Cracking the Passwords of Early Internet Pioneers

Lots of them weren’t very good:

BSD co-inventor Dennis Ritchie, for instance, used “dmac” (his middle name was MacAlistair); Stephen R. Bourne, creator of the Bourne shell command line interpreter, chose “bourne”; Eric Schmidt, an early developer of Unix software and now the executive chairman of Google parent company Alphabet, relied on “wendy!!!” (the name of his wife); and Stuart Feldman, author of Unix automation tool make and the first Fortran compiler, used “axolotl” (the name of a Mexican salamander).

Weakest of all was the password for Unix contributor Brian W. Kernighan: “/.,/.,” representing a three-character string repeated twice using adjacent keys on a QWERTY keyboard. (None of the passwords included the quotation marks.)

I don’t remember any of my early passwords, but they probably weren’t much better.

Posted on October 15, 2019 at 10:38 AM82 Comments


Tatütata October 15, 2019 11:27 AM

Cute! At least most passwords seem publishable…

I would love to learn what Donald E. Knuth’s passwords used to look like.

I created an account on some system which imposes a mixed-case password including at least one number and a non-alphanumeric character. I used a script that generates a string according to the requirement, e.g., z6E7~yiw0Rl’Yq.

Yesterday, hardly four months later the bl**dy site, warns me that the password must be renewed, and asks “security” questions, of which none maps squarely with my case. Mother’s maiden name? In which alphabet? With diacritics? Model of my first car? I loathe cars, and never had one. Name of my first pet? That would depend on the meaning of “my”, and our pets had rather variable identification… And do goldfish have names?

Luckily I had written down the “answers” I had provided to the “security” questions.

Regarding Stuart Feldman, that would be Fortran 77, seventy-seven, and not the original Backus FORTRAN (in all-caps) of 1953. The early compilers (IBM System/360 G-level compiler for the FORTRAN 66 dialect) used dozens of overlay passes to finally excrete object code.

Petre Peter October 15, 2019 11:39 AM

They really thought they could limit the Internet to people they knew so this is not a huge surprise. On the topic of passwords, not sure why if I know my password and just wanna change it, I have to go through the security questions.

Kevin October 15, 2019 11:41 AM

I used to use words I had trouble spelling. Great way to learn to spell a word by typing it numerous times a day. To this day I can spell “necessary” without fail.

Matt October 15, 2019 11:53 AM

“I would love to learn what Donald E. Knuth’s passwords used to look like.”

Knuth just closes his eyes and concentrates and the computer logs him in.

Chase October 15, 2019 11:57 AM

Do not give truthful answers to the security questions when you set those up.

Merchants keep this information and use it to help establish your age and ethnicity and cross-reference with other accounts to identity you.

Pick questions which would never apply to you, and use random words as answers. You’ll need to keep a cheat sheet somewhere, but the notes field in LastPass, etc., serves just fine for this.

CallMeLateForSupper October 15, 2019 12:15 PM

I didn’t have a personal computer in 1980.

At work I had a terminal connected by coax to a distant mainframe. My original PW – assigned by the mainframe gods/gatekeepers – was “gcp47821”; I was told to change it ASAP, following printed rules and procedures. I think the prescription was essentially “Total of at least eight chars, consisting of letters and digits”. So my very first PW was “ao660j971w”.

Over the next 10 years I got two additional accounts on as many mainframes. Assigned PWs were “88awp” and (of the form) [1st-5-char-last-name + 1st-initial + 2nd-initial], both of which had to be changed ASAP. Proscribed PW structure changed during that time, and by the late 80’s each system enforced PW change every 6 months.

Throughout my time with the company, I entered each PW/passphrase (camouflaged, of course ) into my “scientific notebook”, which mysteriously turned up among books, awards, mementos, photos and other personal stuff long after I’d left the job.

Etienne October 15, 2019 12:52 PM

My password was “my3xbeer” for about 35 years. I finally changed it in 2017. Started out on PDP-11/73 running DSD 2.9 Unix.

Now I just use Google Chrome to generate and save my passwords.

Tatütata October 15, 2019 12:56 PM

I checked the historical passwords against a pwned password dataset, and I see that they are still represented:

dmac : 286 times
/.,/., : 158 times
wendy!!! : 17 times
bourne : 3031 times
axolotl : 1270 times

All repeated three-letter combinations from the keyboard seem to be represented. Worth looking into.

Tatütata October 15, 2019 1:11 PM

I now remember that “iefbr14” belonged to the set of my favourite passwords.

The notion of a do-nothing program that serves an crucial purpose profoundly appealed to me. It is theoretically a one-liner program (“BCR 15,14”), but I remember a paper explaining how many things you really need to add to it to get it to work.

Back in the days when my poor “brain” was perverted by the IBM universe, someone explained to me that those newfangled Acorn RISC processors everyone was talking about were really “a 360 in miniature”, or something on those lines.

I recently finally began looking at some compiler-generated ARM assembler code, which I successfully optimised by fixing a very tight innermost loop. The whole thing looked very familiar, my neurons weren’t totally wasted after all.

Etienne October 15, 2019 1:26 PM

“BSD co-inventor Dennis Ritchie…”

I thought he was a co-inventor of Unix? I think he was a east coast Yankee…

Mailman October 15, 2019 1:42 PM

I hope that this has been done under responsible disclosure principles.
After all, nothing says they aren’t still using the same passwords.

Bruce Grembowski October 15, 2019 1:52 PM

My first password was BACKUP for the account 1/18 on a DEC PDP-11/70. In fact, it shipped with that account and password.

Anders October 15, 2019 3:15 PM

Making password longer by 1 unit is more effective than
widening character base by 1 character.

Simple example. Let’s say we have password length 8 units
and character base is 26 characters (uppercase A..Z)

Then to broke that password we have to go through

26^8= 208 827 064 576 possibilities.

If we add one special character to the character base, now we
need to go through

27^8= 282 429 536 481 possibilities

There’s difference, but a small one.

However, if we just make password longer by 1 unit and
character base remains 26, then we need to go through

26^9= 5 429 503 678 976 possibilities

Difference is huge.

My advice – use sentences with spaces between the words for the password,
silly sentences, that make sense to you, maybe in your native non-English
language. Learn Estonian or Latvian or Lithuanian language and use some
nice sentence for the password. Triple caveat – you learn the language,
make good friends and stay secure too 🙂

Jeff October 15, 2019 3:19 PM

My answers to challenge questions go like this:
Q: “What is your favorite color?” A: “what”
Q: “Where were you born?” A: “where”
Q: “Who is your favorite actor?” A: “who”
I never forget my answers. That works fine unless there are multiple who/what/where questions and the system prevents duplicate answers.

