Re: [xml-dev] Venetian Blinds vs Garden of Eden patterns for industry st

["Pete Cordell" <petexmldev@codalogic.com>]

To reset the debate, how would you tackle a schema design based around the 
following example? (Apologies for the naffness of it.  It's getting late and 
I've never been good at examples anyway.)

Let's say that XSD part 2 didn't define xs:date, and instead you needed to 
use a schema library to define date something like:

<xs:schema targetNamespace="http://time.org";>
<xs:complexType name="date"><xs:sequence>
    <xs:element name="day" type="xs:int"/>
    <xs:element name="month" type="xs:int"/>
    <xs:element name="year" type="xs:int"/>
</xs:sequence></xs:complexType></xs:schema>

You could also have a date that was of the romantic type.  This also had its 
own standard schema that might have the form:

<xs:schema targetNamespace="http://relationships.org";>
<xs:complexType name="date"><xs:sequence>
    <xs:element name="name" type="xs:string"/>
    <xs:element name="height" type="xs:int"/>
</xs:sequence></xs:complexType></xs:schema>

I think you have to use namespaces to distinguish these two types of date. 
(You could make the above follow the Garden of Eden pattern if you like.)

Then say you wanted to write your own schema that allowed you to keep track 
of the dates on which various dates were free.  (Told you it was a naff 
example!)  This schema might look like:


<xs:schema targetNamespace="http://opportunities.org";
        xmlns:time="http://time.org"; xmlns:rel="http://relationships.org";>
    <xs:element name="availability">
        <xs:complexType><xs:sequence>
            <xs:element name="from" type="time:date"/>
            <xs:element name="to" type="time:date"/>
            <xs:element name="person" type="rel:date"/>
        </xs:sequence></xs:complexType>
    </xs:element>
</xs:schema>

Note that you can't use <xs:element ref="time:date"/> here because there are 
two time based dates.

As spec'd above, an instance of this with elementFormDefault="qualified" 
would look like:

<availability xmlns="http://opportunities.org";
        xmlns:time="http://time.org"; xmlns:rel="http://relationships.org";>
    <from>
        <time:day>21</time:day>
        <time:month>12</time:month>
        <time:year>2012</time:month>
    </from>
    <to>
        <time:day>24</time:day>
        <time:month>12</time:month>
        <time:year>2012</time:month>
    </to>
    <person>
        <rel:name>Janice</rel:name>
        <rel:height>185</rel:height>
    </person>
</availability>

To me, those namespace prefixes in time:day and rel:name etc. look really 
ugly, and for someone who's not experienced in XML they might be a difficult 
to get right.  It would be cleaner if the XML instance used 
elementFormDefault="unqualified" and looked like:

<availability xmlns="http://opportunities.org"; >
    <from>
        <day>21</day>
        <month>12</month>
        <year>2012</month>
    </from>
    <to>
        <day>24</day>
        <month>12</month>
        <year>2012</month>
    </to>
    <person>
        <name>Janice</name>
        <height>185</height>
    </person>
</availability>

So how would you attack the problem of the above?

Pete Cordell
Codalogic Ltd
Interface XML to C++ the easy way using C++ XML
data binding to convert XSD schemas to C++ classes.
Visit http://codalogic.com/lmx/ or http://www.xml2cpp.com
for more info
----- Original Message ----- 
From: "Philip Fennell" <Philip.Fennell@marklogic.com>
To: "Pete Cordell" <petexmldev@codalogic.com>; <xml-dev@lists.xml.org>
Sent: Thursday, October 28, 2010 12:59 PM
Subject: RE: [xml-dev] Venetian Blinds vs Garden of Eden patterns for 
industry standards XML Schemas



Pete wrote:

> That way you only ever have one namespace declaration (ignoring QName
> issues) irrespective of the types you use from other namespaces, and you
> only have the root element as qualified.  So your XML instance looks like:
> <myNS:root xmlns:myNS="abc">
>     <element1>
>         <elementWithTypeDefinedInAnotherNamespace>
>         </elementWithTypeDefinedInAnotherNamespace>
>     </element1>
> </myNS>
> The instance is appropriately tagged with it's namespace, but it's all
> clearly done in one place and there's no other namespace gotchas to worry
> about.

