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

Re: [ddlm-group] Characterset of non-delimited stringsinsidecompou nd data items. .

I would be fine with disallowing all of [, ], {, and } from appearing anywhere in a whitespace-delimited data value.  The [ and { do not at present cause lexical ambiguity, but there’s something to be said for consistent treatment of opening and closing delimiters.  Do note, however, that the opening [ of a list or { of a table is still required to be separated from any preceding value by whitespace.  Thus

 

_t            [ depth_1[depth_1_still ] ]

 

would contain an error at the second [.  (There is still an error under the current spec or if only ] and } are forbidden in whitespace-delimited data values, but it is at the final ].)  This is still OK, though:

 

_t            [[depth_2] depth_1]

 

 

John

 

 

From: ddlm-group-bounces@iucr.org [mailto:ddlm-group-bounces@iucr.org] On Behalf Of SIMON WESTRIP
Sent: Tuesday, November 09, 2010 9:03 AM
To: Group finalising DDLm and associated dictionaries
Subject: Re: [ddlm-group] Characterset of non-delimited strings insidecompou nd data items. .

 


I would suggest adding [ & { also:

_t            [ depth_1[depth_1_still [ depth_2 x=a1[a2[a3[a4[i]]]];]]


?

Simon


 


From: James Hester <jamesrhester@gmail.com>
To: Group finalising DDLm and associated dictionaries <ddlm-group@iucr.org>
Sent: Tuesday, 9 November, 2010 3:14:55
Subject: Re: [ddlm-group] Characterset of non-delimited strings inside compound data items

Indeed.  I agree that "ease of use of flex" is not a good criterion. A
better way of putting it would be "simplicity of implementation".

Glad you don't object to excluding close brackets.

James.

On Tue, Nov 9, 2010 at 1:23 PM, Herbert J. Bernstein
<yaya@bernstein-plus-sons.com> wrote:
> While I have no particular objection to excluding the close brackets from
> non-delimited strings, personally I think making easy use of flex a
> criterion for the design of CIF2 is not a good idea. -- 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 Tue, 9 Nov 2010, James Hester wrote:
>
>> Dear DDLm group,
>>
>> John Bollinger has alerted me to a glitch in the current DDLm
>> specification, to wit: (i) close bracket characters are allowed as
>> non-final characters in a non-delimited string, and (ii) there is no
>> requirement for whitespace between a datavalue and the close bracket
>> symbol that denotes the end of a table or list value.
>>
>> This means that, in order to decide whether a close bracket character
>> terminates a list or is just another character in the non-delimited
>> string, the parser must look ahead, potentially many levels of
>> nesting.  For example:
>>
>> _t            [outer [inner1 inner2]]
>>
>>
>> The parser does not know that the first close bracket closes the inner
>> list until it has read past the second close bracket.   Or even more
>> confusingly:
>>
>> _t            [ depth_1 [ depth_2 [ depth_3 x=a1[a2[a3[a4[i]]]];]]]
>>
>>
>> While this behaviour is not intractable, it is also not possible to
>> use simple lexing tools (e.g. flex) to handle such syntax.  I would
>> therefore like to propose the following change to the current draft
>> specification:
>>
>> "The characters ']' and '}' may not appear anywhere in a non-delimited
>> datavalue"
>>
>> James.
>> --
>> 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
>



--
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


Email Disclaimer: www.stjude.org/emaildisclaimer
_______________________________________________
ddlm-group mailing list
ddlm-group@iucr.org
http://scripts.iucr.org/mailman/listinfo/ddlm-group

Reply to: [list | sender only]