有关：XBRL可以从XML/JSON的争论中学到什么？

最近Charles Hoffman在yahoo的xbrl-dev讨论组发了一贴“Can XBRL Learn from the XML/JSON Debate? ”。引起了广泛的讨论。出现了不同的看法和意见，有兴趣的可以登录讨论组阅读。大致的观点有：

(1) 使用XML还是JSON作为XBRL数据载体，只是以什么形式“serialization”序列化的问题范畴；如果有良好的内容模型，形式不是主要问题。讨论的关键是要有一个：completeness/usefulness/redundancy of the XBRL model

(2) 讲到内容模型，有提及tuple和dimension。 有人就极力反对使用dimension，因为这增加了XBRL复杂性，为XBRL的推广增加的阻力（US GAAP分类标准使用dimension）。也有人给了折中的意见。

 

Tuples: group facts together. That means that a good use of a tuple is when
you group facts that makes sense to be grouped together.

Dimensions: add properties to fact items. That means that a good use of a
dimension is when the property can be shared across multiple different item
declarations.

IMO, there is no point in thinking that tuples and dimensions are
interchangeable. The only way to make them interchangeable is by doing an
extreme interpretation of the two modeling techniques:

- Using dimensions instead of tuples is like defining a property ?I
belong to a group?. So the group is a virtual property of the facts.

- Using tuples instead of dimensions is like grouping together
redundant information (defining thousands of tuples and then reusing the
same tuple members several times).

Fortunately nobody has ever tried the second option but the first option is
IMHO also a mistaken approach.

Tuples and dimensions can live together because they are different things.

Cheers,

(3) 统一一致的内容模型，有助于数据和软件本身的互操作行。

....