Slashdot Mirror
What Is the Oldest Code Written Still Running?
Consul writes "What is the oldest piece of code that is still in use today, that has not actually been retyped or reimplemented in some way? By 'piece of code,' I'm of course referring to a complete algorithm, and not just a single line." The question would have a different answer if emulation, in multiple layers, is allowed.
Interesting, a quick search on Google reveals that there isn't much on this topic other than people talking about the oldest computer they have. One post talks about some old IBM Series 1's and S/360/30. One good one is to say the computers onboard some of the oldest spacecrafts like Pioneer 10 (1972), Voyager I and II (1977). Although they haven't received anything from Pioneer 10 since 2002. But you could say that the computer in it might still be running.
Somehow I doubt that many of the people that would be running such old computers such as ones from before 1970 would be reading Slashdot. And if you think about it, people conceptulized computers differently back then. I think you'd be hard pressed to find mention of a specific program but more of mention of a computer itself. Its too bad there is such a big disconnect between the generations of computer programmers and administrators.
There's still code running for nuclear power plants that was written in the 60's or earlier; given the challenge of certifying emulators we ran it on the original machines; embedded code in machinery was probably been older. Although, most really old stuff was mechanical not based on ICs.
Some military hardware may be even older; reliability and certainty is often more important than the latest and greatest.
I'm a consultant - I convert gibberish into cash-flow.
I work in a big bank in France and i know we had, a few years ago, some code from 70's in Cobol for some core products.
Well, humans have basically been running the same program for some million years... with some minor software upgrades of course.
I wonder if there are any Jacquard looms still running.
I find it amusing that most conspiracy theorists - whether the conspiracies are true or not is immaterial - tend to write long rambling screeds like that that cause people to lose interest after the first sentence, and then use that as proof that the world is against them.
It's all about the packaging.
For linux tips: http://www.linuxtipsblog.com
...because even if the code hasn't been replaced, you can bet the source control software has. My guess would be old cores of for example banking systems, I know there our company has COBOL code written in the 60s and the system is still in COBOL and in use today. If someone wrote a correct, useful algorithm back then it could very easily still exist today. I can at least assure you that they don't exactly do rewrites very often...
Live today, because you never know what tomorrow brings
IEFBR14, the good old chunk of do-nothing code, the most universal executable used by anyone who ever wrote JCL.
It really does that - nothing. IEF is the code prefix, since all code *must* be prefixed, after all. BR14 stands for "Branch to Register 14", which with the old code linkages conventions means "return and exit". In JCL it's commonly used simply to attach, allocate, and deallocate files. In other words, used for its side-effects with the file allocation parameters. I haven't written any JCL in probably 20+ years, or I'd give an example. Anything I'd show now would likely be too badly riddled with errors to give the true, scrumptious feel.
The living have better things to do than to continue hating the dead.
FANG, from 1972, is probably one of the oldest applications you can still download and run. It's a copying utility for UNIVAC mainframes. UNIVAC Exec 8 was way ahead of its time, with full support for threads, multiprocessors, and concurrent I/O from the late 1960s. FANG was one of the first applications to use that concurrency effectively. You could put in a series of commands to operate on multiple files, and it would do them as concurrently as possible, keeping track of any dependencies in the file copies.
Ada Bryon's Notes on the analytical engine contains the oldest running code today. It can be run here.
Of course Charles Babbage holds the claim for longest vaporware project at 153 years. And also apparently the longest unfixed bug.
We programmed landleveling program on an Apple II+ in 1979. That code has remain pretty much unchanged since a port to GW basic for the PC. That is the oldest code we have written that is still being used. Before that we used a strip programmable HP-Calculator (computer?) to run the numbers.
By which I mean production code, not the 'Hello World!' you did in Jr. High. I'll go first. In the mid 90's I wrote a COBOL program to link a mainframe to a HP printer to print transcripts at a uni. The SYSPROG set up the VTAM lines and I glued the PCL together with COBOL. I checked in about 3 years ago and a friend of mine said they were still running it. So at that time it was pushing 10 years. Which makes me proud actually.