Steve Friedl October 15, 2019 3:47 PM

> BSD co-inventor Dennis Ritchie…”

I thought he was a co-inventor of Unix? I think he was a east coast Yankee…

I believe this is mistaken. UNIX of course was invented at Bell Labs, but the VM stuff was added at Berkeley not long after Ken Thompson (not DMR) was visiting on sabbatical.

Tatütata October 15, 2019 4:07 PM

What Dennis Ritchie has invented is the buffer overflow attack, if you believe this article in Slate: “Future Tense, The Lines of Code That Changed Everything — Apollo 11, the JPEG, the first pop-up ad, and 33 other bits of software that have transformed our world., 14 October 2019

The Null-Terminated String
Date: 1972
The most catastrophic design bug in the history of computing

In 1972, Dennis Ritchie made a fateful decision: to represent text in his new language with something called a null-terminated string. The concept had been around earlier, but he enshrined it in his new language, which he called C, and the legacy of that decision has been with us ever since.

The title image is even inspired by this.

I’m very postjudiced against C and all its derivatives, so I won’t quibble with that proposition. 🙂

Thunderbird October 15, 2019 5:28 PM

In a perfect world, where everyone coded in PL/I instead of C, buffer overflow attacks might be rare. Until they weren’t… The problem (as I see it) is malformed data and interfaces that trust input from callers, whether it is a length-prefixed string or something with a dope vector or whatever. Any sufficiently low-level language will allow you to cobble up something ill-formed in some way that can lead to embarrassing results. Either that, or you implement some kind of high-level data structures right in the architecture which loses a lot of efficiency over a low-level scheme (accounting for the present market share of the Intel 432 line).

Not sure we can lay the buffer overflow at Ritchie’s door.

Magnus October 15, 2019 5:31 PM

“I would love to learn what Donald E. Knuth’s passwords used to look like.”

Knuth just closes his eyes and concentrates and the computer logs him in.

The computer needs a password to log in to Knuth.

Anon Y. Mouse October 15, 2019 6:02 PM


It’s even worse than that. When passwords are required to have
at least one character from a small set — i.e., digits, punctuation —
then the number of allowable passwords is greatly reduced, even when
you account for the possibility there is more than one character from
the special set. Additional requirements — such as a punctuation
character and a digit — further reduce the number of potential

shelby October 15, 2019 7:47 PM

Dave Aronson, “axolotl” could also be a reference to the axolotl tanks from the 1969 science-fiction novel “Dune Messiah”, which are used to clone human beings.

John Smith October 15, 2019 8:21 PM

from Anders:

“My advice – use sentences with spaces between the words for the password”

My experience has been that spaces in passwords will prevent remote logins on some systems – a local login at a machine will work, but a login from across the local network to that some machine won’t.

Confirm it works locally and remotely before relying on it.

x2bike4u October 15, 2019 9:09 PM

Back in the day, toor was the root password we assigned to our customers’ Unix systems.

There were a couple of DG Nova clones that didn’t use passwords at all, just user ids. Not sure if this was how the DG systems themselves worked.

stormwyrm October 15, 2019 9:13 PM

Apparently Ken Thompson’s old password was much, much better than any of those. It turned out to be “p/q2-q4!”. It’s a common chess opening in descriptive notation, and it resisted all attempts at cracking until very recently. Thompson contributed to the development of computer chess and that password would have been easy for him to remember. Good one for the author of “Reflections on Trusting Trust”.

Clive Robinson October 15, 2019 10:24 PM

@ Tatütata,

And do goldfish have names?

Yes they do 😉

Sad but true story told to me by Steve Crook, who also gave me my first Unix account which was on a PDP 11-70 (the password I picked was “pawns”, which got “shoulder surfed” by one of the Profs, who aluded to it with chess moves even after I changed it[0]).

Apparently his sister moved and gave Steve two goldfish in a tank. Being a thoughtful person he asked his sister what their names were, and was told “puddle glumms” which were some kind of cartoon charecters.

Well Steve looked after the gold fish and one day as these things happen, one of them died. Steve thus decided that the one that had died must be “puddle” because now the other one must be “glum”.

But vets can be quite awkward with names as well. To be “family friendly” some record the pets name as the equivalent of a first name and attach the owners surname after it in their records. My second pet rat was a rather charming black rat I named “Rebecca” for no better reason than the joke of calling my first “R4” was lost on most people. Thus when waiting my turn at the vet the assistant would call out for “Rebecca Robinson”.

Well I didn’t think much about it untill several years later when life was way to busy for hanging onto girl friends let alone pets[1]. My boss of the time was Alan Poole, and he came into the office grumpy and late one day so I wandered in for a chat. He related a story about his girl friends cat which was officialy called “Cecil” –which apparently means “blind”– that as is the way with such names had been shortened to “Cess”. Alan was not aware of this vet naming convention when registering the cat with the vet, and only had become aware of it that morning having taken the cat in where upon the receptionist called out for “cesspool”…

[0] I got revenge, it was the 1980’s and his two shiney new research computers were Perqs which had rather nice graphics screens and quite a good PacMan game on them. What he could not work out was how even after he had locked the computer down I was still to be found playing PacMan at lunch times using the guest account that he thought he had removed… The Perq had a problem in that the person who had written the login program had only used a quite short char buffer for the password, and had used the wrong Clib function that did not do bounds checking… I discovered as some one did many years later with SunOS the right key combination to drop you into a root shell… Well they did know when they employed me I had been a bit of a hacker, as over at the college I had set up my own hidden account on their Prime system in the late 1970’s. A habit that I stopped after I hand been involved all be it peripheraly with hacking the “BBC Micro Live” program that an estimated 10million TV’s had tuned into. The “poem” that got displayed was appropriate to this thread and you can read it,

[1] A friend who had, had a relationship with a young lady, that not only got as far as getting a house together but a pet as well had it go bad on him. As he puts it “She got the house, but I got the cat” with a slightly fierce emphasis on the last half.

name.withheld.for.obvious.reasons October 15, 2019 10:36 PM

One feature to use on a local windows OS is to add extended characters to the password string. Hold the ALT key down and enter the decimal value of the character map from the number pad, then release the ALT key. So hold down ALT, entering 7, and release which is CNTRL G (bell).

Tatütata October 15, 2019 10:59 PM

As he puts it “She got the house, but I got the cat” with a slightly fierce emphasis on the last half.

I inherited my late mother’s cat, which has a number of nicknames (the cat, not my mother). Her “official” one is a very cute name in French I won’t betray here. Otherwise, her proportions often inspire me to call her “Botero” or “Barclay’s Bank”. Sometimes I will designate her “AN/FSO-9000g”, as in “ANimal: Feline, Striped, Overweight; 9kg”, in reference to the AN/FSQ-7 which weighed a couple of pounds more than a sawmill.

To be “family friendly” some record the pets name as the equivalent of a first name and attach the owners surname after it in their records.

