Minutes of FIGGY BAR RT Conference.

Date: 04/04/91   Time: 22:36EST

Attendees:
[[LEN] NMORGENSTERN]
[GARY-S]
[[Wil] W.BADEN1]
[[doug] D.GARRETT]
[GILES]
[[Dave] OLORIN]

 Items discussed: Library archiving, conditionals syntax,OOF in limbo, FORTH
                 strengths, portability, extensibility, embedded applications,
                 more sysntax...
Minutes:
<[LEN] NMORGENSTERN> is here.
<[Wil] W.BADEN1> is here.
<[LEN] NMORGENSTERN> A question Gary. Dennis says ARC is preferable to ZIP,
<[Wil] W.BADEN1> hi gary.
<[LEN] NMORGENSTERN> Which do you use?
<GARY-S> Len - A wider mix of systems have .arc ported to them
<GARY-S> I use .arc, .zip. .lzh, on all and .z and .Z on Unix
<[LEN] NMORGENSTERN> Ah, thanx
<GARY-S> Wil - have you had an opportunity to examine the latest bruhou in 
standards over conditionals ? comment ?
<[Wil] W.BADEN1> I don't think that it is appropriate for standardization.
<[Wil] W.BADEN1> My preferences are {IF and }
<GARY-S> The issue now, seems to be 'what' punctuations are used to demark IF 
THEN ELSE
<D.GARRETT> is here.
<GARY-S> I'm not at all pleased with reversed parens
<[Wil] W.BADEN1> In the interpretable mode.  There's no question about compiled 
version.
<GARY-S> lisp users love it, naturally
<[LEN] NMORGENSTERN> Wil, Gary can you explain the questions?
<[doug] D.GARRETT> hello, all.
<[Wil] W.BADEN1> Doug, you told me 600 pages per minute.  That's incredible.
<GARY-S> what questions ?
<[LEN] NMORGENSTERN> About conditionals
<[Wil] W.BADEN1> About interpreted conditionals.
<[doug] D.GARRETT> You strike me as  somewhat curious of other languagues, and..
<GARY-S> What to use to demark them
<[doug] D.GARRETT> oops, sorry
<[Wil] W.BADEN1> "Should Forth have conditional compilation and if so, what 
should the syntax be?"
<GARY-S> Don't apologize, Doug ---- during open FIGGY's it is comment when 
ready !
<[LEN] NMORGENSTERN> Ah. My opinion is very opinionated. 1. It should not be
<[LEN] NMORGENSTERN> standard. 2. If it is, then it must work inside colon 
definitions.
<GARY-S> Thank you Mr. Baden - succinct and correct
<[Wil] W.BADEN1> : { [ASCII] } WORD DROP ; IMMEDIATE
<[Wil] W.BADEN1> : {IF 0= IF [COMPILE] { THEN ; IMMEDIATE
<[Wil] W.BADEN1> : } ; IMMEDIATE
<[Wil] W.BADEN1> That does it all.
<[LEN] NMORGENSTERN> Looks good to me
<GARY-S> You have read Chuck Eaker , Bill Bouma and Mitch Bradley comments ?
<GILES> is here.
<[Wil] W.BADEN1> Briefly.  I think they'r over-engineered.
<GARY-S> agreed !
<GARY-S> Mitch wants C flavouring - that is apparent
<GARY-S> Unix, actually
<[LEN] NMORGENSTERN> I will have to download the thread.
<GARY-S> Cat 10, Topic 4 Len
<[Wil] W.BADEN1> I don't think that conditional interpretation is important for 
applications.
<GARY-S> I rather like your no non-sense syntax, Wil.   nice
<[Wil] W.BADEN1> A complete implementation -- "else" is { }
<[Wil] W.BADEN1> This is also the classical solution.
<[LEN] NMORGENSTERN> It's a matter of degree, Wil. In general, I agree, but
<[LEN] NMORGENSTERN> it is often useful to have stuff for testing that you can
<[LEN] NMORGENSTERN> turn on and off by setting a variable.
<[Wil] W.BADEN1> : jkjkjkjkj   jkjkjj [ aVariable @ ] {IF ...... } sdfdfda ;
<[Wil] W.BADEN1> ack?
