I wanted to sum up some principles about my understanding of how categories are/should be assigned.
The 4 principles extracted by Assif from the assignment of WHS to categories as currently existing are "correct" IMO.
As the other 3 suggestions identify, what we perhaps lack at the moment is a similar set of "principles" as to how the "Categories" themselves should be determined. This is not of course totally independent from the rules as to how the assignment should be done since it is a combination of the 2 sets of rules which creates the total "assignment"
Among those which I remember were things like
a. Try to avoid categories with "too few" WHS - ok some WHS are "unique" but most are comparable in some way with others -and, since identification of such common aspects is one of the main reasons for "categorisation", we should not lose this by dividing into too many different ones. We never formally defined what "too few" might be but I guess 5 or 6 was an ideal minimum if not a hard rule. "Bahai" and "Caribbean" have slipped in with a single rep only and I feel should be reconsidered.
b. Try to avoid Categories with "too many" WHS. If a category gets up to 40 or 50 it begins to lose meaning and hints that a tighter definition might be worth having. "CL" and "Christian" are the worst at the moment and any reconsideration of category definitions should attempt to address these (As Assif does for the CL category!)
c. Try to avoid purely "geographical" categories - ie "where" they are rather than "what" they are. Whilst a geographical location can act as a proxy for Cultures/civilisations it seemed better to use these latter. Some of the current ones which might "sound" Geographical (E.g "Latin American") are in reality "cultural" but there are also a few which perhaps should be renamed to emphasise their "cultural" identity (E.g "Pacific".
d. Try to avoid having too many Categories - around 60 was our target - giving an average of around 15-20 per category
No doubt others were used or "emerged" and Assif has suggested 3 more. The first 2 seem to be an improvement and it would be interesting to "work them through" with some examples to identify and "iron out" difficulties. The 3rd one raises a valid issue and presents 2 possibilities without of course proposing one or the other!!
As I mentioned earlier, the current definitions and assignments were largely created over 10 years ago and rather got "overtaken" as a subject of interest by "Connections" and were never really critically assessed by a "body" of "experts"! I think the current exercise has shown that the ability to compare WHS in terms of "What they are" at approximately this level of granularity (in addition to our other 2 "comparison" lists - by "period of creation" and by "Detailed attribute"- ie "Connections") IS worth having. Also, that the original definitions and assignments have stood up reasonably well in the face of their first real "test" but are certainly both capable of being improved!
One other possibility I would throw in is the possibility of having "subcategories" (1 level!!) within the categories. This would require a bit of work by Els but is not I think a major matter of principle (???? Els!!). "Category" would still constitute the "Index" and could still appear on the page for each site as defining its category (but would it be difficult to create a single "title" made up of Cat and subcat?). However, within a Category, the grouping and sequencing would be by subcategory - thus permitting e.g "Pre-Colombian" to be divided in any number of smaller groups we wanted/found significant or useful. It wouldn't matter then if for instance there was a subcategory of 1. This could be a way of addressing the problem posed by Assif in his 3rd proposal?? We could still easily compare WHS across the wider Category whilst differentiating them within it. So e.g the Meltwaterfalls suggestion for handling the large Christian category which we have adopted for this exercise (ie. dividing them into Churches, Cathedrals, Residential etc) could be easily adopted permanently whilst still maintaining the overall "Christian" category to facilitate comparison with number of sites for other major religions which might not need such a sub-categoristion on their smaller numbers (though it would cause no real problem to adopt similar categories for at least some of them too or, alternatively, ones more appropriate to the religions such as "Temples", "Shrines" etc