My local council (re)introduced pet licences, including for cats. As the fee depended heavily on whether the animal is “fixed” or not, I “borrowed” my sister’s cat documentation, as I never had any vet papers for Botero. (She was however spayed, but isn’t microchipped). It’s still the same family name.

Technically speaking, I committed identity theft… (Or was it my cat?)

Tatütata October 15, 2019 11:42 PM

In a perfect world, where everyone coded in PL/I instead of C, buffer overflow attacks might be rare.

If you coded in PL/I on a saurian IBM machine, most of the universe (eg: early scanned photos!) fitted exactly in fixed 80-column wide records (Image processing on card images! Yeah!), so buffer overflow war kein Problem. You would rarely read streams of bytes like you would do in VMS or Unix or MS/DOS or whatever.

In Fortran 66, the most efficient data type for carting text data about was a REAL*8 vector, which you equivalenced to a 1-byte wide type. (Word alignment faults caused exceptions, but that’s another story).

I almost entirely forgot PL/I, but I was good at it. The downside was that it took forever to compile compared to Fortran. You kind of sprinkled colons everywhere, which caused me problems when I began coding in Pascal which is comparatively a stickler for punctuation.

Clive Robinson October 16, 2019 12:25 AM

@ Tatütata,

“Barclay’s Bank”

+1 🙂

For reasons to complicated to relate, I acted as middle man to my sister getting a very large ginger tom I’d named “wart”[1]

He was an odd but very likable cat, he had as far as I know never miowed but did squeak if he was upset, and had a purr so deep it was like a distant earthquake. He weighed in at a little over 18lb or 8.4Kg, but he was not fat. Far from it he was a lean mean fighting machine when it came to the local dogs, even the local child scaring guard dog was petrified by the sight of Wart who had quite idley shredded it’s nose with three or four well timed swipes.

He could also climb and would often be seen on the roof ridge of houses, especially if they had a bird table. He was not much interested in song birds, but he did find using a pigeon as a stepping stone to get off of a roof ridge, did also give him a snack. He had near decimated both the locsl pigeon and squirrel population by converting them to snacks. I must admit I never saw him play / taunt other creatures, he either ignored them or just ate them.

He could also jump as I’ve mentioned I’m a little under 2meters and he had no trouble jumping from the floor up onto my shoulders, which is rather more than I can say for me, having 18lb “dropping in” on you can not just leave you surprised it can totally floor you.

Most cats I’m not fussed about, they are in most cases patches of fur you avoid sitting or standing on or tripping over when you visit people. I ignore them and if they have any sense they ignore me, but some will head butt me to get attention and depending on what mood I’m in I’ll either flip them on their back or have a wrestle with them.

Wart however was different he could lie down in a way I’ve never seen any other cat or dog could do. Large as he was he would just melt into things totally relaxed and totally at peace with the world. If he wanted to be stroked and trust me it was his choice with adults, he would move in close to you and languidly put a paw on your hand, when he had your attention he would then get comfortable, and then you could start stroking him. Your reward was his extrodinary purr that would put you off to sleep. With small kids he would slouch up with a funny sort of walk you see some pumas and lions do but with a swagger that made you think of a wild west gun slinger, all the while purring, and it was like magic the child would be totaly mesmerized and become quiet and still.

Although he’s nolonger with us, he had a good innings,

[1] Because “he grows on you like a…”.

Patrick October 16, 2019 1:47 AM

Note that these passwords were (as I read in another post on this) limited to 8 characters!

Gerard van Vooren October 16, 2019 2:00 AM

I don’t remember my old passwords as well and I got into this when my company started to complain about password policy that annoyed me a lot. Personally I still like the XKCD password scheme of a lot of words but without any special characters and without capitals [1]. I also still have a small book in where I write down passwords. I hate password mangers [2].


Tom Bäckström October 16, 2019 2:44 AM

Though I find the revealed passwords interesting, I find the approach to publication distasteful. Passwords are personal information and it is unethical to reveal passwords together with the user names.

Anders October 16, 2019 3:51 AM

@John Smith

On what system spaces don’t work?

Sorry, on year 2019 i don’t know any system, that won’t allow
you to use spaces in the password or don’t allow to use passwords
at least 15 units long.

Sed Contra October 16, 2019 4:11 AM

Your password only has four characters?
I also like to live dangerously.
— Austin Danger Powers

Some password checkers rate things like

1111111111111111 … lots of 1s

as extremely safe.

Wael October 16, 2019 12:11 PM


I don’t remember any of my early passwords

Ahaaaaa! So you change your passwords after all 🙂

Security Sam October 16, 2019 1:23 PM

The painstakingly crafted password
Can turn into a double edge sword
Not only can it put the rival on hold
But, you can end up out in the cold.

Timothy McNerney October 16, 2019 1:45 PM

When early MIT AI Lab computers started requiring passwords, Richard M. Stallman (aka “rms”) chose the deliberately insecure, reflexive password “rms”. This did not become a source of consternation until the ARPAnet began connecting research computers across the country. Stallman had accounts at several universities, and his login credentials became well-known among guest users.

But in this era of “security through obscurity,” malicious acts were rare or non-existent. And it wasn’t necessarily “randoms” who were the problem. One memorable case of bridge-burning/foot-shooting was performed by Stallman himself: He so strongly disagreed with the newly-introduced convention of wrapping Lisp “special variables” in stars (e.g. print-base), that he systematically went through the MIT Lisp Machine sources, removed all the stars, and deleted the relevant backups too. It should be no surprise that, when time came for Lisp to leave the lab, and Symbolics, Inc., the very first .com was being formed, Stallman did not get invited to be a founder. The startup soon became his nemesis, as he helped competitors LMI and TI by doing the work of ten programmers, reproducing each new Symbolics software feature “almost clean room” style.

We can be thankful that it wasn’t long before Stallman set his sights on AT&T Bell Labs, single-handedly writing GNU Emacs and GCC before inspiring, Tom Sawyer style, a full-blown effort to scratch-build a work-alike Unix operating system. The rest is history.

Who? October 16, 2019 2:11 PM

Our host password is “schneier” but no one except him can type it successfully. Cracking this password makes John the Ripper fail and its resulting core dump only has the root password for the machine it was running. 😉

Sorry, I cannot resist.

Who? October 16, 2019 2:13 PM

On a more serious matter, I really miss the time Internet was a small community with only a few millions users. Security was not a high priority, there was no spam, and Internet itself was a community of clever people that wanted to help others.

When I joined this network it had only one million and a half users in the world. It was three decades ago.

Clive Robinson October 16, 2019 2:38 PM

@ Who?,

When I joined this network it had only one million and a half users in the world. It was three decades ago.