Anyone else with a story?
putting the 'B' in LGBTQ+
You're welcome. The question just hit me all of a sudden, and I thought it would be a fun mental exercise. I'm actually quite amazed it got accepted. ;-)
-----
"You spilled my egg... I needed that egg."
It came about because the developers of the IBM 360 Operating System suddenly realize a few weeks before it was ro be released that they had no method of actually allocating resources for a job. In a panic, they hacked together a version of their assembler macro language to parse the control statements. so the format of the language was the same as assembler language macro calls
label opcode operands
Spaces were significant, everything had to be upper case, syntax was arcane.The add font dialog is probably the oldest piece of Windows code still running in Windows... it's from Windows 3.1, and still looks the same in Vista!
What time is it/will be over there? Check with my iPhone app!
Is the difference engine actually programmable? Did Babbage actually build a working one? I don't think either is true, but I could be mistaken.
However, those questions lead to other subtly different interpretations of the original question. Is a code/program written on paper enough to qualify, and does it have to be "loadable" software, as opposed to some fixed function device?
If actual implementation counts, then ENIAC, the Harvard Mark I and Konrad Zuse's Z3 (and whatever programs are still available for them) would be the candidates.I can come up with. There's a copy of the Z3 that's ocassionally fired up.
How about a 40 thousand year old year old shrubbery...
It's the oldest living organism, so it's got the oldest bit of unchanged genetic code, and obviously a lot older than computer code for sure.
If I have nothing to hide, you have no reason to search me
It's amazing to me that NASA has the foresight to design such a remote update system years before the concept of a "firmware update" was ever applied to consumer technology. The innovations that have come out of NASA's labs is vastly underappreciated -- one wonders where our technology would be today if we invested more in the space program and less in killing one another (that is _not_ a condemnation of any particular country, pointing fingers doesn't solve problems...if anyone is offended by that remark I apologize).
Of course, it depends on what you count as code and what you count as running.
People have already mentioned DNA, and I guess I'd give that high marks. But maybe we mean things invented by man.
An abacus is a hardware program that is programmable with data and will yield numeric results. So is a sliderule. And there are others like the card sorters for punch cards, which predate programmable computers by several decades and yet performed very useful computation long before general purpose computers. And there are analog computers for predicting the motions of planets or for controlling the locks of the Panama Canal. But maybe we meant code implemented in software.
The Babbage Machine is mechanical so if it stops, does that mean the machine has crashed or does it just have a long cycle time? People have mentioned that, and that's certainly a worthwhile contender.
Mathematics also codes up algorithms, some of which are extremely old, and some of which you might regard as code, and so there might be something there that's competitive. But in a forum like this, full of nerds, I think "math" is too easy an answer and isn't provocative enough to get people thinking, so I'll go with this one:
My personal favorite is just something done in human language. Human language has codified the execution structure of organizations and processes for quite a long time. The US Constitution defines an engine that runs the United States, for example. Roberts Rules of Order is a program that is an interrupt-driven system that runs meetings. Contract law in the US (and perhaps world-wide) reminds me a lot of the structure of bootstrapping TCP (reliable transport of packets under a contract) from unreliable pieces (the contract terms and offers); the whole business of how you can send an offer and what constitutes acceptance in the face of data loss and things arriving in the wrong order is very much analogous to what you see in modern networking systems, but just used to work via pony express instead. So I'd put my vote on one of those. I just don't have the time to work out the timelines to figure out which one came first... probably something in English Common Law. It also depends on whether you want a "framework" or a "packaged application" or whatever, because some of these I've mentioned are in different categories in that regard. These may not be quite as old as some mathematical algorithms, but I bet they're more overlooked.
Now that I think of it, though, I bet food recipes (which are algorithmic in nature) predate even the earliest work of mathematicians, and it wouldn't surprise me if the recipe for making hot tea is the oldest, even if it's been upgraded a few times for changes in available hardware.
Kent M Pitman
Philosopher, Technologist, Writer
1. The US air traffic control system is 1960s vintage and I'd bet that there's still code in it that is unchanged since it was written.
