Minutes of FIGGY BAR RT Conference.

Date: 05/23/91   Time: 22:27EDT
Attendees:
[GARY-S]
[[bob] R.CAVANAUGH]
[[ddJAX] JAX]
[[Len] NMORGENSTERN]

  Open FIGGY BAR of 23 June 1991. Items discusssed include good initial
  reaction to Upper Deck Forth, Why ANSI Forth ?, Forth print resources,
  Forth's lousey acceptance at University C/S departments, comparing C
  to Forth for prototyping - score a point for interactive programming,
  some-not-so good things about Forth, MVP Forth reprised. 

Minutes:
<R.CAVANAUGH> is here.
<C.MORAVETZ1> is here.
<R.CAVANAUGH> Hi. this is my first time at the Figgy Bar. What is happening 
tonight?
<GARY-S> Open FIGGY BARs are largely just that - totally open, name your own 
poison
<GARY-S> Welcome - you're the first and second customers - any (FORTH) axes to 
grind ?
<[Cliff] C.MORAVETZ1> Forth axes? No.
<R.CAVANAUGH> I have a plug if it is OK.  I just purchased Upper Deck Forth for 
the PC V2.00 and so far I have been really impressed.  A lot of bang f
or the buck
<GARY-S> You are the first UPPER DECK user I have heard from - glad to know 
they have a good product
<GARY-S> Why don't you give a brief synopsis of the product (ie: JSR, ITL how 
threaded... etc)
<[Cliff] C.MORAVETZ1> Anyone have news about the ANSI standard effort?
<[Cliff] C.MORAVETZ1> The FIG guest last week didn't have much good to say 
about it.
<[bob] R.CAVANAUGH> I was on Usenet today, and downloaded a synopsis, but 
haven't read it yet
<GARY-S> Cliff - that is precisely why I asked Brad to appear. I wanted some 
thought provoking comments
<[Cliff] C.MORAVETZ1> I had one thing to say, but it seemed like things were 
too busy.
<GARY-S> No, Bob, I mean give us an overview of Upper Decck Forth
<[bob] R.CAVANAUGH> Sorry, lag caught me.  Upper Deck is direct threaded, case 
sensitive.  Next is inline.  The editor is brought in dynamically so it
 does not take any application code space, it has mouse support, color display, 
and lots of the same words as F83
<[bob] R.CAVANAUGH> Also Gary, the support is excellent so far.  I am porting C.
 Lindley's windowing package, and there are a few quirks to the segmen
ted architecture.  Peter Graves was very helpful on the phone, and after about 
10 min we just started "chewing the rag" about Forth in general.
<GARY-S> Say it now Cliff - that's what these open bars are all about
<[Cliff] C.MORAVETZ1> What I can't figure out is why ANSI? It seems to me...
<[Cliff] C.MORAVETZ1> that the Forth standards team did well enough before.
<GARY-S> Bob - any opinion on how hard it would be to port F83/F-PC etc 
applications to UDF ?
<GARY-S> Cliff - ANSI _IS_ the official sanction some vendors need for 
contracct jobs - that's why
<[Cliff] C.MORAVETZ1> The magic four letters $peak!
<GARY-S> DoD for instance has decreed _any_ non-Ada language MUST be Standard - 
either IEEE or ANSI
<GARY-S> Is the support line 24 hours Bob ?
<[bob] R.CAVANAUGH> As far as porting F83, shouldn't be too much hassle.  F-PC 
handles the stack differently so you would have to be careful, especial
ly with code words.  Don't know about 24Hrs, I called this afternoon about 1700 
PST
<GARY-S> sounds like solid support
<GARY-S> any bugs thgat you can't live with ?
<[Cliff] C.MORAVETZ1> June bugs!
<GARY-S> soon :-)
<[bob] R.CAVANAUGH> cliff- I work on DoD projects, and so far we have been able 
to use "C" (not forth <sigh> ), and our complier is ANSI.  Projects co
ming down the pipe will probably require ADA <gag>.
<GARY-S> _or_ if non-Adad, they must have the ANSI or IEEE seal of approval !
<[Cliff] C.MORAVETZ1> I read just last night about the difficulty of running 
Ada under Unix...
<[bob] R.CAVANAUGH> As far a bugs, none that I have been able to find, though 
pse remember guys I got the demo two weeks ago and the main system today
.  I am going to put it through its paces this weekend though ;-)
<[Cliff] C.MORAVETZ1> There's some real problems there.
