50 Years of BASIC, the Language That Made Computers Personal

harrymcc (1641347) writes "On May 1, 1964 at 4 a.m. in a computer room at Dartmouth University, the first programs written in BASIC ran on the university's brand-new time-sharing system. With these two innovations, John Kemeny and Thomas Kurtz didn't just make it easier to learn how to program a computer: They offered Dartmouth students a form of interactive, personal computing years before the invention of the PC. Over at TIME.com, I chronicle BASIC's first 50 years with a feature with thoughts from Kurtz, Microsoft's Paul Allen and many others."

Comments

Was FORTRAN really that hard?

[Bill, Shooter of Bul]

I mean Basic isn't difficult either, but I really don't understand the perspective at the time that FORTRAN was so complex that BASIC and COBOL were really needed for their syntax changes alone. All of the explinations I've read about them, invariably have the line somewhere about FORTRAN being so difficult to understand that only scientists could master it. I understand they were all invented for different problem domains and that's kind of a good reason in and of itself, but sheesh, its not like it was brain fudge.

Re:Was FORTRAN really that hard?

[Brett Buck]

It isn't hard. I think the only issue that anyone has had with it was the column restrictions, and the important development there was interactive computing and a decent text editor, not a new language.

    Bear in mind, also, that most people haven't ever attempted to write in original FORTRAN. Most have seen nothing earlier than FORTRAN 77 and that was tremendously easier and far less irritating than FORTRAN IV. I am old enough to have written FORTRAN IV on card decks, the "good old days" sucked for the most part.

    I always found BASIC to be far more irritating to program in anything more than trivial applications.

      I don't think the language is the problem, it was all other things that where difficult and mostly solved in the 60's and 70's.

      FORTRAN is certainly still in use for Real Programming. I haven't seen a way to use BASIC- in the way it was originally conceived - in 25 years.

         

Re:Was FORTRAN really that hard?

[flyingfsck]

No, it is more a case that access to any computing resources was very limited. Twenty years later, universities still had old time sharing machines running Fortran and BASIC that first year students trained on and COBOL, while PL1 on Sperry-Univac machines was the new thing for the seniors. A typical university only had two to four computers in total. I also started by learning BASIC with punch cards on a mainframe, since that was the only thing that was available. The small computer population explosion only really happened in the 1980s.

Re:Was FORTRAN really that hard?

[Jahta]

Ob. quote from Real Programmers.

"Real programmers don't write in FORTRAN. FORTRAN is for pipe stress freaks and crystallography weenies."

Of course, it also says this about BASIC :-)

"Real Programmers don't write in BASIC. Actually, no programmers write in BASIC, after the age of 12."

Re:Was FORTRAN really that hard?

[tangent]

FORTRAN wasn't the language in 1964 that you think of as FORTRAN today.

Most people's concept of FORTRAN is FORTRAN 77 or its descendants, which was 13 years in the future from BASIC's introduction.

At the time of BASIC's introduction, FORTRAN IV was the current version.

FORTRAN wouldn't be ANSI-fied for another two years as FORTRAN 66, so every version had machine-specific features. Also, because FORTRAN's development was largely driven by IBM until FORTRAN 66, all the non-IBM versions were "nonstandard." Imagine if, today, every computer came with a C compiler and there were no ANSI or ISO standard to constrain its behavior. The last common reference would have been K&R '78.

Another fun feature of early FORTRAN was fixed column layout, common among languages invented in the punched card era. That is, you had to do things like start all statements in column 7 or later, because the first 6 columns had other meaning.

Early FORTRANs also had very primitive program structuring concepts, hardly raised from the level of assembly language.

Read through the Wikipedia article. You'll probably be shocked at how primitive FORTRAN was in the early 1960s.

Re:Was FORTRAN really that hard?

