RE: [ciq] xLink implementation of xBRL

Ram,

I forgot to add – in the alias method – the XLink alias would be the XLink reference URL – so it should just work with existing XLink… again – we’d need to work up some examples from real XBRL / legalXML – so we know it works for sure…

From: David RR Webber [mailto:david@drrw.info]
Sent: Sunday, December 03, 2006 11:25 AM
To: 'Ram Kumar'
Cc: 'Max Voskob'; ciq@lists.oasis-open.org
Subject: RE: [ciq] xLink implementation of xBRL

Ram,

Yes – I apologize – I’ve just not had the cycles I would like to dedicate to this – and unfortunately the W3C really has not addressed this in any systematic way – since they view this as out-of-scope of their prime XML-for-documents-and-web mission. XLink has its advocates and detractors and was somewhat of a W3C in conflict during the development of XLink itself.

I’m mindful of the fact that in the original XML/edi work architecture – this referential information linkage was one role explicitly assigned to the registry – and all the attendant information security and access control that is therefore inherent in that technology. And its just a case of using simple ID-ref values then to link out.

Meanwhile people don’t want to use registries because as yet there is no quick and easy methods there (BTW – sidebar – we’re working on aggressively changing that in 2007 with registry-lite extensions via REST and AJAX support priorities).

So therefore we are back to using XML-based non-registry methods. I’m just REALLY nervous about doing this and then people wrongly thinking that “Oh you need XLink to do CIQ now”. However we structure this – it has to be optional.

Just brainstorming on this – maybe we can package this as a functionality – and then offer extensible methods?

If we have base ID-ref values – then those look-ups can be implemented in a variety of outward ways - XLink, REST / Registry, SQL/RPC, xinclude. We developed this in CAM templates – where you had an optional section to declare your external referencing address method – and assign an alias to it. Then you provide the alias/ID-ref – pairing – and then implementers can build their preferred methods themselves.

Applying this technique to CIQ – you would have an include for the <access-method> xsd – and that then would define the structure of that method you wanted to use. You have one already the <XLink>; people would then be free to define other alternate methods in that <access-method> include?

I’m just thinking aloud on this – I can find the sample <XML> from CAM as well – to compare and contrast what we did there…

Thanks, DW

From: Ram Kumar [mailto:kumar.sydney@gmail.com]
Sent: Sunday, December 03, 2006 12:33 AM
To: David RR Webber (XML)
Cc: Max Voskob; ciq@lists.oasis-open.org
Subject: Re: [ciq] xLink implementation of xBRL

Hi David,

Use of xLink in CIQ is not just for xBRL support only. It is also a requiement for

CIQ to handle external name and address referencing as we discussed before. Given that we have been talking about alternative approach to xLink for the past 6 months and that we do not have any solution that we have come up with, we have to move on.

I do not know what else to do. So, I decided that we stick to xLink and also include key

ref as an option (as was with V2.0).

Any advice?

Regards,

Ram

On 12/3/06, David RR Webber (XML) <david@drrw.info> wrote:

Ram,

I'm not convinced here that this is all good yet! They may have just "knee-jerked" this - and said - OK - then we need XLink - without even looking at use cases or such.

I do have some old contacts with the original XBRL crew - not spoke to them in five years however.

I'd just feel a whole lot better about this if we had concrete examples here - rather than a blanket executive level 50,000ft requirement.

I can see them aligning their own use of xlink methods - but I'm missing how the CIQ part plays - if we're the consistency layer - what else are they expecting to be able to do? As I'd previously noted - we can support external referencing into address lists by IDs - rather than the xlink inline itself - whereas the reverse use case - xlinks within the address - that I'm struggling to relate to how that would work?

DW

"The way to be is to do" - Confucius (551-472 B.C.)

-------- Original Message --------
Subject: Re: [ciq] xLink implementation of xBRL
From: "Ram Kumar" < kumar.sydney@gmail.com>
Date: Sat, December 02, 2006 4:18 pm
To: david@drrw.info

Cc: "Max Voskob" <max.voskob@paradise.net.nz >, ciq@lists.oasis-open.org

Hi David,

Tax XML requirement to xBRL was to provide ability to interoperability with CIQ.

I think this could be the reason why xBRL wanted interoperability with CIQ and

different implementations of xLink by both the groups made it a problem.

Regards,

Ram

On 12/3/06, David RR Webber <david@drrw.info> wrote:

Ram,

OK - I'll have to see what you've done.

I'm still baffled while XBRL think this absolutely has to be in CIQ itself!?

Seems like its entirely external from the use case I posted.

I'm guessing they have some use case where they are XLinking from the CIQ
itself out to something else? We would most certainly need to make that
completely isolated and optional - so people not wanting that do not have to
include it at all?

DW

-----Original Message-----
From: Ram Kumar [mailto: ram.kumar@oasis-open.org ]
Sent: Friday, December 01, 2006 5:06 PM
To: David RR Webber (XML)
Cc: Ram Kumar; Max Voskob; ciq@lists.oasis-open.org
Subject: Re: [ciq] xLink implementation of xBRL

Hi David,

I have successfully included the xlink specs. into CIQ yesterday. I have
also included the reference key approach that we used to have in V2.0. BTW,
it gives the options to either use xlink or standard ref. approach.