<[doug] D.GARRETT> Giles, WHY?!
<[LEN] NMORGENSTERN> Exactly, Wil. And if you want, you can set up a single word
<[LEN] NMORGENSTERN> to do just that, but I can't see it as a standards issue.
<[Wil] W.BADEN1> To whom, Doug.
<[Wil] W.BADEN1> : Test1 ; immediate
<[doug] D.GARRETT> Excuse me, gentlemen.  Not that I'm trying to
<[Wil] W.BADEN1> : Test1 ; [COMPILE] { ; IMMEDIATE
<[doug] D.GARRETT> be anti-social.  I have a few messages...
<[doug] D.GARRETT> that have nothing in common with your discussion about
<GARY-S> So - make them
<[doug] D.GARRETT> whethere a useful programming practice will or will not
<[doug] D.GARRETT> be included in some sort of STANDARD.
<[doug] D.GARRETT> As a practical programmer, making a partime wage...
<[doug] D.GARRETT> I use whatever is well-documented in one place or another...
<[doug] D.GARRETT> and provide THAT 'Standard' to my boss/customer...
<[doug] D.GARRETT> I understand Your interests, but they are not mine, and I 
didn't want
<[doug] D.GARRETT> to interrupt what you were discussing.  Sorry for the 
inconvenience.
<GARY-S> Different opinions and a mix of subjects is exactly what the open 
FIGGY is for
<[doug] D.GARRETT> wow,  hello hello ?? anyone here?
<[LEN] NMORGENSTERN> The problem is not that simple, Doug. If you want your
<[LEN] NMORGENSTERN> program to have the maximum portablity, then you will have
<[LEN] NMORGENSTERN> to write it as a standard program. (If you don't care, then
<[LEN] NMORGENSTERN> of course it doesn't matter.)
<[doug] D.GARRETT> `ah, thank you Gary S.  Ok, I'm out of the closet, no more
<[doug] D.GARRETT> fumbling with trying to /SEN number
<[doug] D.GARRETT> I gather there are two or more opposing points of view...
<[doug] D.GARRETT> as to HOW TO build something useful into the upcoming 
Standard, yes?
<[Wil] W.BADEN1> At least, Doug.
<[Wil] W.BADEN1> Gary -- what's the CAT and TOPIC numbers -- I'll download the 
whole thing in one burst.
<[LEN] NMORGENSTERN> Cat 10 Topic 4
<GARY-S> Cat 10, Top 4 for the past 2 weeks
<[Wil] W.BADEN1> Doug, I think we'r waiting for you.
<[Wil] W.BADEN1> Thanks Len & Gary.
<[doug] D.GARRETT> Hm...  Well, I hesitate to compare
<[doug] D.GARRETT> Forth to MUMPS, because I would never want to sell
<[doug] D.GARRETT> MUMPS to you all.   But in...
<[doug] D.GARRETT> the business I program for, an industry-wide 'standard' for
<[doug] D.GARRETT> extending the language is pretty-well defined and structured.
<[doug] D.GARRETT> so one might look upon the ANS-MUMPS commands, functions, 
global-handling
<[doug] D.GARRETT> facilities as a 'KERNAL', and then...
<[doug] D.GARRETT> the people who write the extensions to meet the design...
<[doug] D.GARRETT> goals try to keep their shop tidy and in order, much
<[doug] D.GARRETT> like a well organized machine shop.  Excuse the analogy, if 
you
<[doug] D.GARRETT> think it is inapporpriate.  Anyway, from...
<[doug] D.GARRETT> MY point of view (I am NOT an engineer, nor did I conceive...
<OLORIN> is here.
<[doug] D.GARRETT> the software system that I extend using MUMPS...
<[doug] D.GARRETT> I look upon however you want to do CASEs or LOOPs...
<[doug] D.GARRETT> as pretty much  hmmm un-attached to my duties in the world.
<[doug] D.GARRETT> When y'all figure out what you want to call standard, I'll 
take
<[doug] D.GARRETT> dutiful note.  But it won't change what
<[doug] D.GARRETT> I'll be doing tomorow or the next month or year.  Sorry.
