12-03-2009, 11:15 AM
I do not want to spill water into anybody's wine here but I am slightly confused as to the purpose of the classification system you developed?
As a system of referring to an individual helmet, it is inconceivable at least to me how somebody could pronounce such a number of "N"s and "Y"s.
As a tool for identifying categories of helmets, it will in the best case only confirm what we already know: After filling in all the data for each helmet and running it through a computer, you will be left with tons of junk data and with exactly the same main categories of helmets (Montefortino, Weisenau/Imperial, Masked/"Sports" etc.) which are traditionally used, except that the system will likely not recognize certain categories (e.g. the Pseudoattic/"Aux Cavalry" category). Even arriving at these categories will require you to look at the helmets and to determine which attributes are really important. The quantity of common attributes is actually not as important as the quality of common attributes ( e.g. a bronze mask helmet will share the exact same NUMBER of attributes with an iron mask helmet and with a bronze Coolus helmet although anybody would agree that it has more in common with the iron mask helmet).
The "traditional" categories are there for good reason and (except for some borderline cases) it is fairly simple to assign a helmet to one of the main categories by simply looking at it. The interesting job is to define subcategories WITHIN each of these main categories and your system will not work for doing that. Actually, given that the extant specimens in each category count in the tens or at the most hundreds, it is actually much more efficient to simply look at each single extant helmet than to feed a list of necessarily broad and unqualified attributes into a data base.
As an addition to the database, the attributes would still be useful if they would allow for specific searches (e.g. of all helmets with enamelled rivets) but that would only be possible if each attribute gets a separate data field of if the system allows for unlimited wildcards (e.g. "search for * * * Y * * * * * * * * N * * * ")
As a system of referring to an individual helmet, it is inconceivable at least to me how somebody could pronounce such a number of "N"s and "Y"s.
As a tool for identifying categories of helmets, it will in the best case only confirm what we already know: After filling in all the data for each helmet and running it through a computer, you will be left with tons of junk data and with exactly the same main categories of helmets (Montefortino, Weisenau/Imperial, Masked/"Sports" etc.) which are traditionally used, except that the system will likely not recognize certain categories (e.g. the Pseudoattic/"Aux Cavalry" category). Even arriving at these categories will require you to look at the helmets and to determine which attributes are really important. The quantity of common attributes is actually not as important as the quality of common attributes ( e.g. a bronze mask helmet will share the exact same NUMBER of attributes with an iron mask helmet and with a bronze Coolus helmet although anybody would agree that it has more in common with the iron mask helmet).
The "traditional" categories are there for good reason and (except for some borderline cases) it is fairly simple to assign a helmet to one of the main categories by simply looking at it. The interesting job is to define subcategories WITHIN each of these main categories and your system will not work for doing that. Actually, given that the extant specimens in each category count in the tens or at the most hundreds, it is actually much more efficient to simply look at each single extant helmet than to feed a list of necessarily broad and unqualified attributes into a data base.
As an addition to the database, the attributes would still be useful if they would allow for specific searches (e.g. of all helmets with enamelled rivets) but that would only be possible if each attribute gets a separate data field of if the system allows for unlimited wildcards (e.g. "search for * * * Y * * * * * * * * N * * * ")
Regards,
Jens Horstkotte
Munich, Germany
Jens Horstkotte
Munich, Germany
