Minutes of the FIGGY BAR RT Conference.
 Date: 10/25/90   Time: 22:48EDT

  The question of the night "What do you think of the changes in ForthNet ?
  Do you even care ?" This prompted a discussion of the recent (dramatic)
  changes in the connectivity of this support resource. Other items discussed
  include implementing strings in Forth, hardware prototyping mag tape 
  controllers, management reluctance to accept Forth.

Attendees:
[[Gary] GARY-S]
[[LEN] NMORGENSTERN]
[[bill] B.CLARK10]
[[Dave] OLORIN]
[[Robert] R.BERKEY]
[[Wil] W.BADEN1]

Minutes:
<NMORGENSTERN> is here.
<B.CLARK10> is here.
<[bill] B.CLARK10> what's up?
<[Len] NMORGENSTERN> brb, all. I have to look up something.
<[Gary] GARY-S> Just chatting - any comment on tonight's question ?
<[bill] B.CLARK10> ForthNet?
<[Dave] OLORIN> is here.
<[Gary] GARY-S> yeah
<[Gary] GARY-S> Hey dave - care to comment on ForthNet ?
<[bill] B.CLARK10> hi Dave,    I'm only dimly aware of ForthNet.
<[Len] NMORGENSTERN> Hi Dave Long time no hear from!!!
<[Dave] OLORIN> Hmmmnnnn. I haven't been doing anything at all in Forth since
SigFORTH (I have to call Art to drop out of TCJ, I just don't have the time)
<R.BERKEY> is here.
<[Wil] W.BADEN1> is here.
<[Len] NMORGENSTERN> Hi all, a big crowd!
<R.BERKEY> Say, Gary, I finally found out what  :)  means...
<R.BERKEY> it's an abbreviation for  :-)
<[Gary] GARY-S> right
<[Gary] GARY-S> By the way, Dave our guest Nov 15 is Allan Furman - talking 
about SigForth
<[Len] NMORGENSTERN> Dave, have you started your new job yet?
<[Dave] OLORIN> Been on the job for more than a month.
<[Wil] W.BADEN1> Dave, tell us about your new job.
<[Gary] GARY-S> Anyone want to write a Forth column in a grass-roots magazine ? 
Dave is vacating such a position
<[Wil] W.BADEN1> Bill, do you know about LA FIG and OC FIG?
<[bill] B.CLARK10> I've been to OC fig, not LA fig
<[Len] NMORGENSTERN> (for strangers to California OC=Orange County
<[Wil] W.BADEN1> Was I at the OC FIG meeting?  If so, when was it?
<[Dave] OLORIN> It's all in C++ so far, with some work being done n TAL in the 
future.
<[Wil] W.BADEN1> TAL ??
<[bill] B.CLARK10> I remember seeing you down there a couple of years maybe 3 
ago
<[Len] NMORGENSTERN> What is TAL
<[Dave] OLORIN> A dedicated language for Tandems.
<[Len] NMORGENSTERN> Thanx
<[Gary] GARY-S> I would think the Tandems are running a BSD derivitive
<[Dave] OLORIN> New Tandems are running Unix. Old Tandems aren't. :-)
<[Wil] W.BADEN1> I rembember that one from the early 70's.
<[Gary] GARY-S> Last request - any comments on the changes in ForthNet ?
<R.BERKEY> Gary, is there really any change yet?
<[Gary] GARY-S> this week-end - yes !
<[Wil] W.BADEN1> So long as the info keeps rolling along...
<R.BERKEY> Or is it more a matter of concern as to what might happen in the 
future.
<[bill] B.CLARK10> what's FortNet, sorry
<[Wil] W.BADEN1> Gary, please epitomize the changes.
<[Gary] GARY-S> ECFB goes down this week-end. OLIS folded up. There is some 
reluctance among the pc-boards to move to Grapevine. Jack Brown may skewer
BCFB.
<[Len] NMORGENSTERN> pmi What is GrapeVine?
<[Gary] GARY-S> the new xCFB hub
<[Len] NMORGENSTERN> Thanx
<[Wil] W.BADEN1> Is there still a usenet connection?
<[Gary] GARY-S> yes - but that is the only segment remaining in tact.  Doug and 
I handle that.