So you are one of those “newbies” I hear so much about 😉

Any one else from the UK remember when London’s County Hall had (D)ARPAnet access back in the early 1980’s?

Or any in the US remember the “Whole Earth ‘Lectronic Link” or just “WELL” back in the late 1980’s?

Happy days.

Anders October 16, 2019 3:13 PM


I remember when Internet came to Estonia, 1992.
No WEB yet, only FTP sites and Gopher 🙂

dancing on thin ice October 16, 2019 3:57 PM

@ Sed Contra

I know that in general the longer a passwords/passphrase is, the harder it is to crack.

It may be security by obscurity but am curious if those with 4 or less characters are even checked, specially if extended or non-standard characters are used like ♬Я☞♣︎ñ♛‽

Anders October 16, 2019 4:17 PM

@dancing on thin ice

This breaks usability. OK, it works on one password, but what
if you need to remember 10..20 passwords? Then you just won’t
remember in which password you had those special chars and
you need to start using password manager. You just can’t
remember those passwords. Also you need always access to
Charmap to enter them.

If you choose longer password sentences, it’s possible to
remember them if the sentence is tied with the system.
For example one password:

This computer has now brand new Ripjaws RAM

No need to use special chars and it’s easy to remember.

BTW, last Flare-On had Keepass hack as one challenge.

Password managers can be hacked, noone can yet hack your brain.

Tatütata October 16, 2019 4:37 PM

Stephen R. Bourne, creator of the Bourne shell command line interpreter, chose “bourne”

That’s the famed Bourne identity!


Hey, Clive, I was using X.25 when the Carter and Schmidt era were still underway, but before the Mitterand and Thatcher began. Later, I was dabbling in RTTY and AX.25, and had permission to use Bitnet. (Only the very select few had access to that mysterious Internet thing). On Bitnet I was printing out messages from Prof. X from overseas and carried them to Prof. Y upstairs, who honestly wanted nothing to do with Prof. X. (Is that social engineering, or spam?)

But no internet access before 1993, when I became a member of a co-op.

Does this make me a newbie or an oldbie?

Tatütata October 16, 2019 6:18 PM

Thanks for the stroking Anders, I was just kidding, I should have written “fossil”.

Since you’re mentioning the time before 26 December 1993, I’m curious about something.

1) Was there a BBS scene, like there was in North America and Western Europe?

2) Interesting things could be heard on the line when one dialled the “8” trunk prefix in the former Soviet Union telephone network, with a series of ANI tones followed by a second dial tone, at least in the main cities.

Was there anything resembling a local phone phreaking scene you could have been aware of?

Anders October 16, 2019 7:06 PM


1) Yes, BBS and Fido too. Very warm memories.

Example of our Dark Corner BBS filelist

2) Phone phreaking wasn’t very widespread here.
There were tricks to get free calls from payphones,
but most of the times they out of order anyway 🙂
For example because of missing handset 🙂
Handset speaker capsule elecromagnets were
used for electric quitar picups so sometimes you
were quite lucky to find working payphone in the

Clive Robinson October 17, 2019 1:31 AM

@ Tatütata, Anders,

I was using X.25 … I was dabbling in RTTY and AX.25

Gosh that takes me back, to many many accounts and wierd, funny and some downright rude passwords not just as a user on bulitin boards but also developing Terminal Net Controlers (TNC’s) for use over radio networks including VHF SMS bulk gateways into GSM networks. Where sometimes passwords were the only way you could “get back” at managment… I remember 1345 being used as a substitute for “Bas”, long befor “l33t” speak turned up with “script kiddy” “wanabies” (god we were so smug back then when even dinosaurs like me still had flesh –all be it putrid– on our bones 😉

Yes, and even earlier times… X.25 came out of work on System X done in part by a war time buddy (Ken Gravit) of my father and maternal Uncle back in the 1960’s when Ken worked for the General Post Office (GPO) up at Dollis Hill and one or two other places[0]. It spawned all sorts of stuff including “View Data” of which the GPO/BT version became known as Prestel, and the French had a version called MiniTell if I remember correctly (The German one is lost in the mists of my memory just past that sign that says “Here be dragons and unicorns 😉

The thing was back befor we went “digital” there was this international system called Telex, using dial up teleprinters using a variation of Baudot signalling on a +-80V line at upto 40mA which realy could wake you up if you made a mistake. It was expensive to use but messages sent over it were given legal status in commerce. Then I found out that some bright spark had connected X.25 to the Telex network via a bridge… So all I had to do was dial into the local X.25 PAD send the appropriate numbers and bingo trouble soon followed…

It was these old 5bit Baudot machines that formed the basis of the Amature Radio Radio TeleTYpe (RTTY) network.

X.25 remained a bit of a mystery to most untill Phil Karn (KA9Q) came up with AX.25[1] using HDLC NRZI into a 300baud Bell 102 modem into a phone patch. To keep things simple AX.25 was designed as a Point-2-Point protocol more similar to ethernet than SLIP, thus switching and routing had to be done by higher level protocols. It’s hardly heard on HF today but you can hear it on V/UHF using either Bell 202 1200baud or mote likely James Miller’s (G3RUH) 9600bps DFSK as the physical layer where it forms the base network for amongst other things the Amature Packet Reporting Service (APRS) or TCP/UDP.

Then the first dial up Internet Service Provider for London “Demon” took Phil’s code and modified it to do IP “And the rest as they say is History”…

Whilst AX.25 is still in the Linux kernel for good reason few Ham’s who use APRS for geo-tagging thrmselves realise where it came from. But with the somewhat amazing spin off of work by Nobel Physicist Joe Taylor’s (K1JT) FT8 has given us JS8call which now can carry APRS messages you can text a friend via HF radio –all be it very very slowely– via HF radio to the likes of New Zealand and beyond depending on the time of day and which great circle path you take (a real bone of contention with Bonnie Crystal (KQ6XA) and her group of self appointed types at HFpack forcing their beaks in uninvited).

You’ve reminded me that, I realy should do as I prommised, –now you can get an app for an Android phone to generate APRS to audio out thus any HT can APRS– and make the equipment for a new packet gateway in South / East London (I’ll let others apply for the NOV to licence it and give it the love such things need).

[0] Ken was an inyeresting chap, he had a photograpas eye for colour and was one of the first to process his own colour film in a home dark room. That eye, also gave excercise to one of his quirks, in that he had disdain for the socks available at the time. At heart he was a little rebel and although wearing a somber suit and drab tie, his way of yelling was that he had “loud socks”… The problem was back then like paint mens socks came in three colour of “dreary / drab”. So he got his wife to teach him to knit, and on the train up in the morning he could be seen either playing poker with some of his regulars, or quietly knitting socks out of his briefcase in wool specially imported from America for making gaudy baby clothes. Likewise on the way home.

