Page references to PDF of version 2.0.2.
"and the encoder" => "and encoders" (cf. "their" below).
P. 420: "Arne Magnusson"
"Arne Magnussen" is the danicised name form:
It is the form used in the 1891 publication of his anonymous tract from 1699 on a Danish witch process: "Kort og sandfærdig Beretning om den vidtudraabte Besættelse udi Thisted til alles Efterretning af Original-Akter og troværdige Dokumenter uddragen og sammenskreven af Arne Magnussen …".
In a Danish context, while "Arne Magnussen" would be the form he generally used himself, his use of the form "Arnas Magnussen" is also attested. In Islandic, he would presumably have used the form "Arne Magnusson" (possibly "Arni Magnusson"), without accents. (Thanks to Guðvarður Már Gunnlaugsson for this information.)
The mixed form is also widely used in Denmark, as in "Arne Magnussons private Brevveksling" from 1920 and Hans Bekker-Nielsen and Ole Widding's biography "Arne Magnusson, den store håndskriftsamler, i 300-året for hans fødsel" from 1963 (the English translation 1972 uses the same form).
I would delete ", the form which he himself, as a life-long resident of Denmark, generally used". Since Iceland was part of Denmark 1380-1944, of course he was a resident of Denmark. If this is not deleted, since he only lived in Southern Denmark about half of his life, we should at least give "long-time" instead of "life-long".
P. 421: "In addition to these specific elements"
These elements must be <faith> to <floruit>, mentioned three pages back; however, on the preceding page, the generic element <trait> is introduced. This gap seems too wide.
Presumably, the paragraph on <persName> on p. 420 should be moved to p. 419, before the paragraph introducing <trait>.
I suggest deleting "to these specific elements".
The paragraph "This element … of the holder" more or less repeats "Where there is confusion … of the holder" on p. 420. This important paragraph should be rewritten and moved to the bottom of p. 415, following "or that of others".
"All the generic elements are also members of the att.responsibility and att.editLike classes. These classes, originally intended to provide attributes ‘describing the nature of an encoded scholarly intervention or interpretation of any kind’ make available the attributes cert, to indicate the degree of certainty, resp, the agency responsible, and evidence, the nature of the evidence used."
"‘describing the nature of an encoded scholarly intervention or interpretation of any kind’" is the definition of att.editLike, not of att.responsibility, so the definition does not cover "These classes".
It is not clear to me what "originally intended to" means here.
"or all the participants in a mutual one."
So <relation name="spouse" active="#H1 #H2"/> is the same as <relation name="spouse" mutual="#H1 #H2"/>? It makes sense (for marriages that are not arranged), but is this overlap intentional? Is <relation name="sibling" passive="#H1 #H2"/> then possible for <relation name="sibling" mutual="#H1 #H2"/>?
P. 427, note:
The link is dead. The most recent revision of WGS84 is from 2004. See <https://www1.nga.mil/ProductsServices/GeodesyandGeophysics/WorldGeodeticSystem/Pages/default.aspx>.
As far as I can make out, EGM96 is something different (<http://en.wikipedia.org/wiki/EGM96>).
The address of Brasserie Georges is not correct. It should be:
<placeName type="street">Cours Verdun</placeName>
P. 429: "the Geographical Markup Language (GML) being defined by the OGC".
I think "OpenGIS Geography Markup Language" is the more correct name. Instead of "OGC", write "Open Geospatial Consortium (OGC)".
P. 429: "now used by Google Maps" gives the impression that Google Maps used something different before, which is not the case. The idea behind this must be that KML was not invented by Google, but by the company Keyhole, which Google bought in 2004. This is only clear to those who know it already.
I suggest deleting "now".