Minutes of a special GEnie Forth RoundTable Conference with Dr. Bob Loewenstein of Yerkes Observatory. Bob has acquired non-commercial rights to the Mac based Object Oriented Forth by Kriya and Charles Duff called Neon, hes enhanced it and renamed it YERK. Dr. Loewenstein joins us to discuss YERK. Entire contents of this transcript copyright (c) 1991 GEnie Forth RoundTable. The contents may be freely copied and distributed in whole or in part provided origination credit is included. Date: 09/19/91 Time: 22:28EDT Attendees: [[bob] FIGGUEST] <-- Invited special guest [[LEN] NMORGENSTERN] [J.T.SAVIDGE] [GARY-S] <---------- Moderator SysOp [[MIKE] M.KELLY36] [[Wil] W.BADEN1] [D.RUFFER] [S.WHEELER] Minutes: It's going to be in the 30's tonight (I'm Texan, too) Is looking at stars enough to keep you up there? sure...Actually, it's the weather. I like snow and Texas didn't have much James, did you get my letter? It is finally getting cool here & I dred it. yes, I mailed my response Monday. Apparently you haven't got it yet. I look forward to it. Basically I don't have an answer for you...you'll have to talk to Kriya since they retain commercial rights, whatever that means. Bob, set your name, kill echo and run status please - then let me know if you'va any questions Sorry JT - you two can resume chatting in a secc <[Bob] FIGGUEST> Ok, gary, I've done those things...My hard disk crashed this morning... <[Bob] FIGGUEST> and I lost the files from the other day. is here. will you be able to wing your opening remarks ? is here. <[Wil] W.BADEN1> is here. is here. Who besides wil is a Mac user in the peanut galley ? <[Len] NMORGENSTERN> I I am. SHOWTIME <[MIKE] M.KELLY36> peut-etre Let's get formal folks The GEnie Forth RoundTable is pleased to welcome as tonight's guest, Dr. Bob Loewenstein of Yerkes Observatory Bob holds a Ph.D. in Space Physics and Astronomy from Rice University. He is a Senior Research Associate in the Dept. of Astronomy and Astrophysics, University of Chicago and Director of Computing and Communications for the Center for Astrophysical Research in Antarctica. He is an infrared astronomer, observing at both ground-based and airborne observatories. He has been programming since 1968. In his last software project he developed was a remote controlled 3.5m telescope. Bob has acquired rights to the Neon Object Oriented Forth (OOF) system for MACs, enhanced it and renamed it Yerk. Please welcome tonight's guest, Bob Loewenstein. <[Bob] FIGGUEST> Let me begin by saying that I am not a computer 'expert.' I need to use computers and software in my line of research and as a consequence have had to create a lot of hardware and software to 'get the data.' Yerk is an object oriented variant of Forth. It allows a programmer to take advantage of Object Oriented constructs and techniques, which makes for a more intuitive program structure and a more maintainable code, as well as fast development time. I was a beta tester for Chuck Duff, Neon's primary author, while he was developing it in mid-to-late 1984. After Neon's demise I continued to refine and update Yerk because I felt it was the best Forth-like Object Oriented language out there. I don't really hold the rights to Neon/Yerk Kriya allowed me to release the sources for educational purposes, but they retain the commercial rights to Neon. One of the conditions I had to agree to for releasing it was to change the name since there is some dispute as to the ownership of the word 'Neon' as applied to software So, with that said, I'm ready for questions. <[MIKE] M.KELLY36> Why " Yerk " ? <[Bob] FIGGUEST> Ok, I've gotten grief over that. I wanted a name that no one would sue me for. Yerk fit the bill. However, some have suggestion FOOPLE, FORTH++ <[Wil] W.BADEN1> I thought the name made sense... It is an abbreviation of Yerkes, and is only four letters so it can replace Neon without increasing line length. <[Bob] FIGGUEST> Right..If I didn't mention it, I reside at Yerkes Observatory in Wisconsin. I figured it was a safer name than any noble gas. <[Len] NMORGENSTERN> How is Yerk for neophytes. In particular, how does it compare in complexity with other Forths that are popular with beginners, such as F83? <[Bob] FIGGUEST> Well, I think almost any Forth is good for beginners Yerk particularly so. It's free, has a good manual with a tutorial and is written on a fairly "standard" fig kernel. However at the present time it runs only on a macintosh. is here. The Objects/classes etc - are they pretty much lineal descendants of PARC research ? <[Bob] FIGGUEST> Somewhat. Yerk isn't a "true" OOL, in that not everything is an object. Many of the classes are patterned on Smalltalk classes. The syntax is 'selector object' rather than the reverse This, to me, is more