The hoard of clutter within someone's memories

Automation and poka-yoke

February 28th 2026

The thing to beat the tedium of typing out things for computing is to create things that automate repeated actions. As the new big-ish project is designed to use XML, finding ways to ease the burden of creating XML files for the end user is always going to be appreciated. Although the examples used are small at the moment because everything is still at a 'proof of concept' phase, automating the production of XML files in this way will work when larger examples need to be created.

And it is not just the workload aspect. Minimising manual work for the user also reduces the scope for possible mistakes, the concept of poka-yoke.

There are many ways to do this, I am going to use bash for this example. The following script creates both an XML file and a dtd file from a number of text files containing input data. If the script automatically assignes an id number using a counter then files containing the details of joints within the framework need only define each joint using lines of x and y coordinates. Different types of joint can be defined using different text files. Again, if id numbers are allocated using a counter within the script then structural members need only be defined by the joint numbers that they are connected to. Forces are defined by joint id, magnitude and angle so each line will be three numbers.

The following bash script creates both xml and dtd files from defined user input files:

#!/usr/bin/env bash

echo "<?xml version=\"1.0\" encoding=\"UTF-8\"?>">frameworktest.xml

echo "<!DOCTYPE framework SYSTEM \"frameworktest.dtd\">">>frameworktest.xml

echo "<framework>">>frameworktest.xml

echo "    <title>Test Framework 1</title>">>frameworktest.xml

echo "        <joints>">>frameworktest.xml

RJointfile=./rigidjoints.txt

SJointfile=./simplejoints.txt

Memberfile=./members.txt

Forcefile=./forces.txt

counter=-1

# Read the input file line by line using a for loop

IFS=$'\n' # set the Internal Field Separator to newline

for LINE in $(cat "$RJointfile")

do

echo "          <rigidjoint>">>frameworktest.xml

    counter=$((counter+1))

    IFS=$','; split=($LINE); unset IFS;

    echo "                <id>$counter</id>">>frameworktest.xml

    echo "                <xcoord>${split[0]}</xcoord>">>frameworktest.xml

    echo "                <ycoord>${split[1]}</ycoord>">>frameworktest.xml

echo "          </rigidjoint>">>frameworktest.xml

done

IFS=$'\n' # set the Internal Field Separator to newline

for LINE in $(cat "$SJointfile")

do

echo "          <simplejoint>">>frameworktest.xml

    counter=$((counter+1))

    IFS=$','; split=($LINE); unset IFS;

    echo "                <id>"$counter"</id>">>frameworktest.xml

    echo "                <xcoord>"${split[0]}"</xcoord>">>frameworktest.xml

    echo "                <ycoord>"${split[1]}"</ycoord>">>frameworktest.xml

echo "          </simplejoint>">>frameworktest.xml

done

echo "        </joints>">>frameworktest.xml

echo "        <strmems>">>frameworktest.xml

counter=-1

IFS=$'\n' # set the Internal Field Separator to newline

for LINE in $(cat "$Memberfile")

do

echo "          <strmem>">>frameworktest.xml

    counter=$((counter+1))

    IFS=$','; split=($LINE); unset IFS;

    echo "                <id>"$counter"</id>">>frameworktest.xml

    echo "                <startjoint>"${split[0]}"</startjoint>">>frameworktest.xml

    echo "                <endjoint>"${split[1]}"</endjoint>">>frameworktest.xml

echo "          </strmem>">>frameworktest.xml

done

echo "        </strmems>">>frameworktest.xml

echo "        <forces>">>frameworktest.xml

IFS=$'\n' # set the Internal Field Separator to newline

for LINE in $(cat "$Forcefile")

do

echo "          <force>">>frameworktest.xml

    IFS=$','; split=($LINE); unset IFS;

    echo "                <jointid>"${split[0]}"</jointid>">>frameworktest.xml

    echo "                <magnitude>"${split[1]}"</magnitude>">>frameworktest.xml

    echo "                <angle>"${split[2]}"</angle>">>frameworktest.xml

echo "          </force>">>frameworktest.xml

done

echo "        </forces>">>frameworktest.xml

echo "</framework>">>frameworktest.xml

echo "<?xml version=\"1.0\" encoding=\"UTF-8\"?>">frameworktest.dtd

echo "<!ELEMENT framework (title, joints, strmems, forces)>">>frameworktest.dtd

echo "<!ELEMENT title (#PCDATA)>">>frameworktest.dtd

echo "<!ELEMENT joints (rigidjoint*, slidexjoint*, slideyjoint*, simplejoint+)>">>frameworktest.dtd

echo "<!ELEMENT rigidjoint (id?, xcoord?, ycoord?)>">>frameworktest.dtd

echo "<!ELEMENT slidexjoint (id?, xcoord?, ycoord?)>">>frameworktest.dtd

echo "<!ELEMENT slideyjoint (id?, xcoord?, ycoord?)>">>frameworktest.dtd

echo "<!ELEMENT simplejoint (id?, xcoord?, ycoord?)>">>frameworktest.dtd

echo "<!ELEMENT strmems (strmem+)>">>frameworktest.dtd

echo "<!ELEMENT strmem (id?, startjoint?, endjoint?)>">>frameworktest.dtd

echo "<!ELEMENT forces (force*)>">>frameworktest.dtd

echo "<!ELEMENT force (jointid?, magnitude?, angle?)>">>frameworktest.dtd

echo "<!ELEMENT id (#PCDATA)>">>frameworktest.dtd

echo "<!ELEMENT xcoord (#PCDATA)>">>frameworktest.dtd

echo "<!ELEMENT ycoord (#PCDATA)>">>frameworktest.dtd

echo "<!ELEMENT startjoint (#PCDATA)>">>frameworktest.dtd

echo "<!ELEMENT endjoint (#PCDATA)>">>frameworktest.dtd

echo "<!ELEMENT jointid (#PCDATA)>">>frameworktest.dtd

echo "<!ELEMENT magnitude (#PCDATA)>">>frameworktest.dtd

echo "<!ELEMENT angle (#PCDATA)>">>frameworktest.dtd

Any other type of joint, such as sliding joints, can be added with another do-loop routine which can be done with minor adjustment with a bit of copy and paste. This is what scripting is meant to do, provide small additional functionality that can have big impacts via automation. This little script will make creating XML files for input faster and easier with fewer opportunities to make mistakes.

Now back to the main feature.