Next in thread → Next in month →

RE: [dita] Attribute change in 1.1

From
Grosso, Paul <>
Date
2006-12-11T14:31:49+00:00
ID
Thread
RE: [dita] Attribute change in 1.1
I agree with continuing to omit the attribute and putting
a note in the spec.

paul

> -----Original Message-----
> From: Robert D Anderson [mailto:] 
> Sent: Friday, 2006 December 08 17:36
> To: 
> Subject: [dita] Attribute change in 1.1
> 
> 
> As some of you may remember, back in June we voted to remove 
> the confusing
> xmlns attribute from DITA elements in 1.1. The attribute was 
> considered by
> all to be a bug - it had a fixed, default value of the empty 
> string. It was
> only in the DTDs, and did not appear in the schema. There was 
> no reason to
> have it in DITA documents or in the DTD. Here is a bit of history:
> 
> Paul Grosso's note questioning why it is present:
> http://lists.oasis-open.org/archives/dita/200605/msg00062.html
> TC minutes with approval to remove it:
> http://www.oasis-open.org/archives/dita/200606/msg00004.html
> 
> We recently noticed that, even though this is a bug, it is technically
> backwards-incompatible. The only ways that this would cause 
> problems is:
> 1. If a user normalizes files, but retains the DTD statement. 
> The default
> value of xmlns="" is added during normalization, but the 
> attribute is not
> valid with 1.1.
> 2. If a tool happens to place default attributes into the 
> file. This could
> easily cause other problems for DITA documents, because all class
> attributes would also be in the file, and thus easy for 
> authors to break.
> 
> Given that the xmlns attribute was considered a bug, it would 
> seem to be a
> mistake to put it back in. So, if there are no objections, 
> I'll ask Michael
> to put a notification of some sort about this in the 1.1 
> specifications.
> 
> Thanks-
> 
> Robert D Anderson
> IBM Authoring Tools Development
> Chief Architect, DITA Open Toolkit
> (507) 253-8787, T/L 553-8787
>
Next in thread → Next in month →