2. Some airline reservation systems are of equally antique origins. Although I'm sure the hardware has been updated in the ensuing years, I'd say there's probably a lot of code that hasn't been rewritten. Back in the '80s when I was doing some work with an airline and asked about that, I was told, "That code is older than you are."
3. Don't know if this is still the case, but back in the late '70s, Navy carriers had computers so old that they were having to scrounge up germanium transistors to keep them operating. They wanted to keep them operating because nobody wanted to pay to rewrite the gazillion lines of reliable and tested assembly-language code that ran on them. If any of those are still around, they'd be my top candidate for having unchanged code still in operation. I'd guess that, in general, military systems (of the non-COTS [commercial off-the-shelf] type) are the most likely "oldest code" candidates, because of the lengthy and expensive qualification process and the long service life of such systems.
DNA.
If you can read this, I forgot to post anonymously.
I was looking for some mathematical routines to port into Python and ended up poking around at http://www.netlib.org/ and http://www.nist.gov/ where there are huge repositories of mathematical functions, most written in Fortran.
One of the most interesting things after perusing much of the code I was looking for, was that instead of using integration routines for calculating things like Bessel functions, Hankel functions, and other differential equation related functions, they simply used look up tables and curve fitting.
I suppose in the 1960's that made perfect sense as computers were so slow. But even today, I don't know why I shouldn't do the same thing. With EM and circuit simulation software its GIGO. There are so many parasitics to model, that you can only ever get an approximation anyway, so what difference does it make if you get a tiny error from a look up table, vs. the "exact" integration routine value?
I saved this post from alt.folklore.computers. Terribly impressive. I'm
not sure his age estimate is necessarily accurate -- the final
incarnation of the Leo ceased to be manufactured in the later half of the
60s.
I don't know if some modern incarnation of the Orange Leo made it past Y2k. If it did, my guess is it will still be around for a long time...
From: Deryk Barker
Subject: Re: Multics
Newsgroups: alt.folklore.computers, alt.os.multics
Date: 1998/11/09
[*snip*]
When my wife was working for Honeywell, in the 1980s, one of the
customers she had dealings with was British Telecom.
BT, at one location, had what they called the "orange Leos".
Now, for those who don't know this, the LEO was the world's first-ever
commercially-oriented machine (1951). Even more amazingly, the Lyons
Electronic Office was designed and built by the J Lyons company,
best-known as manufacturers of cakes and for their nationwide chain of
corner tea shops.
Anyway, an "orange Leo" was an ICL 2900 mainframe (they came in orange
cabinets), emulating an ICL 1900 mainframe, emulating a GEC System 4
mainframe emulating a LEO.
30+ year old executable code over 3 architecture changes....
My uncle used to run embroidery machines in union New Jersey. These were built in the late 1800 and were about 100 feet long 10 feet wide and 2 stories tall with 1000's of needles stitching constantly. Literally were built as part of the building they were housed in.
Where it gets interesting is these were driven by a large mechanical computer that ran from paper punch cards. The device itself was about a 1 meter cube. There were adders, and carry, multiples, and I think even branches and loops. It used to move paper cards back and forth as it created post man patches or frillies part of ladies undergarments.
Don't know if this counts though and I think it's decommissioned anyhow, but it was sure was cool to watch.
I am always doing that which I can not do, in order that I may learn how to do it. - Pablo Picasso
I'm thinking maybe the air-traffic control systems? Aren't there frequent complaints that they are outdated? Or perhaps some kind of defense system, like NORAD, etc.
Pretend I said something meaningful or insightful here.
Of course, one of the main reasons that NASA exists in the first place is to show our military might. If we could send people to the moon it is obvious that we have accurate missiles that can make it across the world and still hit their target. NASA was the "peaceful" way of showing our military power and technological innovations, and it served its purpose quite well.
The problem today is that without a cold war, NASA doesn't have as much of a purpose. It's still around, and still doing neat things, but it isn't where innovation is being pushed as far.
Phil
Most likely not the oldest code running, but how about, perhaps, the oldest widely distributed code, and still being distributed, running code.
Squeak is not just "a" Smalltalk implementation, but it is also, in many ways, "the" Smalltalk implementation.
I may be completely wrong about this, but it makes a good story nonetheless, so I'll make my case.