<[doug] D.GARRETT> please go ahead, tholuhl.  I like to watch you all.  Thanks.
<[Wil] W.BADEN1> Doug, your viewpoint is the proper one.
<GARY-S> Dave - had a inquirey about your OOF - is that + TOTALLY + dead ?
<[Wil] W.BADEN1> I feel that the attraction of Forth is two-fold. ...
<[Dave] OLORIN> Well, uhmmmmmmm.
<[Dave] OLORIN> It's close.
<[Dave] OLORIN> (To being dead)
<GARY-S> Then RIP
<[Dave] OLORIN> I haven't looked at it in over a year.
<[Wil] W.BADEN1> First, it's THE best language for embedded systems, and other 
dedicated applications with limited resources.\
<[doug] D.GARRETT> I heartily agree, Wil.
<GARY-S> As you (wil) once pointed out, it is still tyhe champion "pathfinder"
<[Wil] W.BADEN1> Secondly, it's SO simple that it can be COMPLETELY understood 
after about 2 months study.
<[Wil] W.BADEN1> This lets you do YOUR thing YOUR way.
<[doug] D.GARRETT> Good.  I'm maybe halfway, then.
<[Wil] W.BADEN1> As I also once said, "With Forth systems, when you'v seen one, 
you'v seen one."
<[doug] D.GARRETT> (chuckle) I agree, too.  Every one is ONE, limited 
duplication
<[doug] D.GARRETT> But I am slowly getting used to the second and third...
<[Wil] W.BADEN1> Earlier we were talking about conditional compilation. This is 
not ...
<[doug] D.GARRETT> with their slight differences in vocabulary and i/o stuff
<[Wil] W.BADEN1> appropriate for standardization.
<[doug] D.GARRETT> ah.  I'm slowly getting a handle on what...
<[doug] D.GARRETT> you all mean by conditional compilation...
<[Wil] W.BADEN1> For Forth's eminent domain of bitty-boxes and toasters, we 
shud be ...
<[Wil] W.BADEN1> concentrating on what is minimum universal word repertoire.
<[Wil] W.BADEN1> What can you do without.
<[doug] D.GARRETT> Well, can't you almost arbitrarily chop off...
<[Wil] W.BADEN1> I'd like to report on a study I made recently.
<[doug] D.GARRETT> all but the first 100 common words, say 'This is the Kernal'.
..
<GARY-S> Toaster Ovens and Car emmission controls do not need a FAT Forth - 
that's for sure
<[doug] D.GARRETT> then publish the rest of a workman's toolbox as...
<[doug] D.GARRETT> "recognized extensions in the trade", and let the...
<[doug] D.GARRETT> implementor/programmer use what's needed.  I don't 
understand...
<[doug] D.GARRETT> why there's any controversy over what goes in and
<GARY-S> Some - would have all of us haul around extra tools they feel arre 
essential
<[doug] D.GARRETT> what will get added in the final product.
<[LEN] NMORGENSTERN> Yes you can, the word "arbitrarily" being the operative 
word.
<[LEN] NMORGENSTERN> The place where you would like to cut will be too high for 
some
<[LEN] NMORGENSTERN> and too low for others. Loud howls both ways.
<[LEN] NMORGENSTERN> The problem, as I said before is portability and 
communication.
<[doug] D.GARRETT> Well, why not say hey guys, here is the entire known...
<[LEN] NMORGENSTERN> "What you and your consenting computer do in private is 
your
<[LEN] NMORGENSTERN> own business and noblody else's." John Wavrik
<[doug] D.GARRETT> universe of Forth, with a so-called ANS Standard...
<[doug] D.GARRETT> kernal of primitives.  The rest is all standard programming 
practices
<[LEN] NMORGENSTERN> Doug, have you ever tried to take a program from Forth 
Dimensions
<[LEN] NMORGENSTERN> and make it work on your computer?
<[doug] D.GARRETT> extending the language to meet the...
<[doug] D.GARRETT> needs of the final product.   No, and I would'nt attempt it..
<[doug] D.GARRETT> But I will read WELL-DOCUMENTED source code in any...
<[doug] D.GARRETT> language.  Forth is not good at self-documenting...
<[Dave] OLORIN> I need to run, folks..
