Barcode images with same content can look different

[martinsXXX]

I'm wondering why the datamatrix generated by terryburton.bwipp looks different in some cases to that generated by https://barcode.tec-it.com Web site.
Input data is +E247E7200A0/$$7DBSWIN5.15.0AT2
The call is like this (where %HIBCDATA is replaced with the inputdata):
4.15 4.15 scale 190 54 moveto (%HIBCDATA%) (parse) /datamatrix /uk.co.terryburton.bwipp findresource exec
This is what is the result:

This is the result from https://barcode.tec-it.com/de/DataMatrix?data=%2BE247E7200A0%2F%24%247DBSWIN5.15.0AT2:

They have same #rows and #columns, but look different. When decoded with my handy app, they both yield the same result.
Q1: Can anybody explain why?
Q2: Are both valid?

When we set input data to +E247E72000/$$7DBSWIN5.15.0AT2 (E7200 instead of E7200A), then both results look the same.

Thanks in advance

[terryburton]

They have same #rows and #columns, but look different. When decoded with my handy app, they both yield the same result.
Q1: Can anybody explain why?

There is freedom of choice over how to internally encode data in a Data Matrix symbol. BWIPP attempts to optimise the size of the internal encoding of the given data to minimise the resulting symbol size.

Q2: Are both valid?

The output of a scan of a barcode is essentially a set of bytes. These symbols will scan identically.

Some specialised applications of general-purpose data carriers specify certain restrictions on how the data is encoded, often to ensure stability of the symbol size, however this practise is discouraged both by the specifications and industry best practise because there are better ways to ensure this.