<R.BERKEY> I seem to recall there is now another Forthnet link in Britain.
<[bill] B.CLARK10> Oh!
<[Gary] GARY-S> Was - it has been spotty from day one
<[Wil] W.BADEN1> Then we're in good shape.  You can't argue against the economy.
<[Gary] GARY-S> Lance also may kill the connection from oz
<[Gary] GARY-S> I really think sustained operation needs the input from the 
xCFB's
<R.BERKEY> I still don't understand how Grapevine changes anything.
<[Gary] GARY-S> That is what is in serious jeopardy
<[Gary] GARY-S> The recalcitrance of the other xCFB SysOps
<[Gary] GARY-S> Jack Brown is taking a sabattical and so far no one is willing 
to replace him in his absence
<[Gary] GARY-S> I think we are seeing the accumulative effect of burn-out, too
<[Len] NMORGENSTERN> We all like to think we are infinite, but
<[Len] NMORGENSTERN> we soon find there are limitations.
<[Gary] GARY-S> Oh well  - time will tell, and this is killing other 
conversations
<[Gary] GARY-S> Forth stuff, anyone ?
<[Dave] OLORIN> Only that OOFORTH 3.0 is on indefinite hold.
<[Wil] W.BADEN1> Who's for FORML this year?
<[Gary] GARY-S> ANS TC - want to tear into it, Robert ?
<[Len] NMORGENSTERN> I will be there with one paper and possibly two
<[Robert] R.BERKEY> I've been looking at string parameters.
<[Len] NMORGENSTERN> I am very interested in strings.
<[Robert] R.BERKEY> Wil suggested a while back two names for the end of 
strings, er address.
<[Len] NMORGENSTERN> I have tried to write a decent set of operators, and
<[Len] NMORGENSTERN> always bog down.
<[Robert] R.BERKEY> end and limit, where limit is end plus one.
<[Robert] R.BERKEY> The DO loop naturally takes a string with initial address 
and limit.
<[Gary] GARY-S> implemented with ALLOT ?
<[Robert] R.BERKEY> Limit seems to always be more useful than End.
<[Robert] R.BERKEY> But at the same time, Initial Limit seems to go hand in 
hand with
<[Robert] R.BERKEY> Initial Count
<[Len] NMORGENSTERN> Can't you use BOUNDS    ??? a l BOUNDS  limit origin
<[Robert] R.BERKEY> That's pretty much the end of the comment.
<[Robert] R.BERKEY> Right, BOUNDS converts and Initial Count to Initial Limit.
<[Len] NMORGENSTERN> The big problems I have involve concatenation. How
<[Len] NMORGENSTERN> to you "add" two strings together?
<[Len] NMORGENSTERN> (I mean the syntax, not the actual runtime code)
<[Robert] R.BERKEY> I don't have to do it often.
<[Robert] R.BERKEY> That always an issue, it's hard to solve complex problems 
without
<[Robert] R.BERKEY> specific cases.
<[Len] NMORGENSTERN> Well, I developed a system similar to DOS strings
<[Len] NMORGENSTERN> byte 1 is maximum length, byte 2 is current length
<[Len] NMORGENSTERN> and the rest is the string itself.
<[Len] NMORGENSTERN> I don't really like it, but it will do all the things
<[Len] NMORGENSTERN> that you hve in BASIC
<[Len] NMORGENSTERN> But not really smooth.
<[Robert] R.BERKEY> I'm presently got a sort of string stack that's really a 
linked list...
<[Robert] R.BERKEY> string link  string link.
<[Robert] R.BERKEY> Concatenating is pretty easy with that.
<[Len] NMORGENSTERN> Where do you store the strings? In the dictionary  space?
<[Robert] R.BERKEY> Because just moving the link up allows more characters to 
be added to the string.
<[Robert] R.BERKEY> Above the dictionary, where PAD is.
<[Wil] W.BADEN1> You are all missing what the problem is really about ...
<[Len] NMORGENSTERN> How do you handle garbage collection?
<[Wil] W.BADEN1> dynamic storage allocation and de-allocation.
<[Robert] R.BERKEY> I don't have the need to garbage collect.
<[Gary] GARY-S> malloc in Forth, Wil ?