<[Dave] OLORIN> Take care...
<[LEN] NMORGENSTERN> In your minimal system, EVERY WORD EXCEPT THE 100 IN THE 
KERNEL
<[doug] D.GARRETT> so it's up to the programmer...
<[LEN] NMORGENSTERN> would have to be documented!!!!!
<[LEN] NMORGENSTERN> would have to be documented
<[doug] D.GARRETT> to get off his to embed good remarks so that anyone 
following...
<[doug] D.GARRETT> can read and know what's going on.  I can
<[doug] D.GARRETT> only imagine or guess what the big issue is about the kernal.
<[doug] D.GARRETT> if the Community has a good practice, it is published...
<[doug] D.GARRETT> and made available as a tool, the tool can be copied
<[doug] D.GARRETT> with it's documentation.  So WHAT if it sits on a shelf 
because
<[doug] D.GARRETT> no one needs to or wants to look at it.  If a problem occurs.
<[doug] D.GARRETT> and it concerns something about how somebody implemented a 
loop,
<[doug] D.GARRETT> it's gonna become fairly obvious to whoever works
<[doug] D.GARRETT> with the system to know how to debufg it.
<[LEN] NMORGENSTERN> But, Doug, didn't you just say that you wouldn't try to
<[LEN] NMORGENSTERN> port a program from Forth Dimensions? True, many good 
things are
<[LEN] NMORGENSTERN> published, but often in a Forth that is different from any 
of mine
<[LEN] NMORGENSTERN> and I don't have the time to work them out.
<[LEN] NMORGENSTERN> ga
<[doug] D.GARRETT> If you do it more than once in any piece of software...
<[doug] D.GARRETT> you need to document it.  It's an extension to a primitive 
kernal...
<[doug] D.GARRETT> you need to document it.  It doesn't matter whether you
<[doug] D.GARRETT> point to the 1987 ANS-MUMPS manual on the left shelf, or...
<[doug] D.GARRETT> Mr Brodie's Starting Forth 2d Edition on the right shelf...
<[doug] D.GARRETT> It's in print, nominally readable, and the programmer embeds 
in his
<[doug] D.GARRETT> source code, that he cribbed this and that from...
<[doug] D.GARRETT> pages x, y, and z because of this and that.  If he had to
<[doug] D.GARRETT> change the code, he makes a note of this deviation from...
<[Wil] W.BADEN1> Forth is different from Mumps, etc., becuz wher the money is, 
is in embedded systems. Len and I hav done a lot of translation from co
mp.sci. algorithms to Forth.
<[doug] D.GARRETT> the published "Standard" (left shelf or rioght shelf)...
<[LEN] NMORGENSTERN> Again, it's a matter of time. Wouldn't it be nice to study 
the
<[LEN] NMORGENSTERN> documentation briefly, and just load the screens, not 
examine & debug
<[LEN] NMORGENSTERN> in detail to make sure every word in the thing matches 
your Forth!!
<[Wil] W.BADEN1> But that's the second appeal of Forth (that I mentioned abuv).
<[doug] D.GARRETT> Wil, I think I agree agree again.  The issues surrounding 
the...
<[doug] D.GARRETT> proposed "need" for a standard kernal Forth...
<[doug] D.GARRETT> do not concern small-system designers.  We don't need the
<[doug] D.GARRETT> kitchen sink all the time.  I like a moderate
<[doug] D.GARRETT> size development system with electric windows and 
air-conditioning...
<[Wil] W.BADEN1> The essential idea is "kernel".  Becuz of the extensibility of 
Forth, evry Forther has his own kernel.
<[LEN] NMORGENSTERN> By the way, Doug, there are many that agree with your 
ideas, but
<[LEN] NMORGENSTERN> I think they go too far towards the minimalist side.
<[doug] D.GARRETT> but my target is gonna be about 8k.
<[Wil] W.BADEN1> But a development system for Forth duz not hav to be in Forth.
<[doug] D.GARRETT> I have occasionally imagined a Forth interpretter written
<[doug] D.GARRETT> in MUMPS, but that would be madness
<GARY-S> scuse me - I need to close the steno notes
=== End of Steno notes. ===