implement comments from Mihail

thanks!

Author: exaexa

lib/metadata.tex

@@ -2,6 +2,7 @@
 \author[1]{Miroslav Kratochvíl}
 \author[2]{St.~Elmo Wilken}
 \author[3,4,5,6]{Andreas Dräger}
+\author[7]{Mihail Anton}
 \author{Please Add Yourself Here}
 %\author[999]{Please add yourself if you contribute}
 \affil[1]{Luxembourg Centre for Systems Biomedicine, University of Luxembourg, LU}
@@ -10,5 +11,6 @@
 \affil[4]{Department of Computer Science, Eberhard Karl University of Tübingen, 72076 Tübingen, DE}
 \affil[5]{German Center for Infection Research (DZIF), partner site Tübingen, DE}
 \affil[6]{Cluster of Excellence `Controlling Microbes to Fight Infections,' Eberhard Karl University of Tübingen, 72076 Tübingen, DE}
+\affil[7]{Department of Life Sciences, National Bioinformatics Infrastructure Sweden, Science for Life Laboratory, Chalmers University of Technology, SE-41296 Gothenburg, SE}
 %\affil[999]{Please add your affiliation if you contribute}
 \date{\today}

text/spec.tex

@@ -109,7 +109,7 @@ \subsection{TSV}
 \subsubsection{Numeric and special values in TSV files}
 \emph{Numeric} fields must be converted to decimal representation. The decimal representation may include a leading minus sign, decimal point, and an integer exponent separated by \verb|e| (for example, valid encodings of $-1234.5\times10^2$ include \verb|-123450|, \verb|-12.345e4|, and \verb|-12345000e-2|). Rounding of non-representable and infinite numbers may be performed, but it should preserve the outcome of the report comparison, as described in~\cref{sec:cmpdata}.
 
-Missing numeric values must be represented by an empty string. Missing values typically occur when a value is expected in a field (such as in the place of the optimal value of the model), but it can not be computed for some reason (such as model infeasibility).
+Missing numeric values must be represented by an empty field. Missing values typically occur when a value is expected in a field (such as in the place of the optimal value of the model), but it can not be computed for some reason (such as model infeasibility).
 
 \emph{Solver status} fields contain a special value that describes the outcome of an optimization algorithm. FROG currently describes only two kinds of statuses:
 \begin{itemize}
@@ -167,13 +167,17 @@ \subsection{Metadata file}
  & optional
  & URL of a website or public repository where the model is developed, preferably offering a chance to contact the model authors or submit corrections of the model
  \\
+ \verb|model_reproducibility_url|
+  & optional
+  & URL of a website or a public repository that gives closer information on how to reproduce the results obtained using the model, such as a link to a reproducible notebook or to a test suite
+  \\
 \verb|model_version|
  & optional
  & version string of the model, if available
  \\
 \verb|environment|
  & required
- & free-text description of the software environment used to run FROG analysis tools, such as from \verb|uname -a|
+ & free-text description of the software and hardware environment used to run FROG analysis tools, such as the versions of operating system, the hardware platform and the software tools; if possible, this includes a pointer to a machine-readable information that reproduces the environment, such as a versioned docker image name, or a version-lock file for a package management system
  \\
 \verb|software|$\to$\verb|frog|
  & required

text/test.tex

@@ -71,7 +71,7 @@ \section{Metadata comparison}
 \midrule
 \verb|frog_version|
  & consistent
- & generated report semantics may differ
+ & generated report semantics may differ among incompatible FROG versions; consistency may be checked with respect to semantic versioning~\cite{semver}
  \\
 \verb|frog_id|
  & informative
@@ -99,7 +99,7 @@ \section{Metadata comparison}
  \\
 \verb|model_version|
  & consistent
- & compared models likely differ
+ & compared models of different versions likely differ; consistency may be checked with respect to semantic versioning~\cite{semver}
  \\
 \verb|environment|
  & informative