[jellomizer]

Well there were a few key advantages.
BASIC was an interpreted language, vs FORTRAN or COBOL that required compiling. This gave beginner programmers quicker response in terms of code it run it. vs. Compile find a slew of cryptic errors try to find them and fix and try again.
Secondly BASIC was a line driven language vs. procedural.
10 print "hello"
20 goto 10
yes it created a lot of bad programming habits. But it really explained how the computer processed the stuff better then the procedure. But I am sure some of you taking CS 101 (Those without previous programming experience) often got hung up on function and procedures, loops and nested loops. The BASIC GOTO and if you get advanced use the GOSUB and RETURN makes these concepts easier.

Third BASIC handles free formatting much easier. FORTRAN you needed to follow a Particular set of formatting the first few spaces were reserved for labels or line wrap numbers, limit to 80 columns... In general nitpicking stuff that a started just doesn't want to handle.

As time went on, we really loss the simple computer programming language. BASIC turned to QBASIC then to Visual Basic then to VB.NET which made it a formatted language with all the hold ups the other professional languages has.
There isn't really a good starting language anymore.

Re:Was FORTRAN really that hard?

[MightyYar]

It's not so much that FORTRAN was horrible, but AFAIK it did not lend itself to a simple interpreter and so the gratification was not instant. Compiling is an extra step and it is not a fun step. In addition (again relying on very faint memory), BASIC had easier string-handling functions built in until FORTRAN77 - and most of us were dealing with strings when we were learning to program.

Re:Was FORTRAN really that hard?

[RabidReindeer]

FORTRAN isn't that complex. Originally, it only had 7 statements (or was it 10)? The only halfway complex thing about it was the expression compiler parts.

But BASIC had several advantages. It was intended for interactive use, at a time when most FORTRANs were batch-only. It originally only supported a very limited set of variable names like "A", "B", "C", and so forth, meaning that you didn't have to implement symbol table logic, and the associated storages on machines when 16K RAM space might be considered at lot - instead the variable name was the hash code to a value table. Refinements such as the rich set of built-in functions and extensive string services were also later additions.

The original BASIC was so minimalist that even the first effort from Gates &Co. exceeded it. But it introduced a lot of people to "instant gratification" programming, and thus its influence can be felt in many places to this day.

Re:Was FORTRAN really that hard?

[AdamColley]

Real programmers toggle in the boot code on the front panel.

Re:Was FORTRAN really that hard?

[mwvdlee]

I stopped programming COBOL about 5 years ago, after having professionally used COBOL for about 13 years.
I've never even encountered a COBOL interpreter; it was all compiled.

IMHO, interpreted COBOL makes no sense at all; you use COBOL because you want raw performance.
If performance isn't top priority, you'd be better off using Java, C++ or most other languages.
Perhaps if you need backwards compatibility for obsoleted hardware running legacy code.

Re:Was FORTRAN really that hard?

[hey!]

I've actually programmed in Fortran and BASIC way back in the day (late 70s early 80s). From a language point of view early dialects of Fortran (e.g. Fortran IV ca. 1961 and still in widespread use in the 70s) and BASIC were in fact *very* similar. What was different was that Fortran was *compiled* and BASIC was *interpreted*.

It was common until the mid 1970s for Fortran programmers to physically drop off a deck of punched cards at the operator's window. They'd get their results some hours later, if not the next day, after the operators got around to running the job. Most of the time those results wouldn't the desired computation, but a compilation error. So to be productive in Fortran you had to think about your *entire* problem in advance, carefully preparing your deck to get as much as possible correct before handing the job off.

BASIC was an interpreted language initially. That meant you to type in little snippets of your program, even individual expressions, to see how or if they did what was expected. If you typed in a program and there was a syntax error, you'd know as soon as you hit "return". This allowed a more exploratory approach to programming and learning to program. Of course, you could get the *same* interactive experience in a much more sophisticated language by using Lisp.

I started programming in C in the 1980s, and this use-style distinction between compilation and interpretation remained. A full compile and link of our modest application took something like 30 minutes on the minicomputer we were using, which had a clock speed in the single digit MHz range. So we prepared our source changes *very* carefully, and used our down time to read the full 8 volume Unix manual from cover to cover, over and over again. There was something to be said for such an approach to programming, but it was not for the faint hearted.

