Minutes of the Special GEnie Forth RT Conference with Bill Ragsdale, president of Dorado Systems. Bill discussed his company's approach to 'Targeted Applications in Forth'. Also discussed was various Forth models and what Bill sees as a major failure by the X3J14 ANS Technical Committee as they construct the new standard. Entire contents of this transcript copyright (c) 1990 GEnie Forth RoundTable. The contents may be freely copied and distributed in whole or in part provided origination credit is included. Date: 08/16/90 Time: 22:33EDT Attendees: [[Bill R] FIGGUEST] [GARY-S] [[Ting] C.TING] [[Andy] A.MCFARLAND] [[Brian] B.SUTTON1] [[FRANK] F.SERGEANT] [[Dave] OLORIN] [[Wil] W.BADEN1] [[IRV] I.MONTANEZ] [[Len] NMORGENSTERN] [[Robert] R.BERKEY] [[Dennis] D.RUFFER] Minutes: is here. have you typed /xec to kill echo and /sta to confirm you are job 1 and I am job 2 <[Bill R] FIGGUEST> yes C. H. Ting - have you had a chance to look at Sergei Baranoff's minimal Forth, 'beta' in the library ? <[Bill R] FIGGUEST> who is really Jennifer Marlowe? a log file personalized <[Bill R] FIGGUEST> I thought it might be from Sherlock Holmes! ga is here. The name is from a detective series, she was Mike Hammer's secretary I believe Gary, I just saw it in the library. <[Bill R] FIGGUEST> It had that pulp novel ring to the name. I personally believe you are correct, minimal (assembled) Forth's are necessary for hacker interest Let's just hope 1) you are correct, and 2) their re-emergence is not too late is here. <[Andy] A.MCFARLAND> Hi -- I am new to the real time conference is here. I will hit you with some rules shortly, that we have to maintain for formal conferences, till then it is just like a country telephone - jump in when ready is here. is here. <[Wil] W.BADEN1> is here. 1 minute to show time folks 1 minute and ticking <[Dave] OLORIN> (puts on his soon-to-be-unemployed-programmer hat) is here. is here. Before we begin tonight's conference I need to remind ... everyone this is a formal conference. .. .. .. That means NO interlacing. Request a turn to speak by typing ... !! or /RAI . I will keep a list and call on whoever ... is next. PLEASE follow these simple rules : ... PLEASE make your first response ok (so we know YOU know). ... Then, keep your lines short, ending incomplete thoughts ... with three dots (...) and ending your question/statement with ga . thank you. The GEnie Forth RoundTable is very pleased to welcome this ... evening's special guest, Bill Ragsdale, President and founder... of Dorado Systems, maker of components in computerized management... of building security. All Dorado products are programmed in ... Forth. Bill recently became President and investment advisor... of the Sunshine Growth Trust, a public mutual fund. ... Bill is a member of the ANSI Standards Committee for Forth ... continuing his earlier efforts as Chairman of the Standards ... Team for Forth 79 and 83. ... Bill was the founding President of the Forth Interest Group,... guiding its growth from 1979 to about 1985. Bill and wife ... Anne have two children and reside in Hayward, CA. ... Bill's topic is "Targeted Applications for Forth" ... ( or where it 'Should' be the primary language) - Please help welcome Bill Ragsdale. .. .. .. .. .. .. ga Bill <[Bill R] FIGGUEST> Hi, All Im glad to be here thru the magic of Genie ga <[Bill R] FIGGUEST> We need commercially viable Forth applications that are motivating . .. <[Bill R] FIGGUEST> to the buyer (company) and seller (Forth programmer)... <[Bill R] FIGGUEST> This is needed for Forth to be rewarding to all and... <[Bill R] FIGGUEST> viable in the long term. I propose "tailored applications". <[Bill R] FIGGUEST> These are standard and custom software in a product form... <[Bill R] FIGGUEST> The features are - quick turnaround. <[Bill R] FIGGUEST> - done in a field the programmer knows well... - have customers under a time pressure <[Bill R] FIGGUEST> I would look for applications written from an application <[Bill R] FIGGUEST> shell which the Forth programmer customizes... <[Bill R] FIGGUEST> The prices would be $500 to 1,000 in 1 - 2 hours. <[Bill R] FIGGUEST> or $5,000 to 10,000 when it requires 1 - 2 days... <[Bill R] FIGGUEST> We do such products at Dorado Systems, in this price range.. <[Bill R] FIGGUEST> Any comments <[Dave] OLORIN> At the risk of being burned as a heretic... <[Dave] OLORIN> I