Hydraulic Telegraph | Hacker News Reader

Hydraulic Telegraph

en.wikipedia.org

38 pointstikwidd27 comments

Maybe I'm missing something, but with the Greek system why did they need to use draining water as a time-keeping method? Why not, after some acknowledgement sequence, does the signaler not simply wave the touch a specific number of times which the receiver could use to correlate with a predetermined list of possible messages?

For example, signaler holds touch aloft continuously. Receiver notices and does the same to acknowledge. Signaler now waves the touch seven times. Receiver checks chart for which message 7 waves correlates to. Is that not effectively the same as the described method in the article but with less room for error due to timing?

You also wouldn't need to keep the devices themselves calibrated, as if one drained faster than the other, you'll have a problem. You only need a mutually agreed upon chart of messages which is already a requirement with the water method.

Basically, semaphore. Why not just semaphore? What is the advantage to their method? Is the timing method maybe easier if you were dealing with, say, 50 messages on the vessel? As opposed to coming up with a way to encode 50 different messages?

Light on or off is a clear signal over a much longer distance than a light moving only a few feet. A simple candle can be seen a very long distance on a dark night but just moving it back and forth gets lost. There were no means of visual magnification available which is what makes semaphore possible over usable distances.

So, instead of waving the torch in the open you wave it next to a wall so it appears to go on and off. Same question.

I thought the same as well. Clearly the ancient Greeks weren't morons so there's probably a reason they did it the way they did. Thinking about it more, their system is probably much easier for a novice to use, or some random soldier grabbed for signal duty. There's no need to count and keep track of torch waves. The synchronization protocol seems pretty simple and the process afterwards is equally simple.

Hell, your soldier doesn't even need to know how to count, if that was even a problem for the Greeks.

Now we need a wall on this random hilltop we're signaling from too. Fundamentally, a simpler signal is easier to decode and this only relies on sending two signals start and stop to encode many signals.

In the Roman version, the hydraulic part is a clock for generating and measuring signals having a specific duration.

Yes. In a time before we really knew how to accurately measure short periods of time flowing water was one of the few ways humans had to meter out time.

In theory a pendulum with the same weight and length could have served the same purpose with some limitations on the number of messages it could potentially convey.

/Roman/Greek/ doh!

Reminds me of the paris pneumatic clock system that would be built 40 years after the british telegraph. https://news.ycombinator.com/item?id=19441300

pneumatics are still used in organs today! (Not new ones, but old still deployed ones still have them)

https://upload.wikimedia.org/wikipedia/commons/5/51/Salemer_...

https://en.wikipedia.org/wiki/Tubular-pneumatic_action

This made me recall reading about Douglas Engelbart's project from SRI (whence the computer mouse originated) that used a 5-key chording keyset that had pneumatic actuators for using the keys to output a data stream to the user haptically (from my reading it was unclear if this functionality was ever intended to be put in the mouse as well, which was used in the opposing hand). I wonder if anyone is aware of any existing modern (DIY?) projects to bring such a thing back to life?

Could have used the aqueduct as a telegraph as early as 7th century BCE - put a written message in the equivalent of a rubber duck, drop it in. A spotter on the other end dips it out and delivers the message. How could no one have thought of that?

Maybe they wanted to send messages from one hilltop to another, where there was no aqueduct. Also the system described may be low bandwidth but it's also low latency. Would be useful to relay messages of a security/defense nature.

Could not much the same effect have been achieved by pulling on a piece of string?

You'd have to make a very long piece of string and avoid it snagging, getting cut, or breaking under it's own weight. Also along that distance there's going to be a lot of slack in the system so you'd have to take that up too.

That would only give you one-directional communication, while the hilltop-to-hilltop systems described in the article would work both ways.

Still one-directional communications is useful. Think ordering by mail from Sears.

I was just about to ask HN to come up with some communication devices using only ancient technology.

Although your rubber duck system would require less manpower than a system of signal fires, it would be MUCH slower, and probably even slower than the chains of horse riders that delivered messages Pony Express-style across the Roman Empire. In an emergency, a message could be transferred over 100 miles a day.

https://en.wikipedia.org/wiki/Cursus_publicus

I imagine it would be useful for some information. Crop reports etc for helping predict market fluctuations. All aqueducts lead to the city after all! Where the markets are.

Anyway its nearly free, and the infrastructure was always in place. I'm just surprised it didn't get used.

Oh! And for your ancient-networking subject, I submit:

  * grow trees in rows; cut them in ascii sequences (a stump is a 0). Visible for miles!
  * scratch dated messages in turtles' shells; sample shells continuously and collate
  * spread gossip where the subject is a code e.g. marital infidelity means one thing, sexual preference another. Collect gossip and reconstruct. Use political topics for ack and nack. 
  * Grow crops each year to encode the message: corn is a 1, soybeans a 0. Study crop market reports. Caution: baud rate is low and noise level high, recommend using ECC
  * Become a fashion icon. Adjust skirt lengths microscopically and seasonally to encode multiple bits! A plus: observation and measurement has side benefits

Let me know if any of these suit your purposes. I release them to the public domain.

Friendly reminder that the code block style is basically unreadable on mobile (and not so great on desktop either) because it doesn't wordwrap.

Dunno how to do lists

* Just put an asterix or something

* Don't indent by more than one space

* Hackernews doesn't support proper list formats or really anything beyond italics, urls, and codeblocks.

Or just implement an appropriate RFC[1] :)

[1] https://tools.ietf.org/html/rfc1217

The ancient Greeks beat this in terms of slow delivery:

"In 499 BC, he shaved the head of his most trusted slave, tattooed a message on his head, and then waited for his hair to grow back. The slave was then sent to Aristagoras, who was instructed to shave the slave's head again and read the message, which told him to revolt against the Persians"

https://en.wikipedia.org/wiki/Histiaeus

Need a stable of slaves with different tattoos. Or maybe just two: revolt, or wait for next round. Or put a numbered list on one, and send the next with just a number.

Not even sure it's still available online, but you need the dead media project's working notes.

The problem with rubber-duck aqueducts is that they so rarely go where you want to go.

Reply on news.ycombinator.com