Minutes of FIGGY BAR RT Conference.

Date: 06/07/90   Time: 22:49EDT
  
 The question of the night : When an implementation of ANS Forth becomes
                             available for your system will you use it, roll
                             your own, don't know, don't care ?.....
                             This lead to several opposing views of the TC's
                             actions, the BASIS as it now exists, and view(s)
                             of the future standard. The discussion between
                             Robert Berkey and Dennis Ruffer over decisions
                             made regarding division was especially involved.
                             IF YOU HAVE ANY interest in the future of Forth,
                             this FIGGY is another 'must read'.
        
                             Berkey contends every member of the TC should be
                             'required' to write a "serious" program using the
                             defined standard !
 
                             Does the committee understand F-83 ? Is it just
                             a jumping off point, or even being considered ?

                             Will you be able to write 'standard' code ? Will
                             you even want to ?

Attendees:
[[gars] GARY-S]
[[Wil] W.BADEN1]
[[Dennis] D.RUFFER]
[[Robert] R.BERKEY]
[[jaX3J14] JAX]
[[John] JDHALL]

Minutes:
<[Wil] W.BADEN1> is here.
<NMORGENSTERN> is here.
<D.RUFFER> is here.
<[Wil] W.BADEN1> Let me retrospect.
<[Wil] W.BADEN1> I'll use it.
<[gars] GARY-S> the question of the nite - answer it
<[Wil] W.BADEN1> I did.
<[Wil] W.BADEN1> I said I'd implement and use it.
<[Wil] W.BADEN1> I do already.
<[Len] NMORGENSTERN> Oh, Yes I will use the new standard.
<[Dennis] D.RUFFER> sure, I'll implement it as soon as Forth, Inc lets me :-)
<[gars] GARY-S> What are you running on wil ? still Apple ][ ?
<[Wil] W.BADEN1> Mac.
<[Dennis] D.RUFFER> whose Forth are you using on the Mac wil?
<[Wil] W.BADEN1> Vacilating between PocketForth and MacForth.
<[gars] GARY-S> How soon do any of you project major vendors offering a 
standard kernel
<[Dennis] D.RUFFER> I would suspect that Forth, Inc. will have one ready before 
the standard is actually approved
<[Dennis] D.RUFFER> right now, that will not be until March '92 at the earliest
<[Wil] W.BADEN1> I use my 640K dual-floppy PC clone for Zen 1.2
<[gars] GARY-S> Is the BASIS model still referred to as goofy ?
<[Len] NMORGENSTERN> Who refers to it as goofy?
<[Dennis] D.RUFFER> BASIS 12 is actually going to be pretty good IMHO
<[Wil] W.BADEN1> Last week's Easter algorithm was checked out on MacForth and 
Zen 1.2.
<[gars] GARY-S> Once upon a time the BASIS was referred to by that (or some 
similar) name to discourage premature modeling
<R.BERKEY> is here.
<[Dennis] D.RUFFER> from what I could infer at the last meeting, even the TC 
thinks that it is pretty solid now
<[gars] GARY-S> Bob - will or won't you ?
<[Dennis] D.RUFFER> hi bob, how are you feeling about basis now?
<[Robert] R.BERKEY> Writing standard code is highly impractical.
<[gars] GARY-S> Is this a critique
<[Robert] R.BERKEY> The core has been watered, burden placed on application 
programmers.
<[Dennis] D.RUFFER> true, but the result is consensus
<[Wil] W.BADEN1> The core has been watered?
<[Robert] R.BERKEY> The support of sign-magnitude machines is an empty idea.
<[Wil] W.BADEN1> OK.  We agree.
<[Robert] R.BERKEY> "concensus" means what we have been told to think, bah.
<[Dennis] D.RUFFER> can you program in Forth without thinking now?
<[gars] GARY-S> It is NOT too late to change sign-magnitude support !
<[Wil] W.BADEN1> The word is "consensus" -- what we consent to, not result of a 
census.
<[Robert] R.BERKEY> The extensions, as Bradley notes, are important.
<[Wil] W.BADEN1> For what?
<[Robert] R.BERKEY> It's the core that is poisonous.
<[Robert] R.BERKEY> Committee still doesn't understand Forth-83.
<[Robert] R.BERKEY> There are Standard Programs and Standard Systems.
<[Robert] R.BERKEY> A balance exists between the two.
<[Robert] R.BERKEY> Adding from one takes from the other.
<[Robert] R.BERKEY> I'd like to see every committee member write a serious 
program
<[Robert] R.BERKEY> as required by the standard.
<[gars] GARY-S> That's a VERY interesting concept Bob ! I wonder if any will be 
willing to rise to such a laudible challenge ?!?
<[Robert] R.BERKEY> A Forth-83 + imposes a 20x performance penalty.  WHY?
<[Robert] R.BERKEY> It might be possible to integrate one's and two's 
complement.
<[Robert] R.BERKEY> But the understanding of one' complement is such that
<[Robert] R.BERKEY> the subcommittee chair told me that one's complement is 
_not_ a number
<[Robert] R.BERKEY> circle.
<[Dennis] D.RUFFER> how do you mean "+ imposes a 20x performance penalty"?
<[Robert] R.BERKEY>  + has been reduced from wrap-around numbers to "n"s.
<[Robert] R.BERKEY> Unsigned addition has been removed, in particular.
<[Robert] R.BERKEY> Since every array definition in existence requires unsigned 
addition,
<[Robert] R.BERKEY> every array definition ever written is broken.
<[Dennis] D.RUFFER> how many arrays do you use that are longer than 32K?
<[Robert] R.BERKEY> How many government contracts will call for Standard 
Systems?
<[Robert] R.BERKEY> Dennis, that's not the point.
<[Robert] R.BERKEY> You don't know where your array exists in memory.
<[Dennis] D.RUFFER> yes it is! the fact that arrays manipulate memory addresses 
is where the problem lies, not with +
<[Wil] W.BADEN1> What do you mean "unsigned addition"? -- addition of unsigned 
numbers?
<[Robert] R.BERKEY> It's the same problem as using addresses in a Forth-79 LOOP