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

Re: [ddlm-group] Straw poll results

Dear Colleagues,

   I intend to follow a different approach as far as I can -- having one 
lexer and one parser that conditionally report issues relating to either
CIF 1 or CIF 2, similar to the approach in gcc for deprecated
features.  We went through the business of having separate parsers for
DDL1 CIF and DDL2 CIF when DDL2 CIF came in.  I have found life much
simpler since combining them, and I intend to try to do the same on
this transition.

   Regards,
     Herbert

=====================================================
  Herbert J. Bernstein, Professor of Computer Science
    Dowling College, Kramer Science Center, KSC 121
         Idle Hour Blvd, Oakdale, NY, 11769

                  +1-631-244-3035
                  yaya@dowling.edu
=====================================================

On Wed, 14 Oct 2009, James Hester wrote:

> Dear All,
>
> In this email I address the whole deprecation issue.  Nick: perhaps
> you missed it, but the straw poll went in favour of immediately
> setting a standard, with no deprecation, so the issue of deprecation
> is now somewhat academic.  That new standard looks almost exactly like
> you proposed, except for the addition of UTF8 and remaining doubts
> around whitespace.
>
> For the record, when proposing deprecation I envisioned a 1.2 standard
> which looked like 1.1 except for the addition of brackets and triple
> quoted strings -i.e. with only those changes that broke readers.  At
> the same time as 1.2 was fixed, we also fix a 1.3 standard which
> contains the additional changes that break writers, i.e. reduced
> character set for non-delimited strings and removing embedded
> delimiters. In the notes to the 1.2 standard and in the notes on the
> 1.1 standard, we insert statements to the effect that the 1.1
> character set for non-delimited strings is deprecated and that all new
> applications should use the productions contained in the 1.3 standard.
> That is what I meant by deprecation: formal statements attached to the
> standard stating that certain features are deprecated.  Of course the
> BNF for each standard cannot communicate this information.
>
> The purpose of this deprecation was to try to shepherd the CIF-writing
> software community with us as we change the standard, by giving them
> fair warning and, when the 1.3 standard comes into effect, hopefully
> having a large body of conformant CIF1.3-writing software available.
> However, the downside is that we have two new syntax standards instead
> of one and increased potential for confusion, and in any case there
> will be plenty of CIF1.1 files floating around from archives and the
> consequent need to deal with them.
>
> I believe Brian's proposal of making a single, clean, maximally
> disruptive change to the standard and mandating a compulsory CIF1.2
> header is therefore more reasonable.  It is entirely in the
> community's hands as to when and if they make a transition to CIF1.2,
> and rather than pushing and pulling them as I was envisioning, we rely
> solely on making CIF1.2 easy to transition to (documentation, test
> suites) and attractive as well (dREL, UTF8), without any hint of
> coercion.  We insert deprecation notes into the CIF1.1 standard, and
> explain how to write a CIF1.1 file that is also conformant to CIF1.2.
> That way, any new CIF writing software should at least meet the new
> standard.  We provide simple guides on the differences for parsers and
> actively engage major CIF writing software packages.
>
> The only suggestion I would make is that the new syntax is called
> CIF2.0, in recognition of its disruptive nature.
>
>
>
> -- 
> T +61 (02) 9717 9907
> F +61 (02) 9717 3145
> M +61 (04) 0249 4148
> _______________________________________________
> ddlm-group mailing list
> ddlm-group@iucr.org
> http://scripts.iucr.org/mailman/listinfo/ddlm-group
>
_______________________________________________
ddlm-group mailing list
ddlm-group@iucr.org
http://scripts.iucr.org/mailman/listinfo/ddlm-group

Reply to: [list | sender only]