<[Wil] W.BADEN1> Sure, it's part of LMI's system.
<[Gary] GARY-S> really ?
<[Len] NMORGENSTERN> Do they have a built-in string system, or did you write it 
yourself?
<[Wil] W.BADEN1> More than a year ago I uploaded a 6-line implementation.
<[Len] NMORGENSTERN> Here on Genie?
<[Wil] W.BADEN1> Yes, Len.
<[Len] NMORGENSTERN> I will look it up
<[Wil] W.BADEN1> It's a message, not a file.
<[Len] NMORGENSTERN> Oh. Do you remember the category/topic?
<[Wil] W.BADEN1> In F83X I used + to catenate strings, ! to store.  Inheritance 
was given by the vocabulary.
<[Wil] W.BADEN1> Sorry, Len.
<[Len] NMORGENSTERN> I have F83X, & will look for it.
<[Wil] W.BADEN1> Interesting, + was not commutative.
<[Len] NMORGENSTERN> It isn't in BASIC, etc, either.
<[Wil] W.BADEN1> I'm not advocating malloc for standard Forth.  The 
implementation is application dependent.
<[Len] NMORGENSTERN> Does the allocation/deallocation make it slow?
<[Gary] GARY-S> I was being a bit cheeky
<[Wil] W.BADEN1> You don't need dynamic allocation for toasters.
<[Len] NMORGENSTERN> F-PC has it.
<[Robert] R.BERKEY> Dynamic allocation, Len?
<[Wil] W.BADEN1> A circular string buffer is usually good enuf.
<[Len] NMORGENSTERN> No, but you could allocate memory for a string
<[Gary] GARY-S> Toasters ))))
<[Len] NMORGENSTERN> and then when you no longer need it, you could deallocate
it
<[Wil] W.BADEN1> I wudnt use FPC in a toaster either.
<[Len] NMORGENSTERN> This would handle garbage collection, needed when you
<[Len] NMORGENSTERN> are changing the length of strings frequently.
<[Len] NMORGENSTERN> (It might be slow, however.)
<[Len] NMORGENSTERN> Otherwise you would get cluttered with abandoned pieces of 
strings.
<[Wil] W.BADEN1> What kind of application are you considering, Len?
<[Len] NMORGENSTERN> None, to me it is an important programming problem
<[Gary] GARY-S> Garbage slows system response worse than the time it takes to 
collect and the discard stuff
<[Len] NMORGENSTERN> that I would like to be able to solve, if I ever need it.
<[Robert] R.BERKEY> As I think about it the system I have does lots of string 
manipulations,
<[Robert] R.BERKEY> but it's all handled on a special case basis, and there 
doesn't seem
<[Len] NMORGENSTERN> What system, Bob?
<[Robert] R.BERKEY> to be a need for a general solution.
<[Robert] R.BERKEY> It's test equipment used for testing long distance 
telephone lines.
<[Wil] W.BADEN1> Estimate the amount of storage you'l need, double it, multiply 
by the number of decades you'v been programming, and round up to a power
of 2.
<[Robert] R.BERKEY> Strings are sent and received over a modem.
<[Len] NMORGENSTERN> Ah
<[Wil] W.BADEN1> Right on, Robert.
<[Robert] R.BERKEY> An interesting problem is "reading" the strings that are 
received back.
<[Len] NMORGENSTERN> Then, you don't know in advance how long the string will 
be.
<[Robert] R.BERKEY> One of the programmers discovered that reading the strings 
starting at the
<[Len] NMORGENSTERN> I had not thought about that problem!
<[Robert] R.BERKEY> end was the best for that particular equipment.
<[Robert] R.BERKEY> Sounds like a job for a vocabulary.
<[Robert] R.BERKEY> or, lots of vocabularies.
<[Len] NMORGENSTERN> Granted that the 31-character limit on Forth words is 
arbitrary,
<[Robert] R.BERKEY> That's sort of what they implemented, with discrete DOES> 
structures.
<[Len] NMORGENSTERN> but could Forth dictionary handle these strings?
<[Robert] R.BERKEY> Yes, Len, they would work.  However, this system has no 
classical interpreter...
<[Robert] R.BERKEY> so any vocabulary structures are "to be implemented".
