OK - to try and narrow down into specific cases:
In DITA 1.1, if I specialize map and add no domains, what should @domains be? Is the empty string correct?
In DITA 1.3, for a map with no specializations or constraints, I now see that you've got the "(map map)" token in the RNG (not in the DTD, which is what I originally checked). I assume those were added with a 1.3 proposal, but I do not remember seeing them previously - I thought tokens were only added for specializations. To me, a value of (map map) means "map specialized from map", so it seems ... odd. If anything, I'd have expected just (map) as the token.
Robert D Anderson
IBM Authoring Tools Development
Chief Architect, DITA Open Toolkit (http://dita-ot.sourceforge.net/)
Eliot Kimber ---10/03/2014 15:10:10---Yes, @domains without structural types is valid but incomplete. The map and topic modules have domai
From: Eliot Kimber <>
To: Robert D Anderson/Rochester/IBM@IBMUS
Cc: DITA TC <>
Date: 10/03/2014 15:10
Subject: Re: [dita] Question about domains attribute
Sent by: <>
Yes, @domains without structural types is valid but incomplete.
The map and topic modules have domains contributions defined in the RNG
modules but they're not being put into the generated DTDs: that's
definitely a bug that I will fix--they should be there for completeness.
For the statement "DITA document types are defined via the @domains
attribute" to be true, the @domains value must list structural types in
addition to @domains.
Cheers,
E.
—————
Eliot Kimber, Owner
Contrext, LLC
http://contrext.com
On 10/3/14, 2:52 PM, "Robert D Anderson" <> wrote:
>Ah, of course. That said - I didn't think that the base topic or map had
>domain tokens? They do not appear in the DITA 1.3 document types (I'm
>looking at base topic and base map, which use the fewest domains). So if
>I want a new shell for base map -- what value is legal or required for
>@domains in DITA 1.* on the <map> element, when using no specializaations
>of any kind, and using no constraints?
>
>For the case that was causing me trouble - the structural specialization
>and shell DTD were created back in the days of DITA 1.1, so we did not
>have a structural domain token. I can create that token now and bypass
>the error. Still, it's not an absolute requirement - as you say, it's
>strongly urged, but not a MUST for exactly this reason - I believe my
>shell DTD was valid in DITA 1.1, and should remain valid in 1.2 and 1.3.
>
>Robert D Anderson
>IBM Authoring Tools Development
>Chief Architect, DITA Open Toolkit (http://dita-ot.sourceforge.net/)
>
>Eliot Kimber <> wrote on 10/03/2014 14:33:02:
>
>> From: Eliot Kimber <>
>> To: Robert D Anderson/Rochester/IBM@IBMUS, DITA TC
>><>
>> Date: 10/03/2014 14:33
>> Subject: Re: [dita] Question about domains attribute
>>
>> Starting with DITA 1.2 we strongly urged @domains to list all structural
>> types as well as domains, so the minimum @domains value should be
>>"(topic
>> topic)" or "(map map)" in the case where you have no domains integrated.
>>
>> Cheers,
>>
>> E.
>> —————
>> Eliot Kimber, Owner
>> Contrext, LLC
>> http://contrext.com
>>
>>
>>
>>
>> On 10/3/14, 2:21 PM, "Robert D Anderson" <> wrote:
>>
>> >The DITA spec is pretty clear that structural specializations need to
>> >declare @domains on the root element (specialization of map or topic),
>> >and that this attribute must describe the domains used in a given
>> >document type shell. The spec says of this attribute: "...the @domains
>> >attribute, whose value is a sequence of parenthesized module ancestry
>> >specifications. "
>> >http://docs.oasis-open.org/dita/v1.2/os/spec/archSpec/domainsatt.html
>> >
>> >The question is - what about if I have a document type shell with no
>> >domains? I've got a very simple map specialization - mostly a title,
>> >metadata, and very limited references to other maps. There is no need
>>for
>> >any domains. I've declared the attribute for my root element, but in my
>> >DTD file the attribute is set to the empty string (""). This would not
>> >appear to comply with the language above, because the empty string is
>>not
>> >a sequence of parenthesized modules. Is it correct to set this to the
>> >empty string? I've got a tool that reports an error here because
>>@domains
>> >does not match the current definition. I can get around the error by
>> >setting a value like "none" or "(none)" but those clearly do not comply
>> >either.
>> >
>> >So - what is the correct value for @domains in a document type shell
>>that
>> >does not use any domains?
>> >
>> >Robert D Anderson
>> >IBM Authoring Tools Development
>> >Chief Architect, DITA Open Toolkit (http://dita-ot.sourceforge.net/)
>>
>>
---------------------------------------------------------------------
To unsubscribe from this mail list, you must leave the OASIS TC that
generates this mail. Follow this link to all your TCs in OASIS at:
https://www.oasis-open.org/apps/org/workgroup/portal/my_workgroups.php
