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

DDL2/mm_CIF bugs/suggestions

• Subject: DDL2/mm_CIF bugs/suggestions
• From: James Hester <jrh@xxxxxxxxxxxx>
• Date: Tue, 19 Jul 2005 18:03:14 +0900

I'm thinking that the Volume G editorial process probably caught this
first one:

1. In the ddl2 spec (filename ddl2.c96) note the following:
---------------------------------------
save_ITEM_ALIASES

    _category.description
;
    This category holds a list of possible alias names or synonyms for
    each data item.  Each alias name is identified by the name and
    version of the dictionary to which it belongs.
;
    _category.id                         item_aliases
    _category.mandatory_code             no
     loop_
    _category_key.name                 '_item_aliases.alias_name'
                                       '_item_aliases.dictionary'
                                       '_item_aliases.dictionary_version'
     save_
-------------------------------------
However, _item_aliases.dictionary_version has been replaced by
_item_aliases.version (and all of the official DDL2 dictionaries use
this) and no longer exists in the DDL2 spec, so obviously it cannot be a
category key.

I found this by trying to validate cif_mm_2.0.03.dic against ddl2.c96 -
surely this is not the first time that this has been attempted??

2. Regular expressions again.  The regular expression for type "line"
near the end of the cif_mm dic (version 2.0.03) includes "\n" but not
"\r".  Can I assume this is an omission?

3. A number of the regular expressions contain character classes between
square brackets which include the sequence ".../\{}...".  Python (and
probably Perl, Tcl etc.) unfortunately interprets the "\{" sequence as
an escaped curly bracket, and the backslash is thus lost from the list
(this is in contrast to the POSIX regex standard which states that
backslashes lose their meaning inside square brackets).  When/if the
regular expression in (2) is being updated, perhaps the \{ could be
changed to a "\\{" for robustness as well?

4. If people are serious about using the cif_mm dictionary to validate
files supported by cif_core, I would suggest some sort of mechanism to
indicate when a dataname first appeared in the aliased dictionary (e.g.
_item_aliases.version_first) rather than simply giving the version in
which the dataname appears.  As currently set up, the cif_mm dictionary
will have to carry around a long list of aliases for each dataname, with
one entry for every version of cif_core containing that dataname.
_______________________________________________
cif-developers mailing list
cif-developers@iucr.org
http://scripts.iucr.org/mailman/listinfo/cif-developers

Reply to: [list | sender only]

• Follow-Ups:
  • Re: DDL2/mm_CIF bugs/suggestions (Brian McMahon)

• Prev by Date: DDL1-DDL2
• Next by Date: DDL2 parent-child question
• Prev by thread: DDL2 parent-child question
• Next by thread: Re: DDL2/mm_CIF bugs/suggestions
• Index(es):
  • Date
  • Thread