<[Len] NMORGENSTERN> Sounds like a very interesting challenge!!!
<[Robert] R.BERKEY> The string matching is done in high level, so using well 
built vocabularies
<[Gary] GARY-S> Jump in anytime, Bill
<[Robert] R.BERKEY> would have some advantages.
<[bill] B.CLARK10> just how much string manipulation is threre in a modem 
tester?
<[Robert] R.BERKEY> Um, the modems are communicating with a remote test device.
<[Robert] R.BERKEY> The remote test devices can measure things like resistance, 
hook up to
<[Robert] R.BERKEY> different circuits, etc.
<[Robert] R.BERKEY> On the other side of the data flow is a data base computer, 
as well as
<[Robert] R.BERKEY> the user's terminal.
<[bill] B.CLARK10> Ah!,  do the remote testers really need to send strings or 
can they send junst codes?
<[bill] B.CLARK10> (just codes)
<[Robert] R.BERKEY> So there are really four devices at any one time that are 
sending and
<[Robert] R.BERKEY> receiving strings of various forms.
<[Robert] R.BERKEY> The communication protocols with the remote test devices 
come in several
<[Robert] R.BERKEY> varieties.
<[Robert] R.BERKEY> One of them is simple codes, while others are English-like 
protocols,
<[Robert] R.BERKEY> so that user's with just a terminal can also "talk" to the 
devices, as
<[Robert] R.BERKEY> well as interpret the data stream coming over the modem.
<[Robert] R.BERKEY> ::USR:: DC 1 AP 100 TEST
<[Robert] R.BERKEY> might be an (albeit phoney) example of a message sent or 
received.
<[bill] B.CLARK10> um.      haven't used forth seriously recently, just 
assembler, but I have a new project to do and I'm thinking of forth, of course
 the bod--- boss wants it in 'C',    so have any BASIC or 'C' like string 
handling extensuions been made to forth?
<[Wil] W.BADEN1> There are several.
<[Robert] R.BERKEY> Well, that's just it, here's a Forth system with many 
solutions to the
<[Robert] R.BERKEY> various string issues, yet the Forth per se has no 
centralized string handling capability.
<[Robert] R.BERKEY> Just lots of specific solutions to specific problems.
<[bill] B.CLARK10> my task needs to be up and running yesterday so forth seems 
a natural
<[bill] B.CLARK10> but no one else here knows forth, when I leave they're stuck.
<[Robert] R.BERKEY> A Catch 22, but there's the prototype in Forth and convert 
to whatever solution.
<[bill] B.CLARK10> I guess they can always learn
<[bill] B.CLARK10> the prototype would be the final solution
<[bill] B.CLARK10> it's a one of a kind test fixture for tape controllers
<[Robert] R.BERKEY> What's the processor?
<[bill] B.CLARK10> 68020
<[bill] B.CLARK10> 20 MHz
<[bill] B.CLARK10> nice
<[Robert] R.BERKEY> yep, Nice!
<[Robert] R.BERKEY> Can you run the test application from a remote computer?
<[bill] B.CLARK10> do you know of a good/any 8--- 68000 forth  on genie
<[bill] B.CLARK10> and.....     remote um... an interesting idea.
<[Robert] R.BERKEY> Remote meaning 2 feet away, hahaha.
<[bill] B.CLARK10> I guess so. I could link a PC to the cpu cards serial port
<[bill] B.CLARK10> and just have some simple monitor code in the 68020 cpu
<[bill] B.CLARK10> ummmm... flexibility .     it's definitely woth.. worth 
considering.
<[Robert] R.BERKEY> Oh, I'm sure there are 68000 Forth's here, I just don't 
know much about them.
<[Gary] GARY-S> There are a couple specifically for mounting on C and a couple 
of native code 68K Forths in the library, plus a F83
<[Robert] R.BERKEY> F83 was first done on the 68K and Mitch Bradley also has a 
68K forth.
<[bill] B.CLARK10> I have a modified FIG forth I did 7-8 years ago
<[bill] B.CLARK10> but the new stuff F83... seems so mush easier to use
<[bill] B.CLARK10> I'll scan the software lib after this meeting then.
<[Robert] R.BERKEY> I'm heading out, see you all later.
=== End of Steno notes. ===