The need for updating the ANS Forth standard was discussed. It was felt that the current Forth '94 standard was in need of amendment, but that the current ANS procedure where too formal and, given modern communication technologies, rather out dated.
The conference felt that a less formal mechanism for the raising and discussion of proposals should be introduced. To this end a number of resolutions where adopted by the conference:
- Request for Discussion / Call for Votes system [13/0/2]
- The RFD/CFV system introduced by Anton Ertl on the comp.lang.forth news group should be used as a method of proposing new changes to the standard. The discussion group can also be used to debate the proposal. This process is to be encouraged and managed by Anton Ertl.
- Separate Mail List [14/0/1]
- It was felt that a separate Mail list should be set up for those who either do not have access to, or wish to use the comp.long.forth news group. The distribution list for this email list should not be accessible. All messages sent to the list should be archived and available fore review. It was felt that restrictions should be placed on email addresses in order to prevent unsolicited emails (SPAM), rather than allow for anonymous messages.
Note: An email list has been set up on Yahoo Groups. To join people must send a request to:email@example.com visit the web site:http://groups.yahoo.com/group/forth200x/
- Standards Meeting [9/0/7]
- There is to be a one day meeting prepended to the next euroForth conference to discuss the various proposals. To this end the cut of date for new proposals will be three months prior to the next conference. This is in order to allow the proposals to be integrated into a draft standard document.
A number of project are now taking a base processor design and adapting it for use with the specific hardware requirements. The microcore and b16 projects presented at this conference being two example. Would it be possilbe to develop a minimul hardware core that such hardware based forth systems can adhear to? It is possible that this may lead to a new strand to the Forth-2005 project, a hard Forth standard.
A email list has been set up to allow further discussion on this topic, to join you must send a request to:coreForthfirstname.lastname@example.org visit the web site:http://groups.yahoo.com/group/coreForth-subscribe/
Object-oriented development is a topic which rears its head every few years. There are any number of object orientated packages available for forth, unfortunately they do not agree as to a standard notation all object model.
A number of criteria for any standard object model were discussed:
It was felt that an object model cannot be addressed until such time as a formal data structure model has been developed. Once this exists it was felt that a standard object model could be built in ANS Forth.
- A dot-parser*
- Separate declaration to implementation
- Ability to extend a class after it has been declared
- Both late and early binding
- Visibility: public / private / friends
- Inheritance: multiple / single / interface
* A dot-parser has been implemented in several systems. When a word can not be found in the dictionary, it word is split on the dot, with the left hand side indicating the scope of the word, i.e., the wordlist, and the right hand side indicating the word with the scope.
The review process for the scientific library appears to have come to a stop. It is felt that the harness provided for the scientific library is in need of review. Stephen Pelc has agreed to submit a revised harnessed, based on his DOCGEN system, for discussion under the new Forth200x procedures with the intent that this can be used as the basis for further scientific library submissions and review.