I will be circulating the revised specs. with all documents soon for review
and I am working hard on getting everything done.

Regards,

Ram

David RR Webber (XML) wrote:
> Ram,
>
> Ok - I just looked at that schema fragment - there's no magic there -
> its nothing more than just the W3C default definition of XLink statement.
>
> I think I'm beginning to fathom this out here (in amongst 1001 other
> distractions!).
>
> I understand why XBRL uses XLink - because its essentially a
> "spreadsheet" flattened into XML of an accounting statement and
> report.  Hence each column and each total in the spreadsheet needs to
> be dynamically labelled to indicate what the column or total
> represents in terms of the reporting vocabulary of the particular
> legal requirements for those reporters and reportees.
>
> This is not our world!  I don't see that we are going to dynamically
> label "PersonFirstName" or whatever - with some alternate meaning!!
>
> Therefore - we can expect that our adoptors are interested in
> continuing using our definitions "as is".
>
> However the use case appears to be that XBRL wants to dynamically
> reference to a reporter or customers, or reportees - a piece of
> explicit content as being external to that particular piece of XML
> content - statically defined by an XLink reference.
>
> This looks something like this:
>
> <XBRL>
>    <addresses>
>        <xlink - "points to external URL of content" ID="Jim"
> action="include"/>
>    </addresses>
>  </XBRL>
>
> External addresses XML contains
>                 <address-content>
>                    
>                     <an-address ID="fred"/>
>                     <an-address ID="jim"/>
>                 </address-content>
>
> I'm not sure at all why we have to start putting XLink into CIQ itself
> in order to support them using XLink addressing into content in this
> way?!?
>
> They should be able to use XPath references coupled with a named ID
> completely independently of us.
>
> And in their schema if they put #ANY as the structure returned from
> the XLink to the address - it will accept the CIQ formatted content.
> E.g. in the example above <addresses> is of type #ANY.
>
> Now if they want us to merely add an ID-type item into the CIQ
> structure at some strategic place - that should be fine - other people
> could use that just as is - as a regular ID-type element.
>
> We probably have one of these ID fields already though (I've not looked?).
>
> If this is all they want - then I suggest they clarify this and
> suggest what element they need the ID attribute to be associated with...?
>
> Thanks, DW
>
> "The way to be is to do" - Confucius (551-472 B.C.)
>
>
>     -------- Original Message --------
>     Subject: [ciq] xLink implementation of xBRL
>     From: "Ram Kumar" <kumar.sydney@gmail.com>
>     Date: Fri, December 01, 2006 2:51 am
>     To: ciq@lists.oasis-open.org
>     Cc: "Max Voskob" < max.voskob@paradise.net.nz>
>
>     CIQ TC,
>
>     Enclosed is the xlink specs. that xBRL is implementing soon.
>
>     Max,
>
>     Your assistance is sought here.
>
>     Is it possible to change the ciq v3.0 schemas to include this xBRL
>     specs?
>
>     Thanks
>
>     Regards,
>
>     Ram
>
>     ---------- Forwarded message ----------
>     From: *Hugh Wallis* <hughwallis@xbrl.org <mailto: hughwallis@xbrl.org>>
>     Date: Dec 1, 2006 8:07 AM
>     Subject: RE: Hello
>     To: Ram Kumar < kumar.sydney@gmail.com
> <mailto: kumar.sydney@gmail.com>>
>
>
>     Sure - this should be in a published erratum very soon - hopefully
>     December 7th or 13th.
>
>     This is really the only schema that needs to be shared between
>     XBRL and other specs. Although we have modified other of our
>     schemas they are all in xbrl.org <http://xbrl.org/ > owned
>     namespaces so you are probably not interested in them
>
>     Cheers
>
>     Hugh
>
>     Hugh Wallis - Standards Development
>     XBRL International
>     hughwallis@xbrl.org <mailto: hughwallis@xbrl.org>
>     Tel: +1 416 238 2553
>     Skype: hughwallis
>     MSN: hughwallis@hotmail.com <mailto: hughwallis@hotmail.com>(NOT an
>     e-mail address)
>
>
>
------------------------------------------------------------------------
>     *From:* Ram Kumar [mailto: kumar.sydney@gmail.com
>     <mailto:kumar.sydney@gmail.com>]
>     *Sent:* November 29, 2006 6:19 AM
>     *To:* Hugh Wallis
>     *Subject:* Hello
>
>
>     Hi Hugh,
>
>     Can you send me a copy of the revised xlink specs. (with doc. if
>     available)
>     that the xBRL vendors will implement? CIQ will implement the same
>     specs.to < http://specs.to/>
>     ensure interoperability between the two standards.
>
>     Thanks for all your assistance. Appreciated.
>
>     Regards,
>
>     Ram
>     OASIS CIQ TC
>

--
Ram Kumar
Manager - Technical Committee Development OASIS Post Office Box 455
Billerica,MA 0821 USA
+61 412 758 025 (Direct)
+ 1 978 667 5115 (OASIS HQ)
+ 1 978 667 5114 (Fax)
ram.kumar@oasis-open.org
http://www.oasis-open.org
"Advancing e-Business Standards Since 1993"

ciq message