Re: [xml-dev] XML Turing test

[Thomas Passin <list1@tompassin.net>]

On 2/13/2023 2:10 PM, Roger L Costello wrote:
> Hi Folks,
>
> I have XML documents with this form:
>
> <Document>
>      <Pencil_Manufacturer>______</Pencil_Manufacturer>
> </Document>
>
> The allowable values of Pencil_Manufacturer are: Staedtler, Faber, and 
> Camlin.
>
> I have an application that processes XML instance documents containing 
> those values. For this instance document:
>
> <Document>
>      <Pencil_Manufacturer>Staedtler</Pencil_Manufacturer>
> </Document>
>
> the application outputs: 1
>
> For this instance document:
>
> <Document>
>      <Pencil_Manufacturer>Faber</Pencil_Manufacturer>
> </Document>
>
> the application outputs: 2
>
> And for this instance document:
>
> <Document>
>      <Pencil_Manufacturer>Camlin</Pencil_Manufacturer>
> </Document>
>
> the application outputs: 3
>
> I am invited to do an XML Turing test. I am in one room. A colleague is 
> in another room. The rooms are connected by a closed door. I am asked to 
> determine whether my colleague has a copy of my application.
>
> So I write this
>
> <Document>
>      <Pencil_Manufacturer>Staedtler</Pencil_Manufacturer>
> </Document>
>
> on a piece of paper and slip it under the door. A moment later I receive 
> under the door a piece of paper and it says: 1
>
> Next, I write this
>
> <Document>
>      <Pencil_Manufacturer>Faber</Pencil_Manufacturer>
> </Document>
>
> on a piece of paper and slip it under the door. A moment later I receive 
> under the door a piece of paper and it says: 2
>
> I write this
>
> <Document>
>      <Pencil_Manufacturer>Camlin</Pencil_Manufacturer>
> </Document>
>
> on a piece of paper and slip it under the door. A moment later I receive 
> under the door a piece of paper and it says: 3
>
> I conclude that my colleague has a copy of my application and is feeding 
> my XML instance documents into the application, getting the results, 
> writing down the results, and slipping them under the door.
>
> I open the door and talk to my colleague. I am surprised to learn that 
> he is not using my application. Each XML document that he received, he 
> converted to this form:
>
> <Document>
>      <Umbrella_Manufacturer>______</Umbrella_Manufacturer>
> </Document>
>
> My colleague performed these value conversions:
>
> Staedtler --> Totes
> Faber --> Pogessi
> Camlin --> Dynateck
>
> That is, he converted this XML instance:
>
> <Document>
>      <Pencil_Manufacturer>Staedtler</Pencil_Manufacturer>
> </Document>
>
> to this XML instance:
>
> <Document>
>      <Umbrella_Manufacturer>Totes</Umbrella_Manufacturer>
> </Document>
>
> He then fed that XML instance into his application, which output: 1
>
> He converted this XML instance:
>
> <Document>
>      <Pencil_Manufacturer>Faber</Pencil_Manufacturer>
> </Document>
>
> to this:
>
> <Document>
>      <Umbrella_Manufacturer>Pogessi</Umbrella_Manufacturer>
> </Document>
>
> and fed that XML instance into his application, which output: 2
>
> And he converted this XML instance:
>
> <Document>
>      <Pencil_Manufacturer>Camlin</Pencil_Manufacturer>
> </Document>
>
> to this:
>
> <Document>
>      <Umbrella_Manufacturer>Dynateck</Umbrella_Manufacturer>
> </Document>
>
> and fed that XML instance into his application, which output: 3
>
> The behavior of my application on my pencil XML documents is identical 
> to the behavior of my colleagues’ application on the XML generated by 
> converting my pencil XML documents to umbrella documents.
>
> The pencil XML vocabulary maps to the umbrella XML vocabulary.
>
> So it doesn't matter whether you use the pencil XML instances and my 
> application or the umbrella XML instances and my colleagues’ 
> application. You get the same results.
>
> My colleague’s XML vocabulary and application passed the XML Turing test.
>
> Thoughts?
>
> /Roger

Aside from the fact that this "test" is nothing like a Turing Test, you 
don't explain how your application produces its numbers.  Maybe it 
produces a serial count of submissions.  If so, the other end could be 
exactly the same application as yours.  If you are going to provide us 
with a black box, then it had better be specified well enough as to how 
to handle all or most inputs.  Otherwise how can we tell if the other 
black box will behave in matching ways?