By the 90s this had changed. Compilers were orders of magnitude faster; you'd actually hit "compile" and *watch* the compiler do its thing. A decade earlier that would have been like watching paint dry. Editors became syntax-aware too, and debuggers went from requiring mad voodoo skills to being interactive and usable by ordinary mortals. So now compilation vs. interpretation is a packaging and runtime issue; there's not much to choose between them with respect to how *hard* a language is to use. Naturally someone who cut their teeth in the modern era look at BASIC and Fortran as they were in the 60s and wonders what the big deal was. But it *was* a big deal, at least for people who weren't going to learn Lisp.

Thank you Kemeny and Kurtz.

[DiscountBorg(TM)]

I grew up with a little TRS-80 on which you had to learn BASIC to so much as load a file. In Grade Three I was learning things like coordinate geometry and algebra, while my peers were struggling with their multiplication tables. I remember when my peers were introduced to algebra for the first time, some of them had difficulty understanding how x could be a number, while I was busy making adventure games at home.
Thanks to this head start in life, I now have a job in IT. BASIC gave me a great head start in computer literacy!

Re:Thank you Kemeny and Kurtz.

[DiscountBorg(TM)]

I wouldn't contend for an instant that the kids I grew up around were 'retards'. 8-year-olds can't magically know things without experience.

How many kids have the chance to sit down in front of a computer and learn that the reason a ball goes across the screen comes down to something as simple as x=x+1? Schools won't teach them that until the end of primary.

BASIC does probably teach some bad programming habits but at the same time it's accessible to an 8-year-old, and you're learning concepts that are applicable for life: file management, how to store and retrieve data, syntax, etc etc. If the goal is to introduce kids to ehmm.. basic computing concepts, it worked admirably.

Compare to someone with no knowledge of programming concepts at all whatsoever trying to grasp how to call a function for the first time in their life.

Re:Thank you Kemeny and Kurtz.

[bluefoxlucid]

It is practically impossible to teach good programming to students that have had a prior exposure to BASIC: as potential programmers they are mentally mutilated beyond hope of regeneration.

Re:Thank you Kemeny and Kurtz.

[DiscountBorg(TM)]

Yes, and then from there it's really easy to understand how to calculate things like velocity and gravity, and understand vectors. If I ever have a kid, I'd totally want to get them learning some form of basic coding at an early age. Nowadays it doesn't have to be BASIC, it could just as easily be LUA There's so many useful/abstract concepts you pick up naturally while figuring out these things even outside of basic computing.

Re:Thank you Kemeny and Kurtz.

[DiscountBorg(TM)]

I agree, but this is actually an old tongue in cheek essay, in context it makes more sense perhaps:

"FORTRAN --"the infantile disorder"--, by now nearly 20 years old, is hopelessly inadequate for whatever computer application you have in mind today: it is now too clumsy, too risky, and too expensive to use.
PL/I --"the fatal disease"-- belongs more to the problem set than to the solution set.
It is practically impossible to teach good programming to students that have had a prior exposure to BASIC: as potential programmers they are mentally mutilated beyond hope of regeneration.
The use of COBOL cripples the mind; its teaching should, therefore, be regarded as a criminal offence.
APL is a mistake, carried through to perfection. It is the language of the future for the programming techniques of the past: it creates a new generation of coding bums."

http://www.cs.virginia.edu/~ev...

Re:Thank you Kemeny and Kurtz.

[jeffb (2.718)]

It's different when you cover algebra before you hit programming.

Algebra teacher: "X + 1 = 2X. Subtract X from both sides. X = 1."

Me: "OK"

BASIC program: "X = X + 1"

Me: "No, it isn't!"

Me, years later, in a class using Prolog: "X = X + 1"

Prolog interpreter: "No, it isn't!"