<GARY-S> GEEEZ ! - who would want too ?????? GAGGGGG
<GARY-S> Bob - I will really be interested in your comments
<GARY-S> Can you post some in the Bulletin Board ? If not next Thursdays FIGGY 
will be good
<[Cliff] C.MORAVETZ1> The government wants to!!
<[bob] R.CAVANAUGH> Gary- the last thing that I read was that the TC is really 
close to submitting the strawman.  Do you have any newer information?
<GARY-S> I think that is accurate, Bob, but I have no confirming information
<JAX> is here.
<[Cliff] C.MORAVETZ1> Hi Jax. Read your article in DDJ.
<GARY-S> Jaxon - how you ?
<[bob] R.CAVANAUGH> Jax - Hi! If you didn't get my mailing, I wanted to congrat 
on the two articles in ED and DDJ. Nice stuff!
<NMORGENSTERN> is here.
<[ddJAX] JAX> Thanx Bob, I did ... didn't you give me some code, too?
<[Len] NMORGENSTERN> Hi all
<[Cliff] C.MORAVETZ1> When was the ED article, Jax?
<[bob] R.CAVANAUGH> Jax- no, I don't think so
<[ddJAX] JAX> Cliff -- I think Bob means ESP.
<GARY-S> I _just_ got in a few minutes before the FIGGY
<[Len] NMORGENSTERN> Well, I was (more than) a few after!
<[Cliff] C.MORAVETZ1> I didn't know ESP was a MAGAZINE :)
<[Len] NMORGENSTERN> (Embedded systems programming)
<GARY-S> Jax - understand you and Dave W. finally met
<[bob] R.CAVANAUGH> Embedded systems programming may 1991 pg38 to be precise ;-)
<[ddJAX] JAX> << -- looking up Bob's letter.
<[bob] R.CAVANAUGH> For some reason I keep thinking the mag is "Embedded 
Design".  Brain must be going...
<[ddJAX] JAX> Aha, yes, you expressed the thought (widely shared) that DDJ was 
turning into a "C" rag.
<GARY-S> Bob - since UDF is Direct Threaded I would imagine it is faster than 
many Forths
<GARY-S> IT IS a C rag
<[ddJAX] JAX> They are actually quite eager to dig into the fertile fields of 
embedded control, and thus, into FOrth.
<[Cliff] C.MORAVETZ1> So what if it is a C rag?
<[ddJAX] JAX> They see what a success ESP has been, and are willing to learn 
from their own "children" as it were!
<[Len] NMORGENSTERN> It used to cover more ground than it does now, Cliff
<[bob] R.CAVANAUGH> gary- the benchmark advertised is 10 interations of the 
Sieve on a 4.77 PC in 38 seconds.  I haven't tried it on my 386
<GARY-S> I prefer ESP - and will stay with them
<[ddJAX] JAX> ESP is a great magazine because it knows what it is.
<GARY-S> That certainly sounds respectible, but sieve times are deceptive - as 
Don Colburn proved many years ago
<[ddJAX] JAX> That is typical of the large chain of Miller Freeman publications 
(now including "Guitar Player" and "Keyboard" magazine).
<[ddJAX] JAX> They keep focused on their real audience.
<[ddJAX] JAX> .
<GARY-S> Miller-Freeman gutted a few on the way, too, I think Tyler should get 
the credit
<[bob] R.CAVANAUGH> cliff- don't get me wrong; I like C, but I think that Forth 
is a much better candidate for embedded processing.  I am required to
use C at work on the deliverable product, but I find Forth _invaluble_ for 
testing purposes, and I can't help wondering if it wouldn't shorten the deb
ug cycle if we had it resident in our product.  I would like forth to get more 
than "what is that?"
<[ddJAX] JAX> What is your product?
<GARY-S> Bob - are you aware of the houses that prototype in Forth ?
<[bob] R.CAVANAUGH> jax- real-time communication processing in digital data 
links.
<[Cliff] C.MORAVETZ1> What kind of data protocol, Bob?
<[bob] R.CAVANAUGH> gary - You know, I didn't even hear about forth until a 
friend of mine at work rolled his own and used it in debugging monitor.  I
 didn't even see it once at college, which I think is a sad state of affairs.
<[Len] NMORGENSTERN> Indeed it is!
<[ddJAX] JAX> Gary is right, Tyler is a slick operator. Tyler has a lot of 
*taste* and taste is what makes an editor.
<[bob] R.CAVANAUGH> cliff- Unfortunately, the DoD doesn't like us to get too 
specific
<GARY-S> Very sad - also abundantly "normal" of most University C/S departments
<[Len] NMORGENSTERN> Do you have a Forth Interest Group in Salt Lake-Prove?
