Re: ISSUE-257: Interpretation of percentage tts:fontSize on root (tt) element

Am 22.07.2013 17:20, schrieb Michael Dolan:
>
> Yes, we would need XSD 1.1 to do this – there was a new attribute 
> added to <any> to exclude namespaces.
>
  Yes, it would be good to check, if this would work (but looks good).

> But the real problem is not the use of ##other exactly, but the use of 
> “lax” in combination with arguably non-conformant commercial 
> validation tools that treat “lax” as a top level, “must ignore the 
> schema  even though it is provided”.  That’s not the XSD definition of 
> “lax” which is supposed to be a best effort.
>
> My personal view is that we should go to strict and thus require the 
> foreign namespace schemas during validation.
>
Strict requires to obtain the Schema for the foreign namespace attribute 
and use it for validation. TTML defined behaviour is to prune the 
foreign namespace elements before validation of TTML conformance. I 
think this make sense. It is out of scope for the TTML spec to require 
validation of elements that are not defined in the spec.

With "strict" a  processor must have access to all XML schemas to any 
foreign namespace element or attribute that is defined in a TTML 
document. I assume that this will not always be the case.

Best regards,

Andreas


Am 22.07.2013 17:20, schrieb Michael Dolan:
>
> Yes, we would need XSD 1.1 to do this – there was a new attribute 
> added to <any> to exclude namespaces.
>
> But the real problem is not the use of ##other exactly, but the use of 
> “lax” in combination with arguably non-conformant commercial 
> validation tools that treat “lax” as a top level, “must ignore the 
> schema  even though it is provided”. That’s not the XSD definition of 
> “lax” which is supposed to be a best effort.
>
> My personal view is that we should go to strict and thus require the 
> foreign namespace schemas during validation.
>
> Regards,
>
> Mike
>
> *From:*Andreas Tai [mailto:tai@irt.de]
> *Sent:* Monday, July 22, 2013 7:12 AM
> *To:* Glenn Adams
> *Cc:* Michael Dolan; public-tt
> *Subject:* Re: ISSUE-257: Interpretation of percentage tts:fontSize on 
> root (tt) element
>
> >  I'm not sure it is possible to say ##other minus {ttm,ttp,tts} 
> namespaces, or at least I haven't tried doing this with XSD.
>
> This is not possible with XML Schema 1.0. If the extensibility of TTML 
> elements with "foreign namespace" attributes is implemented with the 
> XML Schema anyAttribute element, this will automatically open the door 
> for potential false positive messages. Therefore documents that use 
> undefined attributes  in (some) TTML namespaces will validate although 
> this is not allowed.
>
> Best regards,
>
> Andreas
>
> Am 17.07.2013 04:04, schrieb Glenn Adams:
>
>     On Tue, Jul 16, 2013 at 6:42 PM, Michael Dolan <mdolan@newtbt.com
>     <mailto:mdolan@newtbt.com>> wrote:
>
>     Agreed. The new XML schema currently permits style attributes, and
>     should be corrected.
>
>     This is a side effect of having added ##other to core attributes
>     in resolving ISSUE-150 [1][2].
>
>     [1] http://www.w3.org/AudioVideo/TT/tracker/issues/150
>
>     [2]
>     http://dev.w3.org/cvsweb/2008/tt/spec/xsd/Attic/ttaf1-dfxp-core-attribs.xsd.diff?r1=1.3;r2=1.4;hideattic=0;f=h
>
>     I'm not sure it is possible to say ##other minus {ttm,ttp,tts}
>     namespaces, or at least I haven't tried doing this with XSD.
>
>         Unrelated, but also about the schema, I thought we had found
>         the authoritative xml.xsd on ww3.org <http://ww3.org> a while
>         back and we were going to defer to that and not include our
>         own copy?
>
>     There is a problem with the xml.xsd you refer to [3]: namely, it
>     types xml:lang as xs:language, but xs:language [4] doesn't admit
>     the empty string, which is a legal value according to XML 1.0 [5].
>
>     [3] http://www.w3.org/2001/03/xml.xsd
>
>     [4] http://www.w3.org/TR/xmlschema-2/#language
>
>     [5] http://www.w3.org/TR/REC-xml/#sec-lang-tag
>
>         Mike
>
>         *From:*Glenn Adams [mailto:glenn@skynav.com
>         <mailto:glenn@skynav.com>]
>         *Sent:* Tuesday, July 16, 2013 4:49 PM
>         *To:* public-tt
>         *Subject:* Re: ISSUE-257: Interpretation of percentage
>         tts:fontSize on root (tt) element
>
>         I was reviewing the recently closed ISSUE-257 today, and after
>         further consideration, have determined it was incorrect, and
>         needed to be reverted which I have done in [1].
>
>         The basic problem with the previous resolution is that TT
>         Style Attributes are not defined for use on the *tt* element
>         (with the exception of the special case of tts:extent).
>         Furthermore, the outermost Content element [2] used during
>         content style inheritance is the *body* element.
>
>         Ultimately, there is no element on which tts:fontSize can be
>         specified that doesn't have a parent element. In the case of
>         *body*, if no tts:fontSize is specified, then region style
>         inheritance comes into effect. So that leaves us with a
>         tts:fontSize potentially specified on *region* as the
>         outermost inheritable font size. Accordingly, if the region's
>         font size is specified in percentage, then we do need to
>         resolve it against something other than its parent, since no
>         ancestor of region participates in style inheritance.
>
>         I will now re-implement a fix for this issue based on the need
>         to resolve percentage on region.
>
>         [1] https://dvcs.w3.org/hg/ttml/rev/0355e9473d7e
>
>         [2]
>         https://dvcs.w3.org/hg/ttml/raw-file/default/ttml10/spec/ttaf1-dfxp.html#element-vocab-type-content
>
>
>
>
> -- 
> ------------------------------------------------
> Andreas Tai
> Production Systems Television IRT - Institut fuer Rundfunktechnik GmbH
> R&D Institute of ARD, ZDF, DRadio, ORF and SRG/SSR
> Floriansmuehlstrasse 60, D-80939 Munich, Germany
>   
> Phone: +49 89 32399-389 | Fax: +49 89 32399-200
> http:www.irt.de  <http://www.irt.de>  | Email:tai@irt.de  <mailto:tai@irt.de>
> ------------------------------------------------
>   
> registration court&  managing director:
> Munich Commercial, RegNo. B 5191
> Dr. Klaus Illgner-Fehns
> ------------------------------------------------


-- 
------------------------------------------------
Andreas Tai
Production Systems Television IRT - Institut fuer Rundfunktechnik GmbH
R&D Institute of ARD, ZDF, DRadio, ORF and SRG/SSR
Floriansmuehlstrasse 60, D-80939 Munich, Germany

Phone: +49 89 32399-389 | Fax: +49 89 32399-200
http: www.irt.de | Email: tai@irt.de
------------------------------------------------

registration court&  managing director:
Munich Commercial, RegNo. B 5191
Dr. Klaus Illgner-Fehns
------------------------------------------------

Received on Monday, 22 July 2013 16:09:17 UTC