don't see a future for Forth outside of small... <[Dave] OLORIN> niche markets. <[Dave] OLORIN> Let me explain. <[Dave] OLORIN> I've spent the better part of the last year... <[Dave] OLORIN> working in a part of the industry which has... <[Dave] OLORIN> been dominated by a language very similar in both... <[Dave] OLORIN> flexibility and fanaticism of its adherents to Forth. <[Dave] OLORIN> The research community, dominated for years by Lisp, <[Dave] OLORIN> and using Lisp specific hardware, is moving *away* <[Dave] OLORIN> from Lisp and into C and (especially) C++. <[Dave] OLORIN> This is a language with far more programmers than Forth, <[Dave] OLORIN> a vastly greater respectability, and an equally vehement <[Dave] OLORIN> user base, and yet it is dying off because of the need <[Dave] OLORIN> for standardization. <[Dave] OLORIN> I don't think that Forth will survive outside of isolated <[Dave] OLORIN> quick turn-around markets (such as you've described). We're <[Dave] OLORIN> even losing the astronomy base that we had for so many years <[Dave] OLORIN> do to the proliferation of high-power workstations. <[Dave] OLORIN> Can Forth survive the sudden loss of resource scarcity? <[Bill R] FIGGUEST> I agree. I think Forth will continue only in niche areas such as <[Bill R] FIGGUEST> robotics and graphics. The reason targeted applications are <[Bill R] FIGGUEST> appropriate is to leverage on the programmers prior knowledge <[Bill R] FIGGUEST> of the application area and because of Forth's quick turnaround. <[Bill R] FIGGUEST> We areI wish the overall outlook was better. <[Wil] W.BADEN1> Bill, when last we spoke face-to-face, you said that most of Dorado's work is now in another language. Is that still so? and what is that other language? <[Bill R] FIGGUEST> No thats not the case. All our <[Bill R] FIGGUEST> applications are written in Forth and all of our <[Bill R] FIGGUEST> test equipment is done in Forth. We are doing <[Bill R] FIGGUEST> 1 to 4 K one chip processors for realtime input and <[Bill R] FIGGUEST> the format conversions to the users special I/O formats. <[Bill R] FIGGUEST> We are stricktly a Forth shop. <[FRANK] F.SERGEANT> How do you prospect for customers? <[Bill R] FIGGUEST> Our customers don't even know we use Forth. We sell the <[Bill R] FIGGUEST> product with customization offerd at a modest cost... <[Bill R] FIGGUEST> generally a 10 piece miminum order. The customizatio <[Bill R] FIGGUEST> comes along at to premium. We are shifting to <[Bill R] FIGGUEST> a $1500 charge in addition to the 10 piece minimum. <[Bill R] FIGGUEST> Our advantage, as a product manufacturer is that we <[Bill R] FIGGUEST> have the ongoing sales relationship, independent of Forth. <[Bill R] FIGGUEST> The programmer also has the problem of prospecting for <[Bill R] FIGGUEST> clients. <[Ting] C.TING> I do not understand how you can produce a $1000 product in 2 hours. <[Bill R] FIGGUEST> We do it. . .. <[Bill R] FIGGUEST> The hardware is a $150 card reader with a standard PC board <[Bill R] FIGGUEST> and aprocessor. The user sends us a sample ID card and or <[Bill R] FIGGUEST> an input format. We customize 1 or two screens of code <[Bill R] FIGGUEST> to move the card digits into the specified output format. <[Bill R] FIGGUEST> we have about 6 shell programs to work from. The total bill <[Bill R] FIGGUEST> to the customer is $1500 which includes the 10 card readers. <[Bill R] FIGGUEST> We have been disappointed with the reorder rate and use <[Bill R] FIGGUEST> an added $1500 one time charge to filter out the buyers <[Bill R] FIGGUEST> who are not committed to a longer term relationship.. <[Robert] R.BERKEY> I didn't understand what kind of "card"s these are. <[Bill R] FIGGUEST> These are security ID cards coded like credit cards. <[Bill R] FIGGUEST> Similar product would be electronic scales,... <[Bill R] FIGGUEST> data colection unit, time and attendace time clocks. is here. I presume you are reading mag-stripe ... are you using the IATA field or ABA ? <[Bill R] FIGGUEST> Yes, in this field... <[Bill R] FIGGUEST> we know a lot about little, a very narrow application field. <[Bill R] FIGGUEST> Most of the work is ABA on Track 2, but we also do IATA <[Bill R] FIGGUEST> on track 1 or 3. any thrift third (track 3 ) ? <[Bill R] FIGGUEST> The thrift or MINTS format on track 3 seems to be fading away. <[Bill R] FIGGUEST> I think people use