Skip to main content Skip to navigation

More about DOSTE and EDEN

Most of the resources that are available to assist in model-building with DOSTE have already been presented in the labs. There is self-evident potential for exploiting DOSTE in relation to improving the modelling of processes, providing richer interfaces for interaction and visualisation, and introducing hierarchical structures. You are encouraged to make use of DOSTE in these roles, and to link its use with the principles of EM as presented in the lectures. Work that contributes to future exploitation of DOSTE (e.g. involving the development of library components to support forms of visualisation and interaction that are not as effectively handled in EDEN), and that highlights future directions of development for DOSTE will be particularly welcome. In some instances, as when overcoming the problems of modelling with "hiatons", or giving fuller support for explicit agency, there are as yet unresolved research questions, and allowance will be made for this when evaluating attempts to tackle these issues.

Because of the attention given to DOSTE in this year's presentation of CS405, the coverage of EDEN has been somewhat less thorough and detailed. To help to redress this, some pointers to supplementary practical resources are reviewed here. Additional advice about the variants and extensions of EDEN that you may wish to use is also included. (Note that many of the links identified here point to practical resources presented in previous years.) In trying out older EDEN models, you may well find it helpful to use early variants of the EDEN interpreter, about which more is said below. Redeveloping some of these older models  on the new tkeden-2.10 platform so that they can exploit the additional power that DOSTE affords is a promising basis for submissions to WEB_EM-06.

  1. For basic knowledge of the eden, donald and scout components of EDEN, consult Labs 1, 2 and 3 from 2007-8. This is an essential prerequisite for basic model-building. (See Seminar 3 in 2007-8 for some answers to Lab 2 together with a supplementary DMT diagram, and for more advanced features of EDEN - some of which are likely to be relevant to your model-building exercise.)

  2. You may well find it useful to exploit additional notations, such as the eddi notation for relational views and tables. An introduction to eddi is provided in Lab 4 from 2007-8. A brief account of how eddi was once applied in teaching a second-year database module exists in the form of an EDEN model prepared for a conference presentation (which you should run on tkeden-1.49 on Linux or tkeden-1.46 on Windows).

  3. It is a good idea to become familiar with some of the models demonstrated in the lectures. There are several links in this year's lecture notes that enable you to do this. If you have difficulty in running these models, please post queries to the CS405 forum.

  4. There are many extensions to EDEN that may be of interest. The CS405 labs from 2006-7 are a good source for studying some of these. You are not required to use any of these extensions in your submitted model, but it may well be that studying / critiquing / deploying one or more of these extensions will be helpful in relation to your chosen theme. They include:

    • The agent-oriented parser (AOP): This is a notation that makes it possible to implement new definitive notations within EDEN itself (and indeed to modify them on the fly). The eddi notation is one example of a notation implemented in this way. (For more details, see Lab 6 from 2006-7.)
    • The dependency modelling tool (DMT): This exists in the form of a Java-based tool for visualising dependencies, and in a version that is implemented in EDEN itself (see Lab 5 from 2006-7).
    • The GEL presentation environment: This environment can be very useful in presenting a model, especially in relation to documenting, recording and replaying interactions. This can be effective even in relation to quite simple models.
    • The Sasami 3D modelling notation: This notation makes it possible to define simple 3D models with dependencies. It is introduced in sasamiexamplesCarter1999, and was improved in one of the contributions to WEB-EM-03 (sasamiprimitivesKnight2007). Its use is illustrated in Charlie Care's planimeter models (kaleidoscopeBeynon2005).
    • Other extensions: Other useful extensions to the EDEN interface include Karl King's Visual Symbol Table (vstKing2005) and variants of his tool for drawing Scout windows with the mouse alone drawScoutKing2004.

As a final word of advice about using EDEN, please bear in mind that - in addition to the radical innovation in tkeden-2.10 - there have been some quite significant changes in the interpreter over recent years. This unfortunately means that there are some minor inconsistencies between different versions of tkeden. The primary things to look out for are: the introduction of clocks (older models may run best with tkeden-1.46); two runset regimes that affect the way in which mouse tracking is handled (between tkeden-1.66 and tkeden-1.67) and may have other side-effects; the revised implementation of eddi that was introduced in tkeden-1.67 - this provides simple eden equivalents for relational operators, but loses some functionality in the process. These inconsistencies are to be regretted, but are hard to avoid as there are currently so few people actively involved in maintaining the EDEN code. Even though it may be difficult to address problems in the EDEN implementations, it is most important that we are made aware of them, and we would appreciate notification via the Forum.