Re: [xml-dev] Help with XML anti pattern

["G. Ken Holman" <gkholman@CraneSoftwrights.com>]

Won't attributes for natural language and names 
restrict the field values somewhat? Of course not 
in raw Unicode characters, but if an exotic name 
requires element structure to specify characters 
of the name, that isn't available in attributes. 
Very (very!) much an edge case, but I've just 
avoided putting natural language into attributes.

I like what William is suggesting with the 
element names as they uniquely label the semantic 
associated with the element values:

<FirstName>Erick</FirstName>
<LastName>Rich</LastName>
<SWName>FancySoft v.1.0</SWName>

If the contractor wants to preserve the 
"name/value" concept, I think attributes would be 
appropriate for labeling the semantic along the lines of:

<value name="FirstName">Erick</value>
<value name="LastName">Rich</value>
<value name="SWName">FancySoft v.1.0</Value>

If you are looking for arguments why what you 
have been given is awkward in XML processing, 
addressing siblings with the same name but 
different semantics can be awkward because of the 
overlap of siblings. Of course one could address 
the immediately following sibling, there are more 
opportunities, I think, for ambiguity.

Consider in the original data that there are 
three <Value> elements that follow as siblings to 
the first <Name> element. Yes, the 
immediately-following <Value> element is the 
right one, but if this approach is meant, also, 
to address absent values, then the 
immediately-following <Value> element might be 
following a different <Name> element.

I consider sibling names with differing semantics 
to be poor structure design and I try to avoid it 
at all costs. I think it increases the cost of 
maintenance and it increases the opportunity for ambiguity.

I hope this helps.

. . . . . Ken

At 2021-11-23 13:49 -0500, Ihe Onwuka wrote:
> If you made FirstName, LastName and SWName 
> attributes of SWMakerInfo wouldn't that satisfy both parties.
>
> On Tue, Nov 23, 2021 at 1:21 PM William David 
> Velasquez <<mailto:wd@creativosdigitales.co>wd@creativosdigitales.co> wrote:
>
> Hi XML folks,
>
> I want to share here a real case to ask for your wisdom and advice:
>
> This week my company started to work in a 
> project involving a customization of UBL 
> Invoices. A team before us, defined a structure 
> similar to this for receiving the name of a 
> person and the name of a software product:
> > <ext:UBLExtension>
> > Â  <ext:ExtensionContent>
> > Â Â  <SWMaker>
> > Â Â Â  <SWMakerInfo>
> > Â Â Â Â Â  <Name>FirstName</Name>
> > Â Â Â Â Â  <Value>Erick</Value>
> > Â Â Â Â Â  <Name>LastName</Name>
> > Â Â Â Â Â  <Value>Rich</Value>
> > Â Â Â Â Â  <Name>SWName</Name>
> > Â Â Â Â Â  <Value>FancySoft v.1.0</Value>
> > Â Â Â  </SWMakerInfo>
> > Â  </SWMaker>
> > Â </ext:ExtensionContent>
> > </ext:UBLExtension>
> This way of modeling data is opposite to the 
> essence of XML, but frequently used by 
> programmers that sells it as "flexible" because 
> it can accommodate future uses with new 
> combinations for elements "Name" and "Value" 
> (and because many programming languages 
> translate object dictionaries to this kind of XML).
>
> The problem arrives when XPath expressions 
> become more complex (although not impossible to write) to access data:
>
> For example, to get the first name:
> > ext:UBLExtension/*:ExtensionContent/SWMaker/SWMakerInfo/Name[ 
> > . = "FirstName" ]/following-sibling::Value[1]
> But this kind of expression is not trustworthy. 
> For example, if the first <Value> element is 
> missing, the LastName will be returned instead.
>
> I'm asking for your help, because we'll have a 
> meeting with the contractor and I need arguments 
> to convince them of changing to a
> "Natural" XML structure like this:
> > <SWMaker>
> > Â  <SWMakerInfo>
> > Â Â Â  <FirstName>Erick</FirstName>
> > Â Â Â  <LastName>Rich</LastName>
> > Â Â Â  <SWName>FancySoft v.1.0</SWName>
> > Â  </SWMakerInfo>
> > </SWMaker>
>
> I'm expecting customer won't be open to accept 
> the change, because the actual structure can do the work.
>
> So, what arguments would you use to convince the customer?
>
> Thanks in advance for all your input. It is very 
> enriching to read your opinions and learn from all your experience.
> Â
>
> --Â
> William David Velasquez
> Creativo de Software
> Creativos Digitales S.A.S.

--
Contact info, blog, articles, etc. http://www.CraneSoftwrights.com/x/ |
Check our site for free XML, XSLT, XSL-FO and UBL developer resources |
Streaming hands-on XSLT/XPath 2 training class @US$125 (5 hours free) |
Essays (UBL, XML, etc.) http://www.linkedin.com/today/author/gkholman |