the IATA on three. Only IBM and some <[Bill R] FIGGUEST> financial institutions seem to use the IATA and MINTS format. <[Bill R] FIGGUEST> Another targeted application area is bar code reader <[Bill R] FIGGUEST> customization. This is hotter than magnetic stripe now. <[Bill R] FIGGUEST> A program could probably do well with <[Bill R] FIGGUEST> a simple bar code reader with an internal one chip <[Bill R] FIGGUEST> processor and do customized applications. ga are you doing flat graphics (bar code) also <[Bill R] FIGGUEST> No we specialize in higher security applications of <[Bill R] FIGGUEST> magnetic striped cards. Nuclear power plants, large <[Bill R] FIGGUEST> office buildings and so on. Another application <[Bill R] FIGGUEST> area would be the access control units that actually <[Bill R] FIGGUEST> control the door. There are about 40 manufactures of these <[Bill R] FIGGUEST> units, but they are slow on the draw for customization. <[Bill R] FIGGUEST> Several firms I know have lost the firmware source code <[Bill R] FIGGUEST> for their systems and can't change! <[FRANK] F.SERGEANT> What microprocessor do you use/prefer (& why?)? <[Bill R] FIGGUEST> We use the Mot 68705 CMOS EPROM and the 68HC11 for larger jobs. <[Bill R] FIGGUEST> presently all our jwork is in EPROM with a little EEPROM <[Bill R] FIGGUEST> on board for start up parameters we change on the <[Bill R] FIGGUEST> production line. The HC11 is particulary fine for forth <[Bill R] FIGGUEST> and for our work. We have found several clients who <[Bill R] FIGGUEST> are also using these,but programmed in assembler or C. <[Ting] C.TING> Bill, You can choose the tool because you own the business... <[Ting] C.TING> But we have to survive by convincing the boss that Forth is ok.. <[Ting] C.TING> Any thought about winning an argument and losing the job? <[Bill R] FIGGUEST> That is the case. , but as sales manager I <[Bill R] FIGGUEST> want the benefits of forth. I walk up to Andrew McKewan, <[Bill R] FIGGUEST> our programmer (and Enginering Manager) with <[Bill R] FIGGUEST> a need and he fills it in 1/2 hours. Sell that <[Bill R] FIGGUEST> quick turn around to the boss and he'll support <[Bill R] FIGGUEST> Forth. There are other ways but we get the results. <[Bill R] FIGGUEST> I havent worked in C environments, but most prior <[Bill R] FIGGUEST> cases I saw 1 to 4 day turnaround on similar projects. <[Bill R] FIGGUEST> We need to sell the results of Forth not just Forth. <[Bill R] FIGGUEST> Forth programmers need to keep one factor in the <[Bill R] FIGGUEST> back of their mind. I, as a Forth progrmmmer want <[Bill R] FIGGUEST> to do things my way. When I make tradeoffs I still have <[Bill R] FIGGUEST> to keep the manager/company point of view in mind. <[Bill R] FIGGUEST> Many of the Forth programmers seem to grow a defensive <[Bill R] FIGGUEST> or defiant attitude that their method or Forth dictates <[Bill R] FIGGUEST> what is to happen on the project. Managers may <[Bill R] FIGGUEST> sometime change the language to change the people they <[Bill R] FIGGUEST> work with in the future. Think it over. <[Robert] R.BERKEY> Have you had occasion to manage more than one Forth <[Robert] R.BERKEY> programmer on a given project? <[Bill R] FIGGUEST> No, I still think it is one programmer/one project. Its <[Bill R] FIGGUEST> like the Texas ranger sent to stop a riot. When asked <[Bill R] FIGGUEST> where his help was he said 'You only have one riot so <[Bill R] FIGGUEST> you only need one Ranger'. <[Wil] W.BADEN1> I believe Work Stations and Windowing Systems have made a tremendous improvement in turn around. Can Forth keep its monopoly on quick turn around much longer? <[Bill R] FIGGUEST> I worry about both of those items... <[Bill R] FIGGUEST> I used to know every bit and flag in my computer system. <[Bill R] FIGGUEST> When forth is running with the virtual Forth machine at least knew all of the Forth internals. .. <[Bill R] FIGGUEST> now with the computers so advanced, we are losing touch with <[Bill R] FIGGUEST> the real hardware and the virtual hardware. I think <[Bill R] FIGGUEST> the trend is not reversable, but it works against small <[Bill R] FIGGUEST> applications and individual programmers. it does not <[Bill R] FIGGUEST> bode well for Forth. <[Ting] C.TING> I think there are still lots of room for Forth applications in.. <[Ting] C.TING> embedded systems... <[Ting] C.TING> Bill, I