Re:Thank you Kemeny and Kurtz.

[MightyMartian]

I learned to program on 8 bit computers (Tandy and Commodore), all of which were running BASIC interpreters, and I had no problem moving on structured programming like Pascal, and ultimately OOP languages like Java.

Re:Thank you Kemeny and Kurtz.

[narcc]

This baffles me. If you compare BASIC to any assembly, you'll find that they're not that different, as far as how you structure your programs.

Does assembly teach "bad programming habits"?

I'm going to call this particular meme "nonsense" until such time as someone can offer up a decent argument.

 

Re:Thank you Kemeny and Kurtz.

[z80man]

Every processor I've ever used has a "JUMP" op-code, synonomous to the "GOTO" statement. If its good enough for the processors machine code, its good enough for me :-)

Re:Thank you Kemeny and Kurtz.

[narcc]

GOTO isn't inherently evil. I'd go as far to say that the GOTO hate is little more than religious propaganda.

It's used in the Linux kernel. There's even a famous post by Linus defending it. He's not alone. Do a quick google search. The results may surprise you.

Happy Birthday

[Captain Emerald]

10 PRINT "Happy Birthday, Basic"
20 GOTO 10

Re:Happy Birthday

[mod prime]

This is not bad programming, this is the most common Use Case.
When I was programming BASIC this was equivalent to

while Break not pressed:
  print 'whatever'
exit program

I liked BBC Basic. And Q(uick)Basic.

[serviscope_minor]

The BBC Model B equipped with BBC BASIC was released in 1981. As well as the usual litany of BASIC like features (i.e. goto), it had proper named procedures and functions with local variables, which allowed structured programming. It didn't have proper block structured if though.

It also had dynamic memory allocation and pointer indirection (not that wretched peek and poke stuff).

It was still tied to line numbers though, but in practive you (a) didn't need them except for computed goto and jump tables and (b) it had a proper renumbering command if you needed to insert space which corrected all the gotos, gosubs and jump tables (but not obviously computed goto).

It had a 5 byte floating point system build in too, which while slow was pretty decent.

Was quite powerful. It also had graphics and sound built in, which made it very nice to play with.

And then I graduated on to QB when I switched to a PC. Mostly QBasic then a pirated version of QuickBasic. Actually my dad was very against piracy but relented when we phoned a Microsoft sales office and they denied all knowledge of such a product and tried to hawk us an early version of Visual Basic.

QBasic was a fantastic system, especially given it was free with PCs, and I challenge anyone to claim otherwise with good justification.

Re:I liked BBC Basic. And Q(uick)Basic.

[Jahta]

The BBC Model B equipped with BBC BASIC was released in 1981. As well as the usual litany of BASIC like features (i.e. goto), it had proper named procedures and functions with local variables, which allowed structured programming. It didn't have proper block structured if though.

Yes indeed. I initially learned to program on a BBC, and I learned a number of good habits in the process.

Re:I liked BBC Basic. And Q(uick)Basic.

[AdamColley]

You both forgot to mention the BBC had a built in 6502 multipass assembler.

So unlike my spectrum I didn't have to reload the assembler every time I made an error and my code stomped all over ram.

Additionally, you had direct access to OS routines from basic, OSBYTE, OSWORD, OSCLI etc.

The BBC was and still is far ahead of anything else as a teaching machine. Simple enough to understand, complex enough to be useful and enough I/O to put a pi with gertboard to shame even today.

Re:I liked BBC Basic. And Q(uick)Basic.

[serviscope_minor]

You're absoloutely correct. It did have those and it was awesome. It meant you could freely mix ASM and BASIC code. Handy so one could do inner loops in ASM and the more complex outer logic in basic without sacraficing anything. I seem to remember that the special variables A%, X% and Y% would be used to set the A, X and Y registers on entry and be set back on exit for easy communication.

Re:I liked BBC Basic. And Q(uick)Basic.

[Jahta]

You both forgot to mention the BBC had a built in 6502 multipass assembler.

