIF2 to CSV/Fortran style elides, 
allow milti-line strings, and drop the triple quotes. The current 
close-quote rule already breaks CIF1 strings with embedded quotes, so 
why not just make regular quoted strings more flexible as in CSV? It is 
simple, and well-established.

If the decision to exclude elides, I don't see any conflicts, because 
any string can use semicolon delimiters with an indented text block. In 
that case, I would prefer that semicolon strings be redefined to always 
include a single leading space, to minimize the need for 
application-dependent dictionaries just to store/retrieve string data.

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

Reply to: [list | sender only]