Minor spelling fixes

acfits.tex

@@ -335,7 +335,7 @@ \section{Introduction}
 \citep[XDF;][]{2001ASPC..238..217S}, FITSML \citep{2001ASPC..238..487T}
 and HDX \citep{2003ASPC..295..221G}; and in the present and future
 (e.g., HDF5 \citep{2011ASPC..442...53A} and NDF \citep{2014Jenness}).
-There are other popular file formats amongst the
+There are other popular file formats among the
 radio and (sub-)millimeter astronomy community such as the Continuum and
 Line Analysis Single-dish Software (CLASS) data format associated with
 the Grenoble Image and Line Data Analysis Software (GILDAS) tools
@@ -462,7 +462,7 @@ \subsection{Poor support for information interchange}
 HDX \citep{2003ASPC..295..221G} was developed around 2002 as a flexible
 way of layering high-level data structures, presented as a virtual XML
 Document Object Model (DOM), atop otherwise unstructured external data stores; this was in
-turn used to develop Starlink's NDX framework, which (amongst other
+turn used to develop Starlink's NDX framework, which (among other
 things) allowed FITS files to be viewed and manipulated using the
 concepts of the NDF format.
 HDF5 \citep{2012ASPC..461..283A} was chosen to accommodate LOFAR's
@@ -935,7 +935,7 @@ \subsubsection{Data Quality}
 
 A data quality model then, should be an ensemble of common statistical
 measures of the type of dataset which may be used to derive
-higher-level judgements of the quality/suitability of the data for
+higher-level judgments of the quality/suitability of the data for
 some other declared purpose. There are many higher types of data
 quality models which will need be created from the lower-level
 measures (image data quality, pointed catalog data quality, etc) and
@@ -1175,7 +1175,7 @@ \subsubsection{No support for declaring byte order}
 (\href{http://ascl.net/1010.001}{ascl:1010.001}) and which use C routines
 to implement other scientific capabilities.  Programmers on little endian
 platforms who work with large data volumes may find that this limitation
-results in a performance penalty as marshalling data to and from
+results in a performance penalty as marshaling data to and from
 the FITS big endian ordering will be required.  This is a frequent problem
 for astronomical programs. Little endianness is found on x86 and x86-64
 processors that are commonly used in universities and research
@@ -1275,7 +1275,7 @@ \subsubsection{No support for parallel write/read operations on HPC platforms}
 \label{subsection_hpc_io}
 
 Large datasets also require parallel read/write operations to be processed on
-parallel high performance computers. FITS can not support optimisation for
+parallel high performance computers. FITS can not support optimization for
 parallel read/write operations.
 
 This has been the driving factor for LOFAR to