I think I see what your trying to do but surely with the above approach you 
are going to have problems when processing fragments of the document because 
you'll lose the context if the fragments are taken as-is. I presume, to 
counter that you're talking about having the 'context' or 'root' wrapper 
being used to contain any fragment of the document which means that the 
container element's content model must allow all possible elements, from the 
schema, as children (or may be a defined sub-set of all). In more complex 
compound documents (XHTML + XForms or XSLT + XHTML + XForms) that'd 
introduce a lot of noise into the document and add a different kind of 
complexity where you're switching namespaces more frequently as you traverse 
the tree in and out of context wrappers. From a readability point of view 
you'd be constantly having to scan back up the tree to get you namespace 
context. Within an XSLT transform that'd be a nightmare.

Regards

Philip Fennell
Consultant
MarkLogic Corporation
Mobile +44 (0) 7824 830 866
email  Philip.Fennell@marklogic.com
web    www.marklogic.com

This e-mail and any accompanying attachments are confidential. The 
information is intended solely for the use of the individual to whom it is 
addressed. Any review, disclosure, copying, distribution, or use of this 
e-mail communication by others is strictly prohibited. If you are not the 
intended recipient, please notify us immediately by returning this message 
to the sender and delete all copies. Thank you for your cooperation.





________________________________________
From: Pete Cordell [petexmldev@codalogic.com]
Sent: 28 October 2010 11:16
To: xml-dev@lists.xml.org
Subject: Re: [xml-dev] Venetian Blinds vs Garden of Eden patterns for 
industry standards XML Schemas

One thing I'm surprised is not more popular is to always use Venetian Blind,
even across namespaces, BUT make local elements unqualified, i.e. use
elementFormDefault="unqualified".

That way you only ever have one namespace declaration (ignoring QName
issues) irrespective of the types you use from other namespaces, and you
only have the root element as qualified.  So your XML instance looks like:

<myNS:root xmlns:myNS="abc">
    <element1>
        <elementWithTypeDefinedInAnotherNamespace>
        </elementWithTypeDefinedInAnotherNamespace>
    </element1>
</myNS>

The instance is appropriately tagged with it's namespace, but it's all
clearly done in one place and there's no other namespace gotchas to worry
about.

The approach mirrors JSON's response to "there's nothing like XML namespaces
in JSON."

I think the issue is that the interpretation of a element depends on the
context provided by the parent and XSLT and other XPath technologies have
difficulty with that.

Spot the outsider!

Pete Cordell
Codalogic Ltd
Interface XML to C++ the easy way using C++ XML
data binding to convert XSD schemas to C++ classes.
Visit http://codalogic.com/lmx/ or http://www.xml2cpp.com
for more info
----- Original Message -----
From: "Pete Cordell" <petexmldev@codalogic.com>
To: <xml-dev@lists.xml.org>
Sent: Thursday, October 28, 2010 10:58 AM
Subject: Re: [xml-dev] Venetian Blinds vs Garden of Eden patterns for
industry standards XML Schemas


