Eurocode 2 Discussion Note N006

Classification of milk and cheese


[Eurocode 2 Home Page] [Eurocode 2 Discussion Group] [Eurocode 2 Proposals] [Eurocode 2 Feedback Form]

ContentsIntroductionComments by:31 Jan 00
Background
Milk classification, version 99/1
Classification of milk products and substitutes
Fat content criteria for milk products
Classification of cheeses by type
Classification of cheeses by fat content
Conclusions and future work
Request for comments and suggestions
 
InformationCheese subgroups, version 93/1
Cheese subgroups, version 99/1
Cheese subgroups, version 99/2
Fat content data for cheeses
CODEX classification for cheeses
Links to external cheese information


Introduction

Eurocode 2 version 99/1 modified the classification of milks by introducing a subgroup for Cow milks and a second one for Other milks. This approach was proposed so that the milk-producing species could be recorded; this was not possible in the earlier version 93/1. However during discussion of Eurocode categories at the meeting of the COST Action 99 - Eurofoods Working Group on Food Description held in Paris, December 1999, it was concluded that division by species should not be made at such a high level in the hierarchy. A higher-level division that reflected fat content was more important.

The classification of cheeses in version 99/1 retained that in version 93/1, in part at least while awaiting comments on the types of revision being proposed overall during the Eurocode Review project. The Paris meeting suggested that the options for dividing cheeses into categories should be considered further in a discussion note.

Since some similar aspects apply to the classification of milk and of cheese, and it is helpful to record the various considerations in defining the classification within the major Milk and milk products subgroups, this note has been produced to collect and combine the discussions.

Background

The following pages provide information on the Eurocode 2 version 93/1 Milk and milk products main group and its development: The Rationale Document also provides useful background on other aspects taken into consideration in the Eurocode 2 version 93/1 Milk and milk products main group; this could also be added to the Website's archive material if this would be helpful.

Milk classification, version 99/1

Eurocode version 93/1 did not record the species for milks, preferring to use the next level in the hierarchy to record fat content. The extra level in version 99/1 allowed milks to be classified by species and then fat content, either as Milk => Cow milk => Cow milk, 3 - 4% fat or Cow milk => Cow milk, 3 - 4% fat => Cow milk, 3 - 4% fat, UHT.

The version 99/1 proposal separated Cow milk and Other milks so that the number of top-level subgroups was kept to a minimum while allowing the fat content for Cow milk to be applied at the next level down. This was seen as a temporary solution in case a unified approach to recording fat content using descriptors is developed. Then a single subgroup could be created for Liquid milks, divided by species at the next level. The subgroup would include all 'fresh' milks, including forms such as sterilised, UHT, etc., with descriptors defined to record such types of processing.

Classification of milk products and substitutes

The Milk and milk products subgroups are based on product type and most of these are subdivided by fat content. In version 99/1, the milk category was divided on the basis of source species as described above, but it was generally agreed that species was the less important criterion. Therefore in version 99/2, most of the group will conform to a categorisation framework that divides it by product type at the subgroup level, then by fat content, and finally, if required, by species. To avoid the need to enumerate every possible species category, a standard set of codes has been defined. Generally within the current classification this would be applied at the fourth level, e.g. 1.10.40.10 - Skimmed cow's milk. However the Processed milks subgroup is further subdivided by product type before fat content is applied and if species is required here, it will be at the next lower level, e.g. 1.15.30.10.30 - Dried whole sheep's milk.

The subgroups of the milk products group are categorised on the basis of product type which can be considered to include the characteristic use of the food item. Thus if a product is used as a substitute for animal milk, it should be categorised as a milk. On this basis, soya milk and cocolait (coconut milk processed for use instead of cow's milk) would be classified in the Milks group as milk substitutes whereas coconut milk itself would be considered a vegetable product.

Version 99/1 introduced an Imitation milk and cream subgroup for substitute products (which included Soya yogurt and Soya cheese). This remains inconsistent with the overall structure of the group, as outlined above, in which the first level is based on product types such as milk, yogurt or cheese. For full consistency, classification of the soya source could be relegated to the fourth level. In that case, Soya milk (1.9% fat) would be coded as 1.10.30.70. Alternatively, the 'Imitation' subgroup could be viewed as a separate product 'type' which would be subdivided on a second product type basis into substitute milk, cheese, etc. (rather as the Processed milks are given a further product type subdivision). Again, for the consistency, this should be categorised as Imitation milk products => Substitute milk => Substitute milk, 1 - 2.9% fat => Soya milk (e.g. 1.50.10.30.70) (endnote), rather than Imitation milk and cream => Soya milk (i.e. 1.50.10), as in version 99/1. Any comments on these alternatives would be very welcome.

Fat content criteria for milk products

It is proposed to continue to use fat content as the criteria below product type in the majority of Milk and milk products subgroups. In versions 93/1 and 99/1, subdivision has been into from 2 to 5 fat content categories chosen to be appropriate to the subgroup and generally topped by a 'greater than' category down to a 'less than' category. In most subgroups the criterion is cited as percent fat in the food, but for cheese the figures are percent fat by dry weight. The fat percentage ranges used for categorisation should as far as possible define distinct products such as semi-skinned or half-fat milk. The ranges used in version 93/1 should at some stage be compared with composition data in several European food tables to check that they remain optimised for this.

