EuroForth 2005 announced talks

Angel Robert Lynas, Bill Stoddart (University of Teesside)

|SuDoku Solver Case Study: from specification to RVM-Forth (part I)
|
|A project is underway to formulate a development cycle from B --- 
|suitably augmenting its implementation language B0 with reversibility 
|constructs --- to a coded implementation in the reversible target 
|language RVM-Forth with translation schemas defined for this final 
|stage. This paper describes the first phase of a case study using the 
|puzzle SuDoku to investigate possible ways of fleshing out such a 
|development cycle. We adopt an experimental approach, using a 
|relatively simple specification as a springboard for what a generated 
|code implementation might look like, and explore correspondences 
|between the specification and implementation. 


M. Anton Ertl, Bernd Paysan

|Xchars: Write programs for UTF-8 *and* fixed-width character encodings
|
|A program written to use the xchar word set can work on a system with
|fixed-width character encodings as well as on a system with certain
|variable-width character sets (in particular, UTF-8).  We have
|implemented this word set in Gforth with support for classical 8-bit
|fixed-width encodings and for UTF-8 (selectable on system startup).
|In this talk we demonstrate the word set, and discuss the current
|state of the word set and its implementation.


N.J. Nelson, K.B. Swiatlowski

|Self Documenting Sequences
|
|When creating automation code for mechanical handling equipment which is 
|specially adapted for installation at a wide variety of different sites, 
|it is common for numerous alterations in the code to be required on 
|site, at the last minute. Under pressure, user documentation starts to 
|diverge from code. How nice it would be, if clear, accurate and readable 
|documentation could be regenerated automatically each time the code was 
|recompiled - and translated, also automatically, into the customer's 
|language!


N.J. Nelson, C. Williams

|First experiences with Microcore
|
|Following the convincing demonstrations at EuroForth 2004, we decided to 
|use the "Microcore" VHDL Forth processor in the design of three new 
|products. This paper will describe our progress in expanding the core 
|design with additional peripherals, performing simulation, board 
|implementation, and early experiments in writing code on the Microcore.


Federico de Ceballos

|                    Simplicity in Forth
|
|In his book "Simplicity", celebrated author Edward de Bono (famous for
|concepts such as "Lateral Thinking" or "Po") put forward ten rules
|that should be used in every system that tries to define itself as
|simple. This paper studies how the Forth language meets these rules.


Juergen Pfitzenmaier:

|              Type checking FORTH -- Part I
|
|The ANS standard for FORTH states in A.3.1.3.3 the one-to-one re-
|lationship between addresses and unsigned numbers. This relation-
|ship  can  be applied in two different ways: Using this relation-
|ship a lot (this is backed by the reading of  the  current  stan-
|dard)  or  making  only sparse use of this relationship.  We show
|that this relationship can make type checking  FORTH  unsound.  A
|few conservative changes to the standard would suppress the rela-
|tionship in most cases and enable a sound type  checking  without
|the need to change an existing FORTH implementation.  We give ex-
|amples showing the problems in the current standard and the  pro-
|posed solutions.
|
|
|
|              Type checking FORTH -- Part II
|
|The  stack  notations  in  the current ANS standard for FORTH are
|neither sufficient to describe the role of execution  tokens  nor
|sufficient  to  describe  the semantic actions of deferred execu-
|tion, or of the words IF and DOES> when it comes to  type  check-
|ing.  A detailed example shows the necessary conservative changes
|to the standard and how IF can be type checked.
|
|
|              Type checking FORTH -- Part III
|
|Full type checking of a FORTH program needs (at least)  one  non-
|conservative  deviation from the ANS standard: the implementation
|dependent size on stack of the data types colon-sys, do-sys, ...,
|nest-sys must be looked up by an environmental query.


M. Anton Ertl, David Gregg:

|Stack Caching in Forth
|
|Stack caching speeds Forth up by keeping stack items in registers,
|reducing the memory accesses for stack items.  This paper describes
|our work on extending Gforth's stack caching implementation to support
|more than one register in the canonical state, and presents timing
|results for the resulting Forth system.  For single-representation
|stack caches, keeping just one stack item in registers is usually
|best, and provides speedups up to a factor of 2.84 over the
|straight-forward stack representation.  For stack caches with multiple
|stack representations, using the one-register representation as
|canonical representation is usually optimal, resulting in an overall
|speedup of up to a factor of 3.80 (and up to a factor of 1.53 over
|single-representation stack caching).