> FWIW I would use:
>
> <xs:element name="" type=""/>
>
> (i.e. Venetian Blind) where type is within the same namespace/schema as
> the element, and:
>
> <xs:element ref=""/>
>
> (i.e. Salami Slice) when the element(/type) being accessed is in a
> different namespace.
>
> To give it a name I'll call it Venetian Slice!
>
> By analogy to a class interface, I see the global elements as the public
> members of an interface, and the global types as the private members.  The
> interface should only publically expose what it has to, and Venetian Blind
> provides a non-narrative way to express that.  Accessing a new namespace
> is analogous to accessing a new class, and so the public interface (i.e.
> global elements) of that namespace have to be used.
>
> That should open up the discussion!
>
> Pete Cordell
> Codalogic Ltd
> Interface XML to C++ the easy way using C++ XML
> data binding to convert XSD schemas to C++ classes.
> Visit http://codalogic.com/lmx/ or http://www.xml2cpp.com
> for more info
> ----- Original Message -----
> From: "Dennis Sosnoski" <dms@sosnoski.com>
> To: <xml-dev@lists.xml.org>
> Sent: Thursday, October 28, 2010 6:41 AM
> Subject: Re: [xml-dev] Venetian Blinds vs Garden of Eden patterns for
> industry standards XML Schemas
>
>
>>
>> If you're thinking of using data binding for working with the data in
>> applications I'd strongly recommend a Venetian Blinds approach. Most
>> data binding approaches turn Garden of Eden into an amorphous foam of
>> tiny classes - which is, after all, exactly what Garden of Eden
>> represents in XML terms.
>>
>> Associating the name "Garden of Eden" with this concept of pureed
>> structure was excellent marketing, but doesn't really give the right
>> impression. I'd suggest "Element Soup" as an alternative. :-)
>>
>>  - Dennis
>>
>>
>> On 10/28/2010 05:31 AM, Lech Rzedzicki wrote:
>>> Hi all.
>>>
>>> I am now involved in developing the next iteration of schemas to
>>> standardise information storage and exchange for trademarks and design
>>> (there is also some cooperation with patent people).
>>> The current standards (TM-XML and DS-XML) use a venetian blind pattern
>>> because it nicely mimics OO model, so it's easy to generate classes
>>> and objects, also it hides element definitions so there's no conflicts
>>> when processing them. Finally the naming convention (not my idea)
>>> means that the element names are not reusable anyway.
>>> Some people are now proposing that we move to "Garden of Eden" design
>>> patter, but I don't see that as a particularly smart move in the
>>> context of our domain.
>>> I have noticed that many of the schemas in the industry, for instance
>>> UBL have gone for Garden of Eden first and and later reverted to
>>> Venetian Blinds.
>>> I am very interested in some thoughts from the trenches as to why one
>>> or the other approach might be more useful in a committee driven
>>> standards with contributors from different domains.
>>>
>>> No need to point me to google either - I have read all the articles
>>> about it [1,2,3] and authored schemas that are in production now both
>>> ways, but still unconvinced either way...
>>>
>>>
>>> References:
>>>
>>> [1] A slideshows on various approaches:
>>> http://dret.net/lectures/xml-fall08/xsd-2#(11)
>>> [2] http://www.xfront.com/GlobalVersusLocal.html - schema best practices
>>> [3] http://www.ibm.com/developerworks/xml/library/x-schemascope/
>>> etc...
>>>
>>> _______________________________________________________________________
>>>
>>> XML-DEV is a publicly archived, unmoderated list hosted by OASIS
>>> to support XML implementation and development. To minimize
>>> spam in the archives, you must subscribe before posting.
>>>
>>> [Un]Subscribe/change address: http://www.oasis-open.org/mlmanage/
>>> Or unsubscribe: xml-dev-unsubscribe@lists.xml.org
>>> subscribe: xml-dev-subscribe@lists.xml.org
>>> List archive: http://lists.xml.org/archives/xml-dev/
>>> List Guidelines: http://www.oasis-open.org/maillists/guidelines.php
>>>
>>>
>>
>> _______________________________________________________________________
>>
>> XML-DEV is a publicly archived, unmoderated list hosted by OASIS
>> to support XML implementation and development. To minimize
>> spam in the archives, you must subscribe before posting.
>>
>> [Un]Subscribe/change address: http://www.oasis-open.org/mlmanage/
>> Or unsubscribe: xml-dev-unsubscribe@lists.xml.org
>> subscribe: xml-dev-subscribe@lists.xml.org
>> List archive: http://lists.xml.org/archives/xml-dev/
>> List Guidelines: http://www.oasis-open.org/maillists/guidelines.php
>>
>>
>




_______________________________________________________________________

XML-DEV is a publicly archived, unmoderated list hosted by OASIS
to support XML implementation and development. To minimize
spam in the archives, you must subscribe before posting.

[Un]Subscribe/change address: http://www.oasis-open.org/mlmanage/
Or unsubscribe: xml-dev-unsubscribe@lists.xml.org
subscribe: xml-dev-subscribe@lists.xml.org
List archive: http://lists.xml.org/archives/xml-dev/
List Guidelines: http://www.oasis-open.org/maillists/guidelines.php