So unlike my spectrum I didn't have to reload the assembler every time I made an error and my code stomped all over ram.

Additionally, you had direct access to OS routines from basic, OSBYTE, OSWORD, OSCLI etc.

The BBC was and still is far ahead of anything else as a teaching machine. Simple enough to understand, complex enough to be useful and enough I/O to put a pi with gertboard to shame even today.

Good point. The built-in assembler was excellent too. The whole BBC Micro project was designed educate people about the computer as a powerful tool they could use, and not just a games machine. And, as you say, they did a damn good job.

DND in BASIC

[null etc.]

Richard Garriott (of Ultima fame) is running an interesting challenge to port his very first RPG computer game, written in BASIC on a teletype connected to a PDP-11, into a web-friendly or Unity version. https://www.shroudoftheavatar....

Re:In 3, 2, 1...

[NapalmV]

The truth about C:

https://www.gnu.org/fun/jokes/...

Now seriously. Pascal was published some 2 years before Kernighan and Ritchie released their masterpiece. Having the opportunity to have a long look at Pascal and yet coming up with something like C shows a very strong character.

No, it wasn't.

[Baldrson]

If IBM had gone to Chuck Moore instead of Bill Gates (or rather, his mom) for their 4.77MHz 8088 PC, your title might have been "50 years of FORTH, the Language That Made Computers Personal".

But, then again, if IBM had done that, the personal computer era might have been bypassed entirely with the network computer launching the equivalent of the WWW in 1983.

Re:No, it wasn't.

[Baldrson]

No doubt BASIC was the path of least resistance, but if you click through my link to 1983, you'll notice that Smalltalk was positioned to execute on a FORTH stack VM which was then reduced to hardware in the Novix chop. Moreover, the technology you see in the Javascript V8 engine had already been published in relation to Smalltalk in 1983.

This was technically feasible at that time. The fact that it wasn't the path of least resistance hardly qualifies BASIC for the credit accorded it by the title of this article.

Vic-20 Tank vs UFO

[NeoNormal]

I bought a VIc-20 in 1982 to use in my woodworking business. I learned BASIC on it by trying to key in the Tank vs UFO game that was printed in the manual. I don't know if it was all of my typos or errors in the printed listing (both likely), but through debugging that ASCII character game, I got started in the direction that took me to working in IT.

Re:In 3, 2, 1...

[AlecC]

No, it just shows habit. C was descended from B, which descended from BCPL. They just did more of the same, instead of going to someone else's syntax.

And, having programmed in Pascal for 15 years. Pascal as defined was not suitable for large projects, whereas C was. Every Pascal compiler had to have some non-standard add-ons to handle modularity. And they were all different. Obviously, the Borland model came to have the status of a de-facto standard, but that was not till some years later. You could not have written Unix in standard Pascal; it was written in standard C. Wirth acknowledged the modularity failings of Pascal in his Modula language family, but by that time he had missed the bus.

LISP instead!

[Anonymous Coward]

Lisp was invented in 1958. Can you imagine a world were personal computers had Lisp instead of BASIC? We would have had the singularity the year after IBM released the AT!

Re:LISP instead!

[istartedi]

I think it's more likely that only a few people would have cared for it. Of those few, an even smaller few would have written something like a BASIC interpreter in Lisp and sold it separately.

Either that, or they would have sold poorly until somebody came out with what we got, which was PCs that shipped with BASIC.

I recall reading a review of one of the less popular systems at the time. The reviewer said something like, "it comes only with assembler, which is useful only for understanding how a computer works". Of course I and a lot of other people ended up programming the C-64 mostly with assembler; but we started with BASIC.

Oh, and I'm surprised I got this far down and nobody quoted Djikstra yet. I like to count myself as one of the many programmers who proves him wrong; although it's my understanding that the BASIC to which he referred was inferior even to the line-numbered versions of the 80s. I wonder if he ever qualified or backed up even a little bit from his infamous quote.