<[Len] NMORGENSTERN> (Provo, that is)
<[Cliff] C.MORAVETZ1> I keep hearing this lament now and then...
<[Cliff] C.MORAVETZ1> Why is Forth so unpopular?
<[Len] NMORGENSTERN> Lack of "critical mass," Cliff
<[Cliff] C.MORAVETZ1> Well, isn't partly do to the fact that Forth is...
<[bob] R.CAVANAUGH> I however, am trying to end-run this.  I gave a friend of 
mine F-PC which he is teaching to his kids in Boy Scouts instead of BASI
C.  If I can get enough people interested in his troop we may have a little 
kernel of "true believers"...Len, I do not know of any chapter here.  I re
ally only know about three forth programmers, of which yrs trly is one.  I have 
been a member of Fig for 2 yrs, just so I can stay current with the la
nguage. Well, I do like deserts
<[Cliff] C.MORAVETZ1> exceptionally unconventional?
<GARY-S> It is non-traditional and instructors can't confine the student's 
usage once the student grasps the assembler - to - High Level reach of Fort
h
<[Len] NMORGENSTERN> (brb -- someone at the door)
<[bob] R.CAVANAUGH> I had plenty of classes in assembler which could have been 
taught in forth.
<GARY-S> You do understand College professors prefer a controlled environment - 
and Forth isn't that ! ???
<[Cliff] C.MORAVETZ1> It seems that Forth is ending up in the same kind of 
situation as APL...
<[Cliff] C.MORAVETZ1> A few true believers, and ignored by the mainstream.
<[bob] R.CAVANAUGH> Besides, L Brodie's _Thinking Forth_ gave me some of the 
best ideas for design in any language!.  I think it should be required re
ading for C/S degrees!!
<[Len] NMORGENSTERN> (Out of print, I believe)
<GARY-S> Both APL and Forth are often referred to in negative ways in the same 
breath by people whose minds end with Pascal bubble sorts
<[bob] R.CAVANAUGH> Hear Hear! And most of them run software departments
<GARY-S> "Thinking Forth" _is_ out of print - damn shame, too
<[ddJAX] JAX> Dave Weinstein was at an auction with me last week (as Gary 
mentioned) and bought an old IBM 5010 APL machine!
<[Cliff] C.MORAVETZ1> Well, for those of you I've met with here before...
<[Cliff] C.MORAVETZ1> I see serious defects with Forth. Don't care to go into 
it now though!
<[Cliff] C.MORAVETZ1> I like to hang out here because I find you guys 
interesting.
<GARY-S> please do Cliff - that is a perfect use of this time
<[bob] R.CAVANAUGH> Cliff- I don't think that anybody here would argue that 
some part (of any flavor) has its drawbacks.  Don't sweat it.
<[Cliff] C.MORAVETZ1> Well, I think I spent about 2 1/2 hours talking to Len 
about it one night.
<[Len] NMORGENSTERN> I remember!
<[Cliff] C.MORAVETZ1> Let's try something different...
<[bob] R.CAVANAUGH> However, I have yet to find an enviroment where I can 
exercise hardware so quickly and freely.  For me, it has a great many plusse
s.  Minuses yes, but then, we could be forced to use ADA...
<[Cliff] C.MORAVETZ1> Instead of me complaining about Forth, and you guys...
<GARY-S> How about some key points and see if anyone has reasonable rebuttals
<[bob] R.CAVANAUGH> what do you want to discuss?
<[Len] NMORGENSTERN> (Nothing is ever so bad that something else couldn't be 
worse, Bob)
<[bob] R.CAVANAUGH> Amen, len
<[Cliff] C.MORAVETZ1> complaining about other languages, What do you guys think 
are the defects..
<[Cliff] C.MORAVETZ1> of Forth?
<[Cliff] C.MORAVETZ1> Is Forth so great that you can't see any warts?
<[Len] NMORGENSTERN> Forth requires rigid discipline with respect
<[Len] NMORGENSTERN> to naming, structuring and documentation
<[Len] NMORGENSTERN> It is so pwerful that you can dig yourself
<[Len] NMORGENSTERN> into a deep hole very quickly.
<GARY-S> No - it is difficult to manage as a group profect language, it tends 
to be cryptic without judicious commenting...