[1] He got involved with AMSAT and the last time I spoke to him he was getting kids involved with high altitude ballooning as science projects and a lead in way to designing CubSat electronic payloads.

Clive Robinson October 17, 2019 2:11 AM

@ Anders,

If you choose longer password sentences, it’s possible to remember them if the sentence is tied with the system.

Reminds me of possibly the longest password I ever came up with for a customer…

Without letter/number substitites it was,


Which is what you would remember, then “auto-replace” as you typed on. So,

    0, replaced o/O’s
    1, replaced i/I’s and l/L’s
    3, replaced e/E’s
    4, replaced a/A’s
    5, replaced s/S’s

And so on. Thus following the basic “Password rules. Oh and the full stop / point was important for the “at least on punctuation mark” rule…

random guy October 17, 2019 4:59 AM

@John Smith

I say don’t use space at all. Space character in password is a nuisance for everyone. If you allow space, you need to set extra rules to prevent passwords starting or ending with space. Most string inputs are trimmed from both sides, most likely by a called function you didn’t write. There are also some symbols that looks like a space, and if you copy/paste your password, you can’t tell what symbol it really is. There are still many special characters you can often use. For example: <°))>< (although brackets are often forbidden because they can be used for code injection)

I also have a funny story with space. In one particular game, I’ve created a character with space in its name. When I tried to delete the character, game wanted me to re-type the name. Guess what, form for typing the name didn’t allow space 😀

Anders October 17, 2019 9:24 AM

@Clive, Tatütata,

What was your first computer?

The problem with Soviet Bloc was
that home computers were virtually non-existent
til second half of the 80’s, and even then remained
inaccessible due to high price.

Engineer’s salary was 120 rubles, 600–650 rubles
was very expensive.

Micro-80 was basically out of reach due to lack of

Radio-86RK changed all that and become hit among
radio amateurs and computer enthusiasts.

Tatütata October 17, 2019 10:25 AM


Ooops! I just realised that the Baltic states escaped the Empire two years before its collapse.

I suppose that early adoption of internet in Estonia is a happy result of patching into the Finnish telephone network?

There are only 80km between between Helsinki and Tallin, which is entirely manageable by microwave, and suitable towers were probably available at either end. It would have been a simple matter of sending technicians with a few crates of equipment.

BTW, East Germany was still operating a century-old telegraph cable to Sweden in 1989. And Deutsche Telekom made lousy technical choices when they rebuilt the network in the 1990s, which later made the provision of DSL a lot more difficult.

My first physical contact with a “computer” was with an IBM 029 keypunch as I was 8 or 9. I loaded a deck of cards (some of which I had tediously punched) in a IBM 1442 card reader, which was taller than me, and pressed START. Wow! Sometimes later came a printout out of an IBM 1403 line printer connected to an IBM 2821 control unit. I was in awe before these beasts that could eat through a whole carton of paper in seconds when the carriage control tape was dirty, and had a furious tendency to leak hydraulic fluid. They also needed a three-phase power supply. The last one I saw was unceremoniously butchered on the spot for gold by knackers circa 1990.

The first computer I could actually call “mine” was a 6502 based <a href=”, circa 1981. I plundered my piggy bank and got 4k of SRAM, and the parts to build a regulated 5V linear supply. Probably the most elaborate thing I ever did with it was an ASCII-to-Baudot serial converter, using the shift register of the 6532 (which shifted out in the wrong direction). A later version used a kind of time scheduling executive, with a 6532 timer generating an IRQ every X millisecond, and the servicing routine doing all the bit-banging I/O, the conversion being performed in the foreground. It would have been an almost trivial job with an UART (a 6551 would have been nice, alas, I just didn’t know where to get one. (I could get 8080 parts, but didn’t know how to interface them).

Greg October 17, 2019 10:44 AM

@ Clive Robinson

After a single substitution (which may be required by the password rules for whatever you’re logging into), following set rules for replacing letters with numbers doesn’t really add security except through obscurity (in that others may not know that you had a set of rules).

What I wonder is whether there are estimates for the entropy of sentence initial passwords. (That is, using “WIwiwtaefteosip.” in place of the previous sentence.) It’s obviously more secure than a word or phrase that totals the same length, and obviously less secure than a random string the same length, but how much can we narrow those bounds?

Anders October 17, 2019 10:54 AM


It was actually a satellite link to Sweden,
for KBFI, financed by Soros.

Put into the Google phrase

“In the volatile fall of 1991 representatives from the Institute of Cybernetics and KBFI”

And you find a book “The Dynamics of Innovation in Eastern Europe: Lessons from Estonia”

But long distance calls to Finland was popular in Fido/BBS time, from there
we got all our software 🙂

Tatütata October 17, 2019 11:35 AM


There is just enough of the book excerpted in Gurgl-books.

The Soros-funded link opened in April 1992. The advantage of the satellite solution is that it could also be applied to the other Baltic states.

But in the very next paragraph I can read:

The Institute of Cybernetics, for their part, would soon establish its own Internet connection to the Nordic world when Eesti Telefon opened a new powerful microwave link to Helsinki, which went online in May 1992, i.e., only a few weeks after the establishment of KBFI’s satellite connections. However, the opening up of EstNet, as the Institute’s network was called, followed only in January 1993, as there were problems getting the necessary permission to import the Cisco equipment. When a new link between Tallinn and Tartu was put into operation, and the grant for the satellite connection depleted by spring 1993, the satellite was shut down. The links between Tallinn and Helsinki were also strengthened considerably through the opening of a new fibre optic cable across the Gulf of Finland at about the same time (information from EsData, 1994).

So my hypothesis wasn’t off the mark.

West Berlin was 180km away from the closest practical point in the West-German “mainland” (in the Harz mountain range), a distance which was difficult to bridge with radio. But from the 1940s increasingly elaborate installations were built, first with low-VHF circuits, and later with impressive troposcatter installations. Such was the cost of maintaining West-Berlin.

In the 1980s these circuits were augmented by satellite backups, and by 1988 a fibre-optic cable operated by the East-German PTT was finally installed. I guess that the attitude must have been “with analog microwave, they can copy it all anyway”…

Quiet accommodations were nothing new. There is the story of this Radio Free Europe or Radio Liberty circuit going over East German lines to West-Berlin, for which they were happy to collect hard currency. But one day a technician patched the feed by mistake for an hour into a GDR radio station.

ps: I mangled the URL again. It should have read: Synertec SYM-1.

Tatütata October 17, 2019 12:00 PM

He could also jump as I’ve mentioned I’m a little under 2meters and he had no trouble jumping from the floor up onto my shoulders …

I had one like that. My mom rescued a barn kitten covered in muck and parasites. When I first saw that thing, I honestly doubted she’d live to see the next day, especially after she fell into a profound catatonic shock as I washed her. OK, it had been a hugely stressful day for her, but I nearly went into shock myself.