Back in the late 70's, the group at Xerox PARC were working on what would become Smalltalk-80. An interesting artifact of the Smalltalk implementation is that rather than being developed as individual facets of source code that are then compiled in to a final distribution, like a C program for example, Smalltalk is distributed and developed by directly making changes in to a running image. At times, that image is replicated and copied to other users.
Smalltalk can certainly be, and is, distributed as solely source code. But even today I do not believe there is a Smalltalk system that can be built from the ground up purely from source code. They typically rely on an instance of an image of Smalltalk to start with.
At the minimum, this is how the PARC Smalltalk-80 is developed.
Since Smalltalk was written on top of a VM architecture, in order to port Smalltalk from one machine to another, you only have to port the VM itself, then an existing Smalltalk image would run on top of that VM. So, the easiest way to port Smalltalk was to start with an existing Smalltalk image, and a VM spec.
In essence, the VM spec is exactly what the famous "Blue Book" was. It was the documentation of the Smalltalk VM.
While PARC released Smalltalk-80 on to the world in 1980, via the August issue of Byte Magazine, they also managed to work with four different companies who were also interested in the Smalltalk technology. One of those companies, as many may well know, was Apple Computer.
In order to facilitate the bootstrapping effort, all of the companies were given an existing Smalltalk-80 image which they would use on their own internally developed Smalltalk VM. At this point, this Smalltalk image could well be considered to be a copy of the original Smalltalk image that PARC itself was using for internal development. Xerox's master, base Smalltalk image.
I personally saw this image running in 1985, when an Apple employee demonstrated it to me at a Mac show being held on our university campus. It was a very enlightening experience. And for anyone who has read the "Orange Book", this image showed the exact same system as documented in that book, with it's browsers, workspaces, inspectors, scrollbars and pop up menus.
Fast forward to the mid-90's, when Squeak was announced to the world.
Squeak is a portable Smalltalk runtime and image, written within itself. It is self hosting and a "full blown" Smalltalk environment. The Smalltalk used to create the VM is a specific subset of Smalltalk that can be compiled down to C, linked with some system primitive functions, and compiled to be hosted on any of several systems. Once the VM is ported, the image "just starts". Squeak is widely ported.
Squeak is a direct descendant of the Apple Smalltalk effort from the early 80's. In fact, it's being build by the same folks who worked on the original Smalltalk at PARC. The early goal of the project was simply to get a portable VM to run the standard Smalltalk image. And by standard Smalltalk image, I mean Apple's Smalltalk image, which is a direct copy of the the image used by Xerox in Smalltalks infancy.
What I assert, though I cannot prove, is that there is code, and perhaps even large chunks of code, within the modern Squeak image, that was placed there by the original authors pushing 30 years ago. I argue that the code is still there, and that it could be code that worked so well, there has not been any call to change it in all these years.
Now, it's fair to assume that the actualy byte codes may not be the same (I imagine they're not, minimally at an object pointer level), but whatever byte codes we have
For that matter, how often does it need to run in order to be "still running"?
If you run the oldest piece of hardware with the earliest software ever written once or twice per decade for historical reasons, is that code "still running"?
Slashdot social media options: AIM, ICQ, Yahoo, Jabber and Mobile Text. Why no MySpace?
The Colossus machines were electronic computing devices used by British codebreakers to read encrypted German messages during World War II. These were the world's first programmable (if not fully), digital, electronic, computing devices. They used vacuum tubes (thermionic valves) to perform the calculations. http://en.wikipedia.org/wiki/Colossus_computer And it still runs (at a museum)
I think that "RNA" was actually in use even earlier and is still used a bit.
http://en.wikipedia.org/wiki/RNA#RNA_genomes
https://en.wikipedia.org/wiki/Inverted_totalitarianism
http://www.mda.mil/mdalink/html/aboutus.html
;)
MDA Mission
To develop and field an integrated, layered, ballistic missile defense system to defend the United States, its deployed forces, allies, and friends against all ranges of enemy ballistic missiles in all phases of flight.
