Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Always Bet on Text (2014)

graydon2.dreamwidth.org

Post reply on HN
Print view View original

41–50 of 349 posts

Re: Always Bet on Text (2014)

#41

Post by cycomanic »

I think the post conflates several quite unrelated concepts under the label text. Also what does he mean by text and information? For example he mentioned the "optical telegraph", which is a semaphore system which used a system of messages which AFAIK were not alphabetic text.

Another example if we compare logographic language systems (e.g. Chinese characters, hieroglyphs ...) to alphabetic systems. I guess we can say both are text, but logographic systems are close to pictures as well (pictographic systems even more so) and they can convey information in much less characters, at the cost that one needs to know a much larger "alphabet".

Similarly if the information we are trying to transmit is actually an image, transmitting the actual image is certainly much less information than transmitting a description of the image. That is also why the comparison with the bird and twitter image falls flat. The image conveys a specific image, not the generic image of a bird. To describe that specific would need many more characters than 4. Similarly if I have a large set of numbers, it's much more efficient to store and transmit in binary format not as text.

Re: Always Bet on Text (2014)

#42

Post by bawolff »

necovek wrote: post #16
virgilp wrote: post #8

I would like to add that text, as a thousand-years-old-technology, has always included illustration too. Even in recent/digital text, you have this bird that doesn't take 4k bytes: [edit: unfortunately hackernews doesn't display the unicode character U+1F426] Text is very powerful, but it doesn't need to always work by itself (arguably works best together with other media)

While I totally agree with the OP, the point to remember is that text is a really terse storage as long as you have the Unicode mapping already stored to transform the binary compressed form to actual pictures of text (vs abstract forms or numeric codepoints when simply stored as "textual" bytes). This does not diminish the value and expressiveness of text, but it needs to be said that in 5000 years time we'll need b…

I dont think you would. If english is still understood, decoding ascii/utf-8 is trivial. Just think of it as a ceaser shift cipher with an offset of 64. Very trivial to decode with frequency analysis if you know its english. Either you understand that 65 represents the abstract concept A, or you have no idea what the letter A is. Either way, a picture of the letter A is not going to be helpful.

Having english still understood after 5000 years seems much harder. But hey, latin is pretty old and we still understand that.

Re: Always Bet on Text (2014)

#43

Post by bawolff »

ElectricMind wrote: post #30

Actually a single letter is "a picture". So Text is a collection of small pictures. There is no such thing as "Text"- we give a meaning. Otherwise it is bunch of weird arrows and curves - pictures.

Alphabetic glyphs and pictures are not the same thing. Glyphs are abstractions. Pictures are literal (ironic).

Re: Always Bet on Text (2014)

#44

Post by cycomanic »

dwb wrote: post #35
ElectricMind wrote: post #30

Actually a single letter is "a picture". So Text is a collection of small pictures. There is no such thing as "Text"- we give a meaning. Otherwise it is bunch of weird arrows and curves - pictures.

A letter is not a picture, at least not in the way we generally talk about pictures – it can be graphically manipulated and re-imagined in almost limitless ways and degrees and still perform its function. For most things-that-we-call-pictures, if you even made a relatively small adjustment to it, it would be a different picture.

That really depends on your language system. What you say is true for alphabetic systems, but quite different for logographic (e.g. Chinese) or pictographic systems.

Re: Always Bet on Text (2014)

#45

Post by bawolff »

msiyer wrote: post #17

If we agree that a picture is worth 1000 words, then... Given the current state of information technology, I agree that we are most efficient at processing text. However, that can change pretty quickly. Storage mechanisms similar to DNA can make the difference between text and multimedia irrelevant. It will happen because nature already does that.

Storage hasnt really been the bottleneck for a long time. A terabyte stores a lot of pictures.

Re: Always Bet on Text (2014)

#46

Post by majewsky »

necovek wrote: post #16
virgilp wrote: post #8

I would like to add that text, as a thousand-years-old-technology, has always included illustration too. Even in recent/digital text, you have this bird that doesn't take 4k bytes: [edit: unfortunately hackernews doesn't display the unicode character U+1F426] Text is very powerful, but it doesn't need to always work by itself (arguably works best together with other media)

While I totally agree with the OP, the point to remember is that text is a really terse storage as long as you have the Unicode mapping already stored to transform the binary compressed form to actual pictures of text (vs abstract forms or numeric codepoints when simply stored as "textual" bytes). This does not diminish the value and expressiveness of text, but it needs to be said that in 5000 years time we'll need b…

> in 5000 years time we'll need both the Unicode specification and those 2-4000 bytes to decipher the author's post.

Nitpick: In all likelihood, an English dictionary would be enough. Even if the Unicode spec is lost, the text can probably be deciphered by using frequency analysis plus the dictionary to associate codepoints with characters.

Re: Always Bet on Text (2014)

#47

Post by cycomanic »

necovek wrote: post #10
samwestdev wrote: post #5

Except nobody really reads text anymore. If you want to deliver a powerful message pics and vids have the most bang for the byte.

That's a really surprising sentiment to me. I despise videos when I want to learn something. Give me a nice text (+ pictures if a subject requires it) so I can follow at my own pace: slow down, skip or skim as needed.

For me it depends a lot on what I want to learn. For some things (e.g. programming) text is much better.

For other things (e.g. repair a part on a bike) text requires images, to really be able to quickly grasp what one has to do.

Finally, for learning e.g. dance moves I really want a movie (or a live instructor).

So like usual "it depends".

Re: Always Bet on Text (2014)

#48

Post by user-the-name »

Ygg2 wrote: post #6
chordalkeyboard wrote: post #2

> Text is the most efficient communication technology. By orders of magnitude. This blog post is likely to take perhaps 5000 bytes of storage, and could compress down to maybe 2000; by comparison the following 20-pixel-square image of the silhouette of a tweeting bird takes 4000 bytes: Author is really excited about really old information technology. He makes some good points.

It's solid technology to be excited about. For example. Graphical programming languages are nowhere. Closest I had was either VHDL or backed by copious amounts of XML like Xtend. And both of those have a textual Component as well.

If you think graphical programming languages are nowhere, that just means you do not have a very broad exposure to programming.

As was already pointed out, graphical languages are massive in game design. They are also used in audio processing, in scientific and engineering software through labView, and in various other places.

They are very popular in fields where we want to open up programming to more people than those who call themselves "programmers".

Re: Always Bet on Text (2014)

#49

Post by user-the-name »

goatlover wrote: post #21
zokier wrote: post #12

Earlier quoted context omitted.

Most programming languages would be better characterized as tree structures than text, being only obscured by unclear (but no less rigid) syntax.

So Lisp? You still type it in as text.

Nearly every language will be processed as an AST, not as text.

Text is just a serialisation format for the AST. For some reason we still insist on editing the serialisation format instead of the actual data representation.

Re: Always Bet on Text (2014)

#50

Post by DanBC »

montebicyclelo wrote: post #13

> text is the most powerful, useful, effective communication technology ever, period. On the other hand, many (technical/mathematical) concepts are more effectively explained using diagrams/images. E.g. the very visual approach taken in Mathologer videos [1] makes difficult, esoteric, mathematical ideas accessible to a wider audience. [1] https://www.youtube.com/watch?v=iJ8pnCO0nTY

Yes. See also Florence Nightingale using rose diagrams to illustrate preventable excess deaths to infection during war.

https://www.florence-nightingale.co.uk/coxcomb-diagram-1858/

Post reply on HN