She recovered and became quite vigorous and nimble, but remained a fairly small cat. Still, her claws would make you writhe when she jumped onto your shoulder — or crawled up your leg while you’re eating. I would bait her with a bit of ripened cheddar placed on the top edge of the kitchen door, where she would jump and dance on the spot to get more. Compare that with my current fur ball who must make elaborate physical and mental preparations just to jump a paltry 50 cm to land on the bed…

Most cats I’m not fussed about, they are in most cases patches of fur you avoid sitting or standing on or tripping over when you visit people.

That’s the problem when they become too relaxed and trusting, they will just lie in the passageway to keep an eye on the action, and you have to go around them. Mine’s tail has an uncanny ability in predicting exactly where I will next set my foot. 🙁 That seems like a good candidate device to engineer an infinite improbability drive around.

The CIA tried to use the couch potato property of cats by hacking bugging devices into them. That is frankly weirder than this recurring nightmare of mine in which I was reprogramming my cat. 🙂

The usual current for a teletype magnet was 60mA, but in the transistor era loops would operate with 20mA (or use a modem, like for the TWX service). I’m not sure where your 40mA comes from.

The transformer for my TTY loop supply was, ahem, “borrowed” from an appliance no one needed anymore. (Not a Soviet pay phone). As I suspected, no one ever noticed. The HV transistor was horizontal sweep device from a TV. At least, it was one of my more carefully put together contraptions, I bribed my kid brother to plug my creations in the mains outlet, as I was too afraid to do it myself. I also chose the testing time carefully, as the fuse box (which was made of thick slabs of slate with exposed conductors) was inconveniently located in the kitchen, on the wall behind the range close to the ceiling.

Long before I had a real grasp of electronics, I created a number of modems, each stranger than the other. I did manage to get them to work, somehow. Before I got a commercially built TNC, I did assemble a 202-ish modem using Exar chips. The circuit was kosher, the assembly method wasn’t. A regional ham club would broadcast a weekly bulletin with all the latest gossip over VHF repeaters (the anti-blabber circuit was somehow disabled), that was about the most interesting thing I did with it.

I collected my assorted brain-droppings, including my industrial designs, in a clearly labelled box, which I left with my parents. Some years later I could never find it again, I suspect my old man threw it out by spite. 🙁

As the Soviets withdrew from Germany, someone went around and collected the improvised TV antennas the forces left behind. They were actually exhibited as “art” in a modern art museum circa 2000. My junk too could have landed in a museum, it was odd-looking enough.

Sed Contra October 17, 2019 1:22 PM

@ Tatütata

Re: inventor of infinite improbability drive

“he was lynched by a rampaging mob of respectable physicists who couldn’t stand him being “a smart …””

Long ago in a faraway amphitheater lecture hall with many, many ranks of seats, a physics prof was lecturing on kinetic theory of gases and how it was practically infinitely unlikely that all the air molecules in the room would bunch up in a corner of the ceiling leaving the rest of the hall on vacuum. As he had just finished that remark, someone in the uppermost rows dropped their pen cap, which bounced and clicked its way down the endless steps running between the rows of seats, until it reached the bottom in front of the lectern, where it bounced back and forth a few times, then came to rest standing up.

lurker October 17, 2019 3:45 PM

@various I’ve been so long using systems that refused spaces in passwords I’ve just got in the habit of not using them. Some systems refused $ / \ . because these were too sacred…

Wayne October 17, 2019 5:26 PM


Sorry, on year 2019 i don’t know any system, that won’t allow you to use spaces in the password or don’t allow to use passwords at least 15 units long.

My former bank “upgraded” their online system, requiring me to change my password. My password methodology is a keyword plus a identifier for the site in question with my own obfuscations for both, and since it is a bank we’re talking about, additional obfuscations. I sign on, change the password, verify it, sign off.

Next time I sign on: Invalid Password.

Call bank, get a reset, sign on, change password, do stuff, sign off.

Next time I sign on: Invalid Password.

Call bank, get a reset, sign on, change password, do stuff, sign off.

There’s something wrong here. I’m not forgetting my password. Call bank. Finally talk to someone in the know. They ask me how long my password is, I reply X. They say, “Oh, it can’t be that long. It can be no longer than Y (a ridiculously short number like 10 or 12).”

I closed my account and changed banks.

Another problem with this bank was they broke a fundamental security rule about databases. When your backend database is unavailable, never reveal information on the frontend that identifies what you’re running underneath. One evening I went to sign on, and the server had crashed. The web page displayed ODBC error messages that said it was connecting to a PARADOX database!

I kid you not! Paradox! Why I did not close my account then, I do not know.

Clive Robinson October 17, 2019 6:23 PM

@ Greg,

following set rules for replacing letters with numbers doesn’t really add security except through obscurity

No but it was not supposed to.

Though the argument could be made that,

1, Long passwords are “considered” all alphas without spaces by an attacker, with the last a default “.”.

2, Then for a password of length N alphas, the brut force size is N^52.

3, Thus if numbers are included but without mapping to alphas this goes up to N^62.

4, Further if their is a mapping whilst the password remains at N^52 (or less). The required brut force search still remains at N^62.

It’s only when you get into XKCD style passwords and “word guessing” attacks from a dictionary does the mapping realy come up. As this is assumed unknown to the attacker then this will effectively increase the size of the dictionary.

But what if the user decides to modify the XKCD model, by rearanging them to make them easier to remember? Well again you will have a password with less entropy especially as you work from the first to last word, but the brut force entropy remains the same as for the XKCD system.

But that modification is turning a random string of words, into something aproximating a sentance. It obviously has less entropy as the number of valid sentences is way smaller than the permutation of just words.

Which brings us to your question,

What I wonder is whether there are estimates for the entropy of sentence initial passwords.

Well yes their are and it’s less favourable than with sentances. Because the length of the password N leakes the number of words in the sentance which very much reduces the brut force search space to a known subset of all possible sentences…

The highest entropy per letter will be not the first but second or third for sentances that do not start with proper nouns, it will then fairly rapidly diminish to quite a small value of just a few bits on average from then on in.

Especially with “known sentances” from the likes of poems etc.

That is if you had N=6 and you were trying “The Cat Sat…” TCSOTM is by far the most likely, with the entropy for the last three words being considerably less than 1bit.

Likewise for N=6 “The Owl And…” is going to be most likely TOATPC. Even more so if it’s the full N=15 for TOATPCWTSIABPGB. This problem has been known by cryptographers for years. With the UK SiS/MI6 telling the SOE that Churchill valued so much a pack of nonsense with “Double Transposition Cipher with Poem Code” was the most secure. Even when the “poem” was unique and had never been published it was still weak especially for short messages and the Cryptanalysis section of the German Radio Security Service had no real trouble breaking them. In fact history shows that simple mistakes would cause the SOE decipheres to fail with “undecipherables” that cost many lives due to the dangers of “resends”.