Re:LISP instead!

[Opportunist]

(I ((am not) so (sure (about that))), (you (may (rest assured) (that (at (some point)) (someone (would have (forgotten (a closing (or opening) bracket)) ... ) ... )? ...

oh fuck, let's start counting... one open, two open, three open, two open...

Re:In 3, 2, 1...

[gaspyy]

I haven't coded in Pascal since the good old DOS days, when I was about 17, but at the time was writing TSR apps, picture viewers (GIF, PCX, BMP, TIF), some graphic-mode UI, including mouse support, even some VGA graphic demos. I can't think of anything that I couldn't do in Pascal (and some ASM, I give you that). In fact, it's the reason I never got too heavily into C...

Basic Got Me My Career

[coolmoose25]

I started working on computers in the early 80's... The first one I used was a TI 99 4a. It had tape drives and a TV set as a monitor, and a horrific keypad (note: not keyboard). Then my brother got a PC Jr. and I started hacking with that and then went off to college. As an engineering major, I learned FORTRAN on punched cards. I hated it! Swore I'd never have a job where I used computers.

Then my brother got the family to chip in and buy me a Tandy 1000a. It came with DOS, Deskmate, and Basic. I started programming in Basic using the concepts I had learned in FORTRAN. By the end, I think I had dumped about $5,000 into that computer. Printers, memory upgrades, floppy upgrades, hard drive, monitor, etc. And still was able to do amazing things with Basic and with BAT files.

My first job was with Arthur Andersen. COBOL. Batch COBOL. 2.5 years of it. Learned it in 6 weeks, and spent the rest of my career there either coding it or writing tech specs for it.

Went to work at an insurance company coding SqlWindows, a now obscure 4th gen programming language. But hey, it was Windows programming. Spent 10 years there in a variety of roles.

After that I set up my own web development shop... Wrote classic ASP which is essentially Basic for the web. And then went to work at another insurance company, writing, you guessed it, Microsoft VB.net. Granted, VB.net was a far cry from the original basic, and probably would have been better off learning C#. But that was Microsoft's strategy with .Net - recycle old VB programmers and old C programmers using the CLR. At the end of the day, not much difference between C# and VB.net. Now I don't code anymore, I'm a VP at that insurance company. But I owe a lot of my career for having a tool like Basic available to me in my formative years. Sure, it teaches you some bad coding habits. But just like anything else, you learn from that, and others, and classes (and objects for those who like puns). Those who say that you can't be a good programmer after having learned basic are either elitist snobs or idiots. Sometimes you have to do it wrong first to see how doing it right makes all the difference. So Happy Birthday Basic - I love ya' baby.

Re:In 3, 2, 1...

[WillAdams]

``Pascal as defined was not suitable for large projects...''

Unless of course, one is Dr. Donald Knuth, then one creates a brand new programming paradigm: http://www.literateprogramming...

and writes programs such as TeX: http://www.ctan.org/tex-archiv...

Somewhere, I have a copy of the Oberon language manual printed out --- it's quite cool, and very concise.

Re:Looks like a duplicated thread

[Opportunist]

That's not really a good idea, Basic is hardly thread-safe!

Re:In 3, 2, 1...

[AlecC]

I do not deny that Knuth, and Wirth, created other, very cool, programming languages later. But I stand by my statement that Pascal, as originally defined, was not suitable for large projects, a failing that Wirth himself recognised.

Idiots

[Anonymous Coward]

So, if you are surrounded by idiots as early as middle school, you'll get better grades.

When I was a kid, I wanted to be a machinist. I love working with my hands.

In Middle School and High School, there were these "idiots" who took shop barely passed Algebra and took jobs that gave them credit to graduate.

My parents didn't want me to be a blue collar worker and demanded I go to college. Part of it was that they wanted something more for me - blue collar jobs were being sent down South (Carolinas, GA, FL, etc ..) at the time and the "college boys" had their cushy salaried jobs and were the ones laying people off and sending jobs to the South - "those Southerners took our jobs!" (The good ole days before Globalization).