readable. Typically a method is a verb and the object a noun, so verb noun reads better than noun, verb. I forget how many classes are in the standard Yerk application, ..about 50 or so. <[MIKE] M.KELLY36> What affect ( if any ) will the ANS standardization process have on Yerk ? <[Bob] FIGGUEST> Probably very little. For one thing, Yerk is loosely used by those people who have picked it up. There is no organization yet I am mainly interested in keeping it working on all Macintosh platforms for my own work. If a user's group starts up, with a clamor for standardization, that is another thing. But for now, I'm happy that it works on all Macs. And if you program using object oriented structure, the ANS probably wouldn't apply as it would to other real Forths. How is Yerk used in your work and are there any problems with System 7? <[Bob] FIGGUEST> I've been using Yerk (Neon) since 1984. We had a mac since they came out, but couldn't program it since nothing really was available then. We looked at MacForth, but heard about Neon (not called that then). We wanted to use the mac as a data acquisition and display fronted..also as an instrument control computer. I first programed up a data acq and display for our research done at Mauna Kea observatories and on the Kuiper Airborne Observatory. Then, as things began to role, we wrote quite a bit for lab use and eventually wrote a controller for one of the Yerkes telescopes. We made that remote controlled, and I have just finished version 1 of a remote control interface to a 3.5 meter remote control telescope in New Mexico. This involves some image processing. Charles Duff left Kriya to develop Actor - does he still invlove himself with YERK ? <[Bob] FIGGUEST> Not really. I see him every so often since we are friends. He visited me last year. He's too involved in Actor now. Are you involved with the VLA in New mexico <[Bob] FIGGUEST> No, but one of the programmers there used to work here at Yerkes on our 3.5m project. I have not been to the VLA since I'm not a radio astronomer. (editor note: VLA is the Very Large Array of ganged radio telescopes in South Eastern New Mexico) <[MIKE] M.KELLY36> Being an OO neophyte, what advantages does an OOF have over a ( your words ) " real Forth " ? <[Bob] FIGGUEST> Well, I've been programming in Forth for about ten years. Forth is fun, and all that, but it always was confusing to me to try to read other people's code. If Forth isn't written well, it is a bear to figure out. The control system for the Multiple Mirror Telscope was written in Forth, but it was so un-maintainable that I've heard they've rewritten it in C. OO imposes a structure that Forth doesn't require. It provides an intuitive framework for program design. This makes the program more easily maintained. Objects are chosen as intuitive constructs and are easily remembered. The late binding provides a powerful tool that frees you from using CASE statement-like structures. Particularly for Instrument control, OOL is good since Instruments are objects..and you send messages to them. <[MIKE] M.KELLY36> Understood, but are there not tradeoffs for the " intuitive framework " ? <[bob] FIGGUEST> I would say there are differences...you might call them... tradeoffs. One tradeoff might be OO takes up more memory I forgot to mention inheritance for a major benefit of OO over Forth. <[MIKE] M.KELLY36> But for embedded applications ... too big ( as a rule ) ? <[bob] FIGGUEST> I don't think so. A typical application on the Mac is about 120K. The basic Yerk with floating point is about 60K. ga <[Wil] W.BADEN1> Kurt Schmucker, in "Obj Ori Prog for Mac" sez that Neon is the fastest performing implementation of SmallTalk. Would you comment? <[bob] FIGGUEST> I remember reading that way back. Neon isn't an implementation of smalltalk, but it is an OOL implementation. It is fast, but not as fast as a compiled language. Typically about 2-3 times slower than, say, a compiled C program...But like any Forthlike language, 90% of the time is spent in 10% of the code. You can still optimize that part. (Maybe it was 80/20). I am really not familiar with other OO implementations except for MacApp. Closing remarks please Bob <[bob] FIGGUEST> Well, I hope people have a chance to get a copy of Yerk if they don't already have it. It is transportable to other machines and several people have written to say they are attempting to port it to PC, Sun, etc. If anyone has any questions about Yerk, feel free to get in contact with me. I think it is a powerful language, and it is free. We didn't get a chance to discuss the application issues of YERK, but this has been an interesting conference. Thanks for taking the time to chat with us Bob. This conference is officially closed. === End of Steno notes. ===