CaveXML Discussion Summary

Martin Laverty, 2002-10-01

This page lists the topics discussed to Sept 2002 on the CaveXML discussion list. The summaries usually only give an indication of the status and key arguments deployed: you should read the original postings to appreciate the arguments, counterarguments, protagonists and nuances.


Topics covered:

  1. Comments
  2. Cross sections
  3. Dates
  4. DTDs and Schemas
  5. Entity model for cave surveying
  6. Forerunners of CaveXML
  7. Internationalisation
  8. Modularity
  9. Original book
  10. PCDATA vs CDATA
  11. Prototype definitions
  12. Referencing
  13. Requirements of Software using CaveXML
  14. Scoping
  15. Survey methodology

Threads to refer to for topics discussed:

Cross sections

Summary: Although not required for a basic line survey, many contributors thought the inclusion of cross section data very important. Most are satisfied with Left, Right, Up, Down measurements (with discussion on whether these should be applicable only at survey stations or allowed more generally with addition distance and orientation parameters) but some are more ambitious.
See Thread: LRUD: 'From' or 'To'?
Date: january_01
See Thread: Cross Section libraries
Date: january_01
See Thread: cross sections
Date: january_01

Referencing

Summary: Considerable attention was devoted to use of the ID type for referencing eg a station name in leg readings and in results. An ID must be unique within an XML file. XMLSchema gives a wider concept ...
See Thread: ID specification?
Date: july_01

Comments

Summary: The concensus is that CaveXML should allow comments at all levels. There was discussion as to whether to call it text or comment
See Thread: Comments
Date: january_01

Original book

Summary: It is thought by many to be important that CaveXML be capable of retaining images of the original survey sheets rather than just transcribed data. It has also been suggested that the order of data recording, even at the column level, be reflected in the XML but few good reasons have been advanced to do this.
See Thread: Images in the data file
Date: january_01
See Thread: CDFO: Raw data and XML
Date: january_01
See Thread: Data format/visualization?
Date: september_02

DTDs and Schemas

Summary: Much discussion has been couched in terms of DTDs, but the use of different types of schema languages has been mentioned. It remains to be decided how prescriptive the definitions will be with the aim of enforcing best practice as opposed to a permissive approach reflecting the diversity of approaches which have been used.
See Thread: Validating proposed CaveXML DTD's
Date: march_01
See Thread: Relax, TRex?
Date: march_01
See Thread: Data format/visualization?
Date: september_02

Prototype definitions

Summary: Much discussion has been couched in terms of DTDs, but the use of different types of schema languages has been mentioned.
See Thread: First cut
Date: march_01
See Thread: Latest DTD attempt
Date: march_01
See Thread: Second cut at DTD
Date: march_01
See Thread: Latent DTD attempt
Date: june_02
See Thread: Measurement Element
Date: september_02

Scoping

Summary: The ability to allow a high level definition of values applying too nested sub-elements will be useful, but this should be capable opf being overridden lower down. This will apply to things such as corrections and units. There was discussion as to whether specific groups would be needed or desirable.
See Thread: defaulting units for measurement
Date: february_02
See Thread: Groupings
Date: february_02
See Thread: Finding groupings
Date: february_02
See Thread: Nested Groupings
Date: february_02
See Thread: Nested Surveys
Date: september_01

Modularity

Summary: It is hoped that it will be possible to use the Xinclude and Xlink standards to enable reuse of existing data
See Thread: Linking CaveXML files...
Date: july_01

Entity model for cave surveying

Summary: An entity relationship model is a well-established way of analysing data requirements for database systems. As it is likely that cave survey data could be stored in such a way it was suggested that this would be a good way of analysing requirements for CaveXML. It does not mean that CaveXML will be linked with any particular database type - XML is increasingly seen as the common data interchange type between different systems.
See Thread: Survey entities and fields
Date: october_01
See Thread: About data model
Date: january_02
See Thread: Cave/CaveSystem/Project/Survey element/entity
Date: june_02
See Thread: Data model : branches and nodes... and more
Date: june_02
See Thread: Entities and fields
Date: june_02
See Thread: Types of data sets
Date: june_02
See Thread: ERD based on?
Date: june_02
See Thread: Comments on the data model
Date: june_02

Internationalisation

Summary: Although discussion has all been in English it is recognised that CaveXML should be designed explicitly with international use in mind. XML goes a long way to providing this capability, but terminology is a point which differs even between Anglophones eg UK leg, bearing USA/Australia shot, azimuth
See Thread: Internationalization
Date: april_02
See Thread: Internationalization
Date: may_02

Forerunners of CaveXML

Summary: A request was made for information about the latest version of the specificartion for Hierarchical Tagged Objects, a proposal made in 1995 and implemented as an output format by Survex until 2001, when it was dropped due to lack of use. No source was found. A thought provoking discussion of previous cave data standardisation efforts was contributed. The Land Survey specification was pointed out as a kindred effort.
See Thread: HTO
Date: january_01
See Thread: Past Words
Date: january_01
See Thread: Land Survey XML
Date: april_02

PCDATA vs CDATA

Summary: Rather esoteric, but essentially deciding that it could be useful to allow for storing binary data such as images directly rather than through a link. Some form of binary to text encoding would be needed.

Survey methodology

Summary: It is necessary to know something of how surveys are actually done to guide identification of terminology, entities, and relationships for incorporation in CaveXML.
See Thread: DataFileVersion
Date: february_02
See Thread: Stations are primary
Date: february_02
Date: march_01
See Thread: Spanning trees (again)
Date: february_02
See Thread: Unique Station names
Date: february_02
See Thread: Station names
Date: february_02
See Thread: What is a Station?
Date: february_02

Dates

Summary: It was concluded that the ISO standard for dates should be adopted ie: ccyy-mm-dd
See Thread: Dates: explicit fields and missing values
Date: february_02

Requirements of Software using CaveXML

Summary: Some people have wondered whether CaveXML should be designed with any particular software in mind. Should particular constructs be used because they might be easier to implement, or are already implemented in a particular way? Some authors of existing software have suggested that they will not use CaveXML unless their specific requests are heeded. The concensus is that CaveXML should not be biased by such considerations but should seek to encourage best practice.
See Thread: Required ability of software?
Date: february_02