Because of these known issues with “Poems” what you call “sentence initial passwords” are likely to be very very weak if an attacker knows you are using them.

The problem is that of “shoulder surfing” via CCTV or even microphone. Even if you can not see the keys the user is pressing if you video record them you can often work out from their latter typing the probability of which keys they are pressing for their password. This can be done even if you only have a rear view, as long as the attacker can see your shoulders and elbows you are likely done for, unless you have the required level of OpSec “paranoia” to actually deliberately change the way you type when entering passwords.

It’s just one of the reasons I’ve mentioned in the past “umbrellas” those nice big golf ones, even when you think you are safe at home, you don’t realy know if someone has put a miniture camera and microphone in some object in your room being under a golf umbrella you have to hold with your shoulder and leg or table edge, will change your typing style as well as hiding your hands from overhead “in smoke detector” cameras and hanging over your shoulder to the extent it also hides your shoulders and elbows from cameras in room objects.

Unfortunately with WiFi present especially that which operates at 5GHz or above it can see through the umbrella and effectively map your hand and finger movments… The same will apply to certain varients of 5G and 6G which is probably less than a decade away.

Unfortunately technology is moving so fast, it’s rather more than a full time job trying to keep OpSec rules sufficiently current to be secure from it.

Clive Robinson October 17, 2019 7:13 PM

@ Anders, Tatütata,

What was your first computer?

It’s a good question…

The first finished commercial computer that was of any real use for Word Processing / high level language (Pascal and Fortran) and serious usage via modems etc was the Apple ][ which I stll have it cost me all up nearly 5 months take home pay. Not the most expensive computer I’ve bought (one was 8months equivalant that had specialised video input cards and 8 very fast SCSI hard drives, and an expensive RTOD I used to develop a video encryption system).

But I’d built a number of computers prior to the Apple ][ one was based around the SC/MP micro chip and another based around two 74LS 4bit bit slice processor chips on hand wrapped “wire wrap sockets on a Euro-Card which was simillar to the S100 system.

I was also possibly amongst the first people to build an all 5V home computer due to being able to source 5V only ROM and RAM byte wide chips directly from the Japanese manufacturer as “samples” through the company I worked for that was subcontracted to build medical imaging equipment.

Somewhere I still have a tube of UCSD P-machine processors that in theory can run programs written in Pascal or Fortran on the Apple ][. However the P-machine support chips had difficulty interfacing with other chips due to odd bus timing.

Oh interesting fun fact, not so long ago somebody did some tests on “historic computers” from the Apple ][ Single CPU running at 1MHz with 64Kbytes DRAM to modern multi core running at multi GHz with 4GByte RAM. The tests were “response times” in a simple text editor.

The Apple ][ came well out ahead of the pack. Although it had a few wiggles in the curve, the general trend is the more powerfull compiters get, the slower they get in responsiveness… Yup that nice new shiny high end gaming desktop is actually a dog compared to a 1970’s 8bit business compiter when typing in text etc (which is still what most computers are used for).

Oh and another statistic, office workers were most productive in 1973… Since the introduction of computers both main frame and personal, real productivity has declined… As pointed out by a friend just a few days ago when talking about it, “It’s bl**dy marketing sticking their noses in with corporate synergy cr4p and nonsense”…

I must admit on trying to find counter arguments that did not sound “lame” I failed misserably. So if you’ve some good counter points now would be a good time to let me know 0:)

Greg October 17, 2019 11:15 PM

@ Clive Robinson

Because the length of the password N leakes the number of words in the sentance which very much reduces the brut force search space to a known subset of all possible sentences…

Sure, but the number of 15-word sentences is astronomically higher than the number of 16-letter sequences you can get by putting full words together and sticking a period at the end. It’s not nearly as high as the number of completely random 16-character strings using letters and punctuation, but the point is to have something more memorable than that.

As for the poem code, my understanding is that the insecurities had more to do with the mechanics of encryption and fact that poems were reused than with a lack of entropy in those poems.

I don’t fully understand the poem code, but as an analogy think of a Vigenere cipher. If you use a completely random 20-letter string as the key to a sufficiently long Vigenere cipher, you can break it with a few seconds of frequency analysis on a computer. That doesn’t change the fact that the initial string had 94 bits of entropy and would have taken a lot longer to brute force if it hadn’t been used in an insecure cipher.

But that modification is turning a random string of words, into something aproximating a sentance. It obviously has less entropy as the number of valid sentences is way smaller than the permutation of just words.

It reduces entropy if you only permute the words into something more memorable, because then different RNG results can give the same final password. If instead you make it more like a regular sentence by adding more words between the random ones (and keeping the order), then you are adding entropy. Just like how you reduce entropy if you always replace all i’s and l’s with 1’s, but you increase it if you randomly choose some subset to replace.

Clive Robinson October 18, 2019 2:29 AM

@ Greg,

Never type with a tired brain trying to get back into it’s home time zone…

In my above to you my “N^52” etc should be “52^N”

With luck my brain might be back on UTC+1 just in time to get hit with the change to UTC as we “fall back” from BST to GMT…

Clive Robinson October 18, 2019 6:25 AM

@ Greg,

Sure, but the number of 15-word sentences is astronomically higher than the number of 16-letter sequences

That is because of “alphabet” size. The minimum recomended dictionary size for the XKCD method is 2^10 or 1024, the alphabet size for upercase alphers is 26. Of course the down side of using thr initial letter of words is that in effect the “EAT ON IRISH” or “A SIN TO ERR” frequency distribution comes up, with the fact that many words map to the same Capital letter.

That said the point I was making though, is that using whole words instead of just their initials makes the length of the password significantly variable and unrelated to word size. Thus knowing N does not leak the number of words. Not knowing the number of words means in a brut force search space would have to go through say all sentances made of two very long words[1] which through all sentances of words using shorter and shorter words. Thus it gets you back to very nearly the same size as the brut force search on an assumed dictionary.

As for the poem code, my understanding is that the insecurities had more to do with the mechanics of encryption and fact that poems were reused than with a lack of entropy in those poems.

With regards the double transposition poem code, yes the double transposition made short messages weak. Also sending an agent into the field for long time periods made the reuse of a poem way beyond safe limits a racing certainty… But the reasons it was “realy realy” weak was that the use of poems learnt at school etc was highly predictable. So the Germans could guess the line of the poem. Thus the real entropy of the KeyMat was very very small. Break it once and all past present and future messages were laid wide open, worse in fact than accidental reuse of OTP KeyMat, because all keys are related due to the poem unlike the OTP. Which all ment brave people tourtured or died needlessly as the seniors in SiS/MI6 knew they would and probably planed it so due to “turf wars”…

… think of a Vigenere cipher. If you use a completely random 20-letter string as the key to a sufficiently long Vigenere cipher, you can break it with a few seconds of frequency analysis on a computer.