Years later, I was patting myself on the back for making 6 figures when I bumped into an "idiot" I went to school with. Well, I met him for lunch at his $5 million tool and die company - and I let my parents know about it.

Well, today he's lost a bit of business because of off-shoring but is still doing well and he's still respected for being a business owner.

I'm unemployable with savings dried up, investments gone, and people telling me that I'm no good and stupid - Maybe so.

Of course today, being a machinist is pretty much "monkey pushes the button" for these high tech CNC machines - the designers write the programs.

I digress.

tl;dr: "Stupid people" have their place and don't be surprised if they are more successful than you.

PS, think about what you are doing in IT/Software development. What do you really offer society and humanity?

Looking at the "apps" and "technology" coming out of Silicon Valley, I have to say, they offer no value. The just contribute to our mindless consumer society.

Google, Yahoo!, Apple, etc ... are just consumer products and services that exist for us to ... consume. No value.

My corner mechanic offers more to society than all of you in Silicon Valley. Same goes for the nurse at my hospital. We in software and IT like to think we offer so much but really, what do we do? Or what does you job do? All of you working at Facebook are a waste. Same goes for you Google "engineers" - you are nothing but marketing people.

Just a waste.

The best part about Basic

[mark-t]

Basic was my first programming language, and I actually spent almost 10 years using it before moving onto more structured languages like C, but it wasn't too long after I learned Basic that I found that my favorite features of the language were the ones that enabled me to extend it with my own customizations, which I would have to write in assembler. If I remember correctly, the relevant basic keywords in the implementation that I used were 'usr' and '&... practically turning it into another language with all of the extensions that I would throw in.

Dartmouth College

[bill_mcgonigle]

editors:

s/University/College/

(hoping that wasn't an editorial 'correction' since TIME got it right)

Also, there's a party all day on campus tomorrow.

Re:use COBOL because you want raw performance

[mwvdlee]

Then you've probably never even touched COBOL in your live.

All shops I've worked for are actually pretty hardcore about performance, testing against other languages whenever a compiler is updated.
Nobody wants to program COBOL code, it's just that it's a produces very good binaries for the specific (financial) tasks it was designed for.
Or do you really think all those COBOL programmers still active are only doing it out of legacy support?

Re:"Technically Feasible"

[Baldrson]

The reason I promoted Forth as a graphics communications protocol alternative to NAPLPS is the Western Electric Videotex terminals for the Viewtron service were so limited in RAM and ROM (far more limited than the first 4.77MHz IBM PC) that it needed a highly compressed representation of the firmware for decent graphics performance. Forth provided that and it would have further allowed dynamically downloading tokenized Forth stack functions (called 'words'). I talked about this with the guys at PARC late 1982 (prior to PostScript spinning off). Forth was lousy as a high level language but great as a macro-assembler for a hardware Forth chip -- a macro-assembler in which implementing the Smalltalk -- or Simula -- as an OS user interface would have had a very small memory footprint compared to a full-blown Smalltalk environment. Performance wise, the NOVIX chip was 16,000 transistors and was a full 16-bit wide FORTH hardware machine running at 24MIPS whereas the 8088 was 29,000 transistors but had an 8 bit bus running at about 2-3MIPS. Both were 3um technology. This initial implementation wouldn't have had enough memory to allow the full optimizations, such as type inference w/JIT etc that were then known possible and are now realized in the V8 JS interpreter, but it would have been at least as fast as BASIC as well as providing a clear migration path to vastly superior software development. As for price, well, Tesler was at PARC and had published on type inference -- and Jobs got him away. There's no particular reason something like small Simula core couldn't have been quickly implemented atop the Forth machine and Chuck Moore's low level drivers etc. In any event, IBM could have gone to Xerox and offered to buy PARC from them and Xerox would probably have sold it cheap. Xerox undervalued the work at PARC.