1. Retain, recruit, and develop a high-performing and accountable workforce.
2. Deliver near-term additional defensive capability in a structured Block approach to close gaps and improve the BMDS.
3. Establish partnerships with the Services to enable their operations and support of the BMDS components for the Combatant Commanders.
4. Substantially improve and demonstrate the military utility of the BMDS through increased system integration and testing.
5. Execute a robust BMDS technology and development program to address the challenges of the evolving threat through the use of key knowledge points.
6. Expand international cooperation through a comprehensive strategy to support our mutual security interests in missile defense.
7. Maximize mission assurance and cost effectiveness of MDA's management and operations through continuous process improvement.
Because, when an organization is going to burn through more cash than you or I will see in several lifetimes, you can bet your bippy they'll have some fancy words out front.
Get thee glass eyes, and, like a scurvy politician, seem to see things thou dost not.--King Lear
The roots of NASTRAN go back to 1962, with the earliest versions of the code as writing by Robert MacNeill (of the company MacNeill Schwendler, now MSC, of which he is no longer a part). It was released as NASTRAN (NASA STRuctural ANalysis) in 1968 as public domain code, but the most commercially successful version is MSC.Nastran. It's nearly 1M lines of mostly Fortran 66 and a few thousand lines of C. NASTRAN has such clout that most compiler vendors need to continue supporting antiquated constructs in Fortran just so the code can compile. It's I/O model is based on tape drive access. They owe their success in great part to buying up their rivals, following the model of Microsoft.
NASTRAN has been the industrial standard for structural analysis for 40 years, but its days may be numbered unless it can adapt to newer computational models.
In 1982, I did a semester break job working for the railways in an African country. The switching was all done with circuits that used relays for the logic. The "UPS" was a roomfull of car batteries.
That switching system was made in the early 50's and is still running (on occasion) today. The greatest thing about it that you can actually fix individual relays, which is good a country with no real infrastructure where repairs need to be done by hand, and also because relays are not exactly easy to come by these days.
Ooo! I forgot to mention the best part of this code:
C ***** REPLACE HOLLERITH LABELS WITH SPECIES NUMBERS IN JCHEM *****
IF(JCHEM(M,J).NE.ISPEC(I)) GO TO 6
JCHEM(M,J) = I
It replaces the Hollerith string (stored in an integer-typed variable) with an actual integer. Because hell, the string "H2O2" and the number 6 are the same type, so why not?
Typecasting is easy when your only type's an integer.
The oldest program could be the fundamental theorum of calculus. One of the intended purposes of Calculus was to have a language that a machine could use to generate proofs. The machine was never made, but the code lives on.
It is called RNA - it ran already before DNA, which has been around for 4+ billion years or so, and it is still running in all known lifeforms.
I worked on that program. Sorry, you are wrong. We had milestones and hit them. And our stuff worked, even when modified on the fly - c.f. the shootdown of the trashed intelligence satellite by an SM3ER with out software on an Aegis crusier.
Q.E.D. You are wrong.
Essentially they just modify the executable itself rather than having the code and recompiling it. The types of people who do this also tend to be good at things like debugging programs by reading a raw core dump. From the quintessential article on the matter: "For this reason, Real Programmers are reluctant to actually edit a program that is close to working. They find it much easier to just patch the binary object code directly, using a wonderful program called SUPERZAP (or its equivalent on non-IBM machines). This works so well that many working programs on IBM systems bear no relation to the original Fortran code. In many cases, the original source code is no longer available. When it comes time to fix a program like this, no manager would even think of sending anything less than a Real Programmer to do the job-- no Quiche Eating structured programmer would even know where to start. This is called "job security"." http://www.pbm.com/~lindahl/real.programmers.html
Funny. When Israel and Saudi Arabia were being fired upon by Saddam Hussein's scuds, not knowing what was in the warheads, they thought land-based missile defense was a pretty good idea. And now that Japan has seen North Korea both detonate a nuke (albeit likely the size of a school bus) AND lob a missile completely across Japan, they rather appreciate the idea of sea-based missile defense. And we need look no further than Vladimir Putin's hostile reaction to the proposed eastern Europe missile *defense* system to see that not everybody thinks the idea is so God-Damned funny.
Don't trust anyone under thirty.