If you are talking about the debased form, it’s a case of “key reuse” thus after a string of length 40 you have in theory everything you need to know. You will realise the key length, thus align the individual Ceaser ciphers and use simple frequency analysis on each.

Double transposition and Vigenere are very different ciphers. Transposition ciphers mearly shuffle the position of the letters, thus the natural language statistics still remain loud and clear in the cipher text. In theory you “just anagram them out”. Vigenere in it’s debased form is a substitution cipher with multiple shifted substitutions. Thus it’s a weak polyalphabetic cipher, and this shows in the ciphertext statistics. Polyalphabetic ciphers become much stronger when the substitution ciphers evolve over a period longer than the message. Depending on how they evolve they can become unbreakable as they become a variation of the OTP. However as WWII and the Enigma machine showed they can be broken if more than one message uses the same key, or there are weaknesses in the alphabet selection process or in the alphabets them selves (all of which it failed with in one way or another).

[1] According to some sources the language with the longest word was German. It had a word of over one hundred letters to do with EU food labelling rules. Unsurprisingly the word has been “deprecated in use”. But the reality like many long words it was a compound word of a number of smaller words. The important take away though is the human mind is limited in capabilities, and accurate spelling of words drops off quite quickly after around eight letters in a word. Some people mistakenly believe that the ability to accurately spell words of twelve to fifteen letters in the English language is a sign of inteligence (it’s not, any more than it is to have a good hand on macramé / macrame / macramie which can be correctly spelt in atleast three ways 😉

Greg October 18, 2019 6:56 AM

@ Clive Robinson

Of course using known sentences (from famous poems for example) has low entropy. This is true even if you use full sentences. There are about 160 billion words in the Google Books corpus, which means something on the order of 10 billion sentences. If you know I’m using a sentence from a book, then the space of possibilities is the same size as every string of seven lower case letters.

The strength of using sentences is that the number of sentences that have never been written (or even spoken) is and will always remain far, far greater than the number that have been recorded somewhere. And for that same reason, I don’t think we can reasonably draw conclusions about the initials of a random sentence from the insecurity of initials from a reused poetry verse.

(Though good point up above about using the second letter instead of the first. That’s a good way to increase security without making it less memorable.)

Clive Robinson October 19, 2019 4:25 PM

@ Tatütata,

Otherwise, her proportions often inspire me to call her “Botero” or “Barclay’s Bank”.

Your above words came back to me this morning.

I’d been over in the suburbs of East London early in the morning helping out a Scout Hall and Radio Club preparing for Jamboree On The Air (JOTA) by teaching someone how to “Foxhunt” down a hidden radio transmitter so they could show the scouts later in the day.

I was on my way back to the “dead tree cave” and due to “rail works” and a “political rally” I ended up going a different route through Waterloo Station.

As you come out of “Upper Waterloo” you jet met by this “Jumbotron wall” of advertising. Upon which was a Barclay’s Bank “blue” background no words but four video pictutes of cats at play just before it switched into a new advert.

So if I ever say “there are no such things as coincidences” just remind me gently 0:)

Who? October 19, 2019 7:08 PM

Sorry, last message was a reply to Clive Robinson. Missed the “@” at the beginning of the post.

Clive Robinson October 19, 2019 8:47 PM

@ Who?,

Yep. I joined Internet at the beginning of eternal september

Ahh the summer of 93, was a good time for me, for lets just call it “social reasons”.

a fact unrelated to AOL

Yes AOL, and their disks… A thousand land fills in their own right, and even the butt of a joke or two in SiFi cartoons.

Atleast the floppy disks were reusable, and I’ve still got a few hanging around. But the CD’s there are only so many “coffee coasters” and new age “Dream catchers” you can use. Hey even Turkey Farmer’s had more “peck mirrors” than they could tie up.

Even being into my third decade did not stop me succumbing to the temptation to use then like marshal arts throwing stars, to see if you could get them to explode[1]. One enterprising person I know built their own version of a “clay pigeon target launcher” for the CD’s to discover that they flew to fast to shoot down with a shotgun.

[1] There are certain materials polycarbonate being one that “store stress” and can release it all very violently in effect exploding outwards. Most people only see it with the likes of tempered glass (rather ironically also called safety glass). Especially that, which goes through reprated temprature stressing like some glass cookware or glass shower screen doors. Often something very very minor will trigger the stored stress in what looks and feels like a miniture explosion. Thus throwing AOL CDs “playing card style” against a wall you would very occasionaly get rewarded by one fragmenting into tens of major pieces and a myriad of glittering shards. These had edges like finely honed razors and points sharper than any needle, and you had to brush them up with care. Because if you got a tiny shard or flake under your skin it would stay there like a splinter or small thorn tip and would be just as annoying.'s_Drop

Clive Robinson October 20, 2019 5:24 PM

@ Anders,

Do you still use your Apple II?

Yes I still use it, rather more than I should, because it’s about 2/3rds my age, and I’m no spring chicken that’s for sure.

What everyday tasks is possible to do with it today?

The editor I use on it whilst limited in some respects, is still way faster than a whole bunch of modern IDE’s. I still “support” software I wrote on it in Pascal and Fortran, even though I’ve dropped broad hints to those using it that I and the Apple ][ have something in common in that neither of us has as firm a grip on the pearch as we once did and that sooner rather than later one or both of us are going to topple off. I also technically still support some hardware I developed for it as well, not that you can get the parts these days.

Oh and the serial card works well enough for me to run a terminal emulator on it and talk to a 486SX box that acts as a terminal concentrator when any other of the terminals plays up.

But very very occasionaly I fire up my first commercial product, a game of “Moon Defender” not unlike a certain other spave invader game. But unlike that game where the missiles came straight down, these missiles could come in at quite low angles.

Then there is the Sargon Chess game… So yes it does get used from time to time.

Clive Robinson October 20, 2019 9:53 PM

@ Anders,

Apple did not have ProDOS back in 1980, so it’s not “native” to my Apple ][ with the original drives.

So no I don’t use it, but I, guess I should look at it as a way of migrating people out 😉

Anders October 21, 2019 3:02 PM



I think i should find myself too now Apple II for playing 🙂
Back in ’80s behind the Iron Curtain it was unreal dream.
Now it’s even more than realistic.

Leave a comment


Allowed HTML <a href="URL"> • <em> <cite> <i> • <strong> <b> • <sub> <sup> • <ul> <ol> <li> • <blockquote> <pre> Markdown Extra syntax via

Sidebar photo of Bruce Schneier by Joe MacInnis.