<GARY-S> but the freedom to wrap the language around your project instead of 
the other way around is damn hard to ignore
<[bob] R.CAVANAUGH> When I first started with the language, the total lack of 
DOS support. You had to write your own file handlers, and I could not go
 from an editor like PE to L&P's line editor.  It kept me from really picking 
up the language for several years.  Every time I had an application, it
was easier and cheaper to use my C compiler
<[bob] R.CAVANAUGH> Should I use multiple one lines?
<[ddJAX] JAX>
<[ddJAX] JAX> Right now I am working in MVP Forth! Talk about the "bad old 
days" :-)
<[bob] R.CAVANAUGH> I also think the way good syntax looks is the most elegant 
and readable code I have seen
<[bob] R.CAVANAUGH> ex: turn on led 5 seconds turn off led
<GARY-S> Direct hardware access is not a shabby thing to have either
<[Cliff] C.MORAVETZ1> Well, direct hardware access is very doable with C. Not 
interactive, tho.
<[bob] R.CAVANAUGH> Also, if the programmer is careful, code _can_ be portable 
(for Forth-83)
<[bob] R.CAVANAUGH> cliff- You have hit it on the head for me
<GARY-S> Then - the real decision is presonal preference. Thank goodness for 
that !
<[bob] R.CAVANAUGH> I have spent hours testing on a portable
<[bob] R.CAVANAUGH> and running back and forth to the main machine to change 
one line
<[bob] R.CAVANAUGH> of testing code.  I can just forget and recompile at the
<[bob] R.CAVANAUGH> bench with forth
<GARY-S> I personally hate to recompile for EACH code segment debug
<[bob] R.CAVANAUGH> What about the time you forget to initialize that last 
little pointer and the stepper motor scraped across the platen?
<GARY-S> That is also why many Unix programmers program in Shells instead of 
'C', the interactive environment is a smooth singing sirene
<[Len] NMORGENSTERN> pmi, what is Shells
<[bob] R.CAVANAUGH> Though I must admit, Borland C products are a nice tool for 
some applications. Their debugger is hard to beat
<[ddJAX] JAX> There are three great things about Forth vs C in Embedded Systems:
<[ddJAX] JAX> 1) Forth is cheaper
<GARY-S> Shells = interactive command interpreters in Unix
<[ddJAX] JAX> 2) the Forth environment is simpler ..( i do Forth on an XT with 
two floppies, C on a 386 networked to a 286)
<[bob] R.CAVANAUGH> That assumes you have the resources and Unix platform
<[ddJAX] JAX> 3) You don't have to come up with all the faked data sets at 
compile time, you can just interpretively dump them on the stack at run tim
e.
<[ddJAX] JAX> .
<[bob] R.CAVANAUGH> we have a diagnostic display for our systems written in 
Forth.  I can
<[bob] R.CAVANAUGH> look directly at memory real time, and write colon 
definitions to
<[bob] R.CAVANAUGH> make the changes I want at the system on the fly. Jax is 
right. We use
<[bob] R.CAVANAUGH> Chordata 4.77 pc clones for "portable" testing.  They work 
fine under forth
<[bob] R.CAVANAUGH> P.S jax- ours is a mod to MVP. I agree with you 100%!
<[bob] R.CAVANAUGH> Well guys, I have to sign off.  This was my first time and 
its a lot of fun. See you all soon - Bye
<[Cliff] C.MORAVETZ1> So often when Forth is touted, it's always the embedded 
applications...
<[ddJAX] JAX> The big problem is I used MVP a long time ago and forgot a lot of 
things!
<[Len] NMORGENSTERN> See you again Bob.
<[Cliff] C.MORAVETZ1> that are the showcase.
<[bob] R.CAVANAUGH> has left.
<[ddJAX] JAX> bye bob
<[Cliff] C.MORAVETZ1> Bye, Bob.
<GARY-S> In defense of Glen Haydon, MVP was a class act when written, and well 
supported
<[ddJAX] JAX> Oh yes, and portable.
<[ddJAX] JAX> But one gets spoiled by modern Forth tools!
<[ddJAX] JAX> Glen is still staunch in support of MVP ...
<GARY-S> All life marches on - it's a rule, you kknow <grin>
<GARY-S> Let's wrap this one up
<[ddJAX] JAX> but he promises he is going to give serious consideration to MVP 
ANS-Forth when the time comes.
<[Cliff] C.MORAVETZ1> How about modern data processing? I mean the big dual IBM 
4090 installation.
<[ddJAX] JAX> .
<[Cliff] C.MORAVETZ1> Like the kind where you have 5 billion records!
<GARY-S> Several insurance companies do their actuary work on such machines 
using Forth !
<GARY-S> I'm closing the notes and splitting folks
=== End of Steno notes. ===