would like to ask you, as the author of figForth, ... <[Ting] C.TING> that Forth resulted from metacompilation vs from straight <[Ting] C.TING> assembly which one should we provide to users... <[Ting] C.TING> I think metacompiled Forth hurts the acceptance of Forth ... <[Ting] C.TING> to beginning users. <[Bill R] FIGGUEST> I think that either is ok if the nucleus is less that 12,000 bytes. <[Bill R] FIGGUEST> Beyond that point it should be meta compiled. This issue <[Bill R] FIGGUEST> has nothing to do with the acceptance of Forth, just the <[Bill R] FIGGUEST> effort the user expends for startup and maintenance. <[Bill R] FIGGUEST> People can learn the internals either way... <[Bill R] FIGGUEST> and this is great. I do think the Forth virtual <[Bill R] FIGGUEST> machine should be standardized. and the current <[Bill R] FIGGUEST> standards committees is on the wrong track... <[Bill R] FIGGUEST> They should be looking at absolute portability of <[Bill R] FIGGUEST> applictions, be it large or small. The richness <[Bill R] FIGGUEST> of the language can come later. <[Wil] W.BADEN1> Forth will survive. You have identified where and how -- dedicated applications with precious resources. Congratulations. <[FRANK] F.SERGEANT> What did you learn working with Charles Moore? <[Bill R] FIGGUEST> I only had the opportunity to work with Chuck for the <[Bill R] FIGGUEST> time I served on the Board of Directors <[Bill R] FIGGUEST> of Forth Inc. I had to learn From reading his code. <[Bill R] FIGGUEST> That was probably the best mentorship I ever had. <[Ting] C.TING> figForth was enormously successful. F83 and FPC much less so... <[Ting] C.TING> Can we repeat the history? <[Bill R] FIGGUEST> I think F83 ... <[Bill R] FIGGUEST> eclipesed figForth. It was more robust n <[Bill R] FIGGUEST> on more machines and had a better variety of <[Bill R] FIGGUEST> tools. MVP Forth comes in a close second, for in <[Bill R] FIGGUEST> its day it had great, distributed sponsorship. By that <[Bill R] FIGGUEST> I mean more sponsors that just the original vendor. <[Bill R] FIGGUEST> FigForth got off to a good start and still helps people. <[Bill R] FIGGUEST> get going, but we collectively designed it to be just a platform and <[Bill R] FIGGUEST> not to threaten any other's commercial efforts. Its time <[Bill R] FIGGUEST> came and went. We can't put the Genie back in the bottle.. <[Bill R] FIGGUEST> so other Forth language generations will come. I do <[Bill R] FIGGUEST> harp, though, on absolute portability at the level of the <[Bill R] FIGGUEST> lowest 100 to 150 words. Irv - you have last question <[IRV] I.MONTANEZ> BILL, WHAT ROUGHLY IS THE PERCENTAGE MAKE-UP OF YOUR FORTH <[IRV] I.MONTANEZ> LANGUAGE APPLICATIONS WORKLOAD, AND IN WHAT AREAS. <[Bill R] FIGGUEST> We use F83 for product development... <[Bill R] FIGGUEST> and for test equipement. A couple of production units, <[Bill R] FIGGUEST> label printers for example, are still in MVPForth because <[Bill R] FIGGUEST> of the computer they run on. We tried FPC but it was just <[Bill R] FIGGUEST> too big to 'get our arms around' and there were side <[Bill R] FIGGUEST> effects of changes that we were impatient to resolve. So <[Bill R] FIGGUEST> we stick with the safe and sane F83. In our business areas <[Bill R] FIGGUEST> we usea multiuser systesm in Pick.. closing comments please, Bill. <[Bill R] FIGGUEST> Thanks for the chance to chat this evening... <[Bill R] FIGGUEST> I offer this challenge to the Forth community, or at least <[Bill R] FIGGUEST> the one persone who might make it happen... <[Bill R] FIGGUEST> <[Bill R] FIGGUEST> How can we hold an interactive RoundTable like this, <[Bill R] FIGGUEST> programmed in Forth, by facsimilie? ... <[Bill R] FIGGUEST> I want to feed in hand written material and have it <[Bill R] FIGGUEST> distributed in realtime to the participants. <[Bill R] FIGGUEST> I think we can write quicker and with aids <[Bill R] FIGGUEST> like diagrams more effectively than we type. <[Bill R] FIGGUEST> So a fax roundtable is the next generation. Thanks <[Bill R] FIGGUEST> all for the fellowship we've enjoyed over the years. <[Bill R] FIGGUEST> Forth is **more** than a language. Bill Ragsdale, on behalf of the GEnie Forth RoundTable, ... 'Thank you' for a stimulating conference. ...

All may stay and talk, but this conference has officially ended. === End of Steno notes. ===