sDNA isn’t the only spatial network analysis software out there!  Here are some others.

• SANET is a toolbox for analysing events that occur on networks or alongside networks.
• SANET is a reinvention of traditional spatial techniques, such as clustering and correlation analysis, for use on network rather than Euclidean space.  SANET serves a slightly different purpose to sDNA; the two software packages complement each other nicely.
• MIT Urban Network Analysis is a similar tool to sDNA, measuring closeness and betweenness on spatial networks.
• Theoretically, the major difference is that UNA uses Address Points (i.e. buildings, origins, destinations etc) as the fundamental unit of analysis, which means the user must either have, or simulate, data which is external to the network itself.  Besides entailing more work for the user, this also means it is difficult to compare results from analyses of two different sets of address points.  sDNA, by contrast, measures the network geometry alone (though if the user wishes to take account of a building layer in the analysis, they can assign custom weights to an sDNA analysis).
• A practical difference is that UNA, although released as a free tool, requires ArcGIS Network Analyst to run (sDNA doesn’t).  This costs up to $2500 on top of ArcGIS itself.  UNA does not support Autocad.
• Depthmap and Confeego are sDNA’s predecessor software developed in-house by Space Syntax Ltd.  The former is for building-level (interior or exterior small-scale) analysis, the latter for city-level analysis.  Depthmap is a stand-alone program, while Confeego runs inside of MapInfo GIS.
• Both tools were originally developed for internal use only at Space Syntax, whose business model was to use proprietary spatial analysis as a unique selling point for their design consultancy.  Depthmap has recently been released as open source software, however, educating clients in the techniques of spatial network analysis is still not a primary concern for Space Syntax (it is for us).
• Technically, both Depthmap and Confeego use axial lines, rather than network links, as a fundamental unit of analysis.  Originally (before the advent of widespread digital mapping data) these were drawn by hand, and the system worked pretty well, but you wouldn’t want to do it now!  More recently, algorithms have been developed to convert existing map data (such as that provided by the Ordnance Survey) into axial lines; however, this leads to issues with repeatability as minor differences in representation of a geographic feature can lead to major differences in the axial line representation and hence the overall analysis.  Another consequence of axial line analysis is that more computation is required; meaning that during a design process, the design-analyse-review loop takes considerably longer, and the spatial network design process becomes less interactive.