Forth standards meeting at EuroForth 2008

On 2008-09-25 and 2008-09-26 there was a Forth standards meeting, right before EuroForth 2008. The participants were:

Willem Botha, CCS, South Africa
Federico de Ceballos, University of Cantabria, Spain
M. Anton Ertl, TU Wien, Austria
Ulrich Hoffmann, Heidelberger Druckmaschinen, Germany (observer, late)
Peter Knaggs, University of Bournemouth, England
Bernd Paysan, Germany
Stephen Pelc, MPE, England
Bill Stoddart, University of Teesside, England
Andrew Haley, Red Hat, England
Leon Wagner, Forth Inc., USA
Willi Stricker

At this meeting the observers were allowed to vote. Anton Ertl chaired most of the meeting.

The following topics were discussed (not always in this order):

Minutes of last meeting and standards document

The minutes of the last meeting contain a mistake: They claim that the throw-iors proposal was referred for further clarification, but actually it was accepted (this is also reflected in the standard document). It was decided (10:0:1) to accept the minutes as is, and write a correction in the next minutes. However, later Peter Knaggs decided to actually make the correction in the 2007 minutes (which were a draft until then anyway).

The standards document was reviewed and accepted as reflecting the results of the last meeting (8Y:0N:3A).

Minutes, draft documents, and reports were published on the www.forth200x.org web site. Anton Ertl promised to write a 2007 and a 2008 report for the Vierte Dimension (German FIG newsletter). Peter Knaggs has not contacted Paul Frenger about publishing the minutes in SIGPLAN Notices yet.

Accepting proposals into the standard document

The following proposals which had gone to a CfV were discussed and voted on:

Separate FP stack: The proposed changes to the standard document were discussed, and additional changes were asked for (additional documentation requirements, and specifying the combined stack as obsolescent). These were added during the meeting, and the result was accepted 10Y:0N:0A.

Removing obsolete words

Should words that were called obsolescent in Forth-94 now be removed as obsolete? There was consensus that they should, the question was how:

Remove them completely
Keep a list of names in an appendix
Keep names and specification in an appendix
Keep names and specification in an obsolete wordset

After some discussion, it was decided (11Y:0N:0A) to keep the names and the specification in an appendix.

RfDs

The following proposals (currently in RfD phase) were discussed:

2VALUE and FVALUE

No progress has been made there because the original RfD submitters have not submitted new RfDs adapted to the changed TO wording. The original RfD submitters should be contacted (by Anton Ertl) and be asked to submit a revised RfD (and eventually a CfV).

Escaped strings

This RfD was also seen as being mostly ready for a CfV (although some participants did not remember if there were some details that should be fixed first).

Synonyms

The discussions centered mostly around some of the ambiguous conditions in the proposal (in particular, that synonyms would not be tickable and POSTPONEable). Several participants had a strong feeling that these ambiguous conditions should be removed before the RfD becomes a CfV. Otherwise, the proposal was found acceptable.

Extended characters

This proposal was now deemed to be mature enough to go to a CfV (probably after a final RfD). A question came up whether it should be in a new wordset or in an existing one.

Enhanced locals

Leon Wagner reported that in the code of Forth, Inc. customers '{' is actually used in compile state several hundred times, so Forth, Inc. would not implement a syntax based on '{'. Also, BigForth uses

{
... |

in a different way than proposed. There was some discussion on how to resolve these conflicts, resulting in the following syntax: {: ... | ... -- ... :}. Another RfD will come out with this change.

Another point in the discussion was whether to allow several locals definitions. Bernd Paysan intends to write an RfD for that.

Local buffers

There was quite a bit of discussion on that issue. After the Forth200x meeting Leon Wagner and Stephen Pelc worked out a completely new syntax which would allow writing code like this:

[: value a 5 value b 10 cells buffer: c :]

Directories

It was pointed out that someone uses F" for some other purpose, so a different name would have to be used. Several other points that were already made in the email discussion were repeated. In an off-line discussion Stephen Pelc outlined his preferred solution of first standardizing macros and then building directory-path words on top of that.

Big proposals

The internationalization/localization and cross-compilation proposals (which are not yet RfDs) were discussed (but not very deeply, as most or all participants were not very familiar with them, and they are big), and it was suggested that they should be developed into RfDs.

For the cross-compilation proposal some revision of the existing proposal was promised (apparently there were some advances since that proposal was originally written); also Stephen Pelc, Leon Wagner, and Bernd Paysan should sit down together and work out the RfD.

Future RfDs

Federico de Ceballos promised that he would submit RfDs on the issues that he has worked on.

Next meeting

There will be another meeting before the next EuroForth. There was some discussion whether it should be two days, and whether there should be another meeting before the Forth-Tagung.

Anton Ertl