Talk:Naming Conventions

From XBRLWiki

Revision as of 08:09, 12 October 2012 (edit) Hommes (Talk | contribs) ← Previous diff		Current revision (10:57, 25 October 2012) (edit) Hommes (Talk | contribs) (→Comment-18)
Line 32:		Line 32:
	=== Comment-11 ===		=== Comment-11 ===
-	RH: Is there a list of owners? Is there a desired limit to the length of the ns-prefix?	+
		+	=== Comment-12 ===
		+	RH: Is there a desired limit to the length of the ns-prefix?
		+
		+	=== Comment-13 ===
		+	RH: Referring to custom codes through a linkbase mechanism is fine, but using a label linkbase is not. These are not labels as intended by XBRL 2.1 (for human consumption). Why not using a proprietary linkbase with a dedicated arcrole expressing the exact meaning of the relationship?
		+
		+	=== Comment-14 ===
		+	RH: Why are lower and upper case on domains needed? Rather keep URI's in lowercase to prevent any problems.<br/>
		+	RH: Is it wise to have all the member schemas named 'mem.xsd'?<br/>
		+	RH: What is the added value of another layer in the location that expresses the same thing as the name as the schema? (met/met.xsd, why not just met and add the extension .xsd for the location)<br/>
		+	RH: What is the ns-prefix for the hier.xsd?<br/>
		+	RH: Is it wise to have all the member hierarchy schemas named 'hier.xsd'?<br/>
		+
		+	=== Comment-15 ===
		+	RH: The assigned letter 'p' is used for both percentages and decimals. Both are also supported by the measure 'xbrli:pure'. How the differentiate between the two? I suggest using a) different letters and b) using UN-Cefact measures for percent: un66411:P1 which frees up xbrli:pure for decimal or integer.
		+
		+	=== Comment-16 ===
		+	RH: Can we put 'any xml datatype' on the Typed domain?<br/>
		+	RH: There will be no measure on a typed domain member, that can be left out?<br/>
		+	RH: The source document describes an extra attribute (@model:domain) to be included to identify the Qname of the domain. Why is this needed, and where is this extra attribute placed?
		+
		+	=== Comment-17 ===
		+	RH: Why make the URI on ELR so complicated? Simply use URN's. E.g. urn:eba:perspectives:1. In namespaces the URI is really an identifier, identifying the elements in the schema, but in linkroles the URI is just a name, there is no pointer from outside XML that can point to it (like with an URL which is mostly derived from the URI by placing the file name there or just the file extension added).
		+
		+	=== Comment-18 ===
		+	RH: Dimensions and domains are both made up by two letter capitals. Are they allowed to overlap? MUST domain CA be in dimension CA?
		+
		+	=== Comment-19 ===
		+	RH: A tableGroup hierarchy is created in a generic linkbase with a custom arcrole. Yet the linkbase is named tab-pre.xml. This name is hinting at a parent-child presentation linkbase. Better use something else.<br/>
		+	RH: The custom arcrole for tableGroup - table has @cyclesAllowed=undirected. I think that should be 'none'. There should be no cycles in any form between tables and tablegroups because tablegroups can already be parent and child.

---

Current revision

Contents 1 Comments 1.1 Comment-01 1.2 Comment-02 1.3 Comment-03 1.4 Comment-04 1.5 Comment-05 1.6 Comment-06 1.7 Comment-07 1.8 Comment-08 1.9 Comment-09 1.10 Comment-10 1.11 Comment-11 1.12 Comment-12 1.13 Comment-13 1.14 Comment-14 1.15 Comment-15 1.16 Comment-16 1.17 Comment-17 1.18 Comment-18 1.19 Comment-19

Comments

Comment-01

RH: Do we have a limited list of 'owners' that can be prescribed?

Comment-02

RH: How do we number the rules uniquely?

Comment-03

RH: I would like to emphasize that having reasons for each rule prevents a lot of questions. I.e. The reason for folder names to be lower case is to prevent problems between software running on Unix or Microsoft server.

Comment-04

RH: In a picture supplied in document 'eba-dpm-xbrl-mapping' more subfolders are presented than are explained in the text. Maybe DTS authors are free in creating extra layers within the 'dict' and 'releasedate' folders?

Comment-05

RH: I do not understand why the dictionary folders are not part of a version or release date. And why it is necessary to have a folder per schema. If there are multiple fam.xsd, met.xsd etcetera there may be a use otherwise a 1:1 has been created.

Comment-06

RH: A lot of new (to XBRL) terms are introduced, must they be linked to the definition page?

Comment-07

RH: Are the new terms agreed upon by the participants or still under review?

Comment-08

RH: There will be no divide in label and reference linkbase(name)s based on the role?

Comment-09

RH: The naming convention on D-linkbases is incomplete.

Comment-10

RH: The naming convention on P and C-linkbases forces children to come from the same schema or split linkbases per children origin. Is that the intention or is there a better algorhytm for the naming convention?

Comment-11

Comment-12

RH: Is there a desired limit to the length of the ns-prefix?

Comment-13

RH: Referring to custom codes through a linkbase mechanism is fine, but using a label linkbase is not. These are not labels as intended by XBRL 2.1 (for human consumption). Why not using a proprietary linkbase with a dedicated arcrole expressing the exact meaning of the relationship?

Comment-14

RH: Why are lower and upper case on domains needed? Rather keep URI's in lowercase to prevent any problems.
RH: Is it wise to have all the member schemas named 'mem.xsd'?
RH: What is the added value of another layer in the location that expresses the same thing as the name as the schema? (met/met.xsd, why not just met and add the extension .xsd for the location)
RH: What is the ns-prefix for the hier.xsd?
RH: Is it wise to have all the member hierarchy schemas named 'hier.xsd'?

Comment-15

RH: The assigned letter 'p' is used for both percentages and decimals. Both are also supported by the measure 'xbrli:pure'. How the differentiate between the two? I suggest using a) different letters and b) using UN-Cefact measures for percent: un66411:P1 which frees up xbrli:pure for decimal or integer.

Comment-16

RH: Can we put 'any xml datatype' on the Typed domain?
RH: There will be no measure on a typed domain member, that can be left out?
RH: The source document describes an extra attribute (@model:domain) to be included to identify the Qname of the domain. Why is this needed, and where is this extra attribute placed?

Comment-17

RH: Why make the URI on ELR so complicated? Simply use URN's. E.g. urn:eba:perspectives:1. In namespaces the URI is really an identifier, identifying the elements in the schema, but in linkroles the URI is just a name, there is no pointer from outside XML that can point to it (like with an URL which is mostly derived from the URI by placing the file name there or just the file extension added).

Comment-18

RH: Dimensions and domains are both made up by two letter capitals. Are they allowed to overlap? MUST domain CA be in dimension CA?

Comment-19

RH: A tableGroup hierarchy is created in a generic linkbase with a custom arcrole. Yet the linkbase is named tab-pre.xml. This name is hinting at a parent-child presentation linkbase. Better use something else.
RH: The custom arcrole for tableGroup - table has @cyclesAllowed=undirected. I think that should be 'none'. There should be no cycles in any form between tables and tablegroups because tablegroups can already be parent and child.