Classification of cheeses by type

In the Crete workshop discussion, a preference was stated for categorising cheeses by criteria such a method of production rather than on fat percentages which were often very similar for different cheeses. It was also suggested that individual cheeses should be assigned to their place in the classification so that codes could be looked up using the cheese name in the index. However the Rationale Document and the version 93/1 classification applied a very simple classification by type (into Hard cheese, Soft cheese, Fresh cheese). Although the Rationale Document states that this approach was taken after the investigation of several alternatives, there may be advantages in a revised cheese type criterion. Subdivision by fat content, discussed below, may also be improved by revising the cheese type categories.

Version 99/2 proposes seven cheese subgroups: Fresh cheese, Soft cheese, Semi-hard cheese, Hard cheese, Blue cheese, Smoked cheese and Processed cheese. Descriptions of these types have been drafted and comments are welcome on any advantages or disadvantages compared to the set of three simple types (Fresh, Soft and Hard). A possible advantage of distinguishing between Semi-hard and Hard may be that it separates out cheeses with differing percentages of water, as suggested in the Crete workshop. An alternative faceted classification of cheeses contained in the CODEX General Standard for Cheese uses separate factors for firmness, fat content and the principal curing characteristics.

In practice the subdivision into cheese types should probably be moved down a level, below a general subgroup category of 'Cheese'. Are there other possible cheese categories that should be included, for example Whey cheese?

Classification of cheeses by fat content

Version 93/1 further subdivided the three cheese types by fat percentage criteria appropriate to each type. This resulted in some specific cheeses being assigned to two or even three categories, for example Brie as 'Cheese, soft > 60% fat', 'Cheese, soft 46 - 60% fat' and 'Cheese, soft 30 - 45% fat'. Where products spanned two categories, this might have been because typical fat contents approximated to the boundary between the categories. For example, as some sample data show, many hard cheeses (many of which would be assigned to the 99/2 'Semi-hard cheese' category) have close to 50% fat on a dry-weight basis. When the product type categories for cheese are finalised, the set of data should be extended and organised by the revised categories so that appropriate fat content ranges can be identified. Where products spanned three categories, the data used may have covered distinctly different products, perhaps including low-fat varieties of the cheese which should be given more specific names.

Conclusions and future work

The proposals for handling information on milk source species in version 99/2 have introduced the idea of a careful structuring of the categories in the Milk and milk products main group. Subdivision is first by product type, then fat content and finally source species. However within the 'Processed milks' subgroup, a second product type level is required and, as noted above, the same approach be taken with a 'Cheese' subgroup that has subcategories such as 'Soft cheese'. This suggests that a structure with two product type levels, followed by fat content and source species levels should be implemented throughout the main group. This might be useful in further subgroups, for example creams might be subdivided into categories for fresh, UHT and soured cream. The approach should also simplify retrieval and data aggregation based on criteria lower in the hierarchical structure. It assumes that category assignment will be computer-supported to avoid the input of long codes and indeed should allow more flexibility in the support application through the possibility of a semi-faceted implementation.

The Core Classification as presently defined includes the subgroups and sub-subgroups and thus equates with the two product type levels proposed for the Milk and milk products main group. Therefore the type categories will be defined in the present project but the subsequent task of subdividing these into suitable fat content categories is considered future work. This will involve identifying the most useful (and most easily applied) fat ranges for each product type category. It should be based on an extensive international collection of fat and moisture data for individual cheeses that will have been assigned to product type categories. Extracts from the data collection can now easily be extracted into HTML pages, as demonstrated by the sample data.

If and when the use of longer but more structured codes is shown to improve the categorisation, the approach might be applied to other main groups. The gains in flexibility are considerable, allowing new categories to be created by users from existing criteria and permitting intermediate criteria to be skipped as unassigned. For example, the code for the general food 'Buffalo cheese' can be generated as required rather than being predefined (endnote). Other groups would require different sequences of subdivision criteria to maintain compatibility with generally accepted food grouping hierarchies, but a structure key for each group would allow retrieval systems to locate comparible criteria in different groups, for example the source species criterion. Relevant (or available) information identifying the food could be recorded even where intervening criteria are unknown and a wide range of aggregation groupings could be derived, including very general categories such as 'Sheep products'. The necessary revision and clarification of the Milk and milk products classification, particularly with regard to milks and cheeses, provides an opportunity to test the further possibilities.

Request for comments and suggestions

Comments on the above notes, together with any further information or suggestions for specific categories, would be very welcome. Please return your comments using the project feedback form, if possible by 31 January 2000.


Endnotes

Unspecified criteria

It is possible that the information on criteria at the intermediate levels may not be known but it is required to record the low level information, in this case the source. The new specification of codes only assigns number in the range 10 to 79 to specific categories, leaving the codes 00 to 09 and 80 to 99 available for special purposes. One clear requirement is for a code representing 'unspecified' at an intermediate level. For example, '00' could be assigned to this. Soya milk of unspecified fat content would then have the code 1.10.00.70 (or 1.50.10.00.70 under the alternative proposal). Any study requiring to record species but not fat content would then be able to use the classification, e.g. Buffalo milk has the code 1.10.00.14. In effect, use of the '00' code converts the Milk and milk products group to a faceted classification. Alternatively, if codes with unspecified criteria are preferred to sort to the end, a high number such as '85' could be used. The Unassigned code could also be used to skip the second level product type criterion where this is unknown or not relevant. For example, Buffalo cheese might be assigned the code 1.40.00.00.14.   – Return: milk classification;  conclusion

 


 


Ian Unwin
Updated:  21 January 2000