Baconian - courtesy of Sir Thomas by avastmick · Pull Request #35 · arosspope/cipher-crypt

avastmick · 2018-03-20T06:59:55Z

This was a fun one!

Under the hood, it uses a simple substitution cipher, mapping alphabetical characters to 5-bit binary equivalents - so 'a' maps to AAAAA, 'b' to AAAAB and so on.
There are two code mappings, traditional that maps 'I' and 'J' to the same value, and 'U' and 'V' to the same. This was due to the fact in Elizabethan England, these were variants of the same character and not distinct, as they are today. I also added a distinct mapping that has a full alphabet mapped to discrete values.

Next, the fun part: hiding the produced substitution ciphertext in a decoy text. I do this by mapping alphabetical characters to italics if the binary is given. I use the mathematical italic utf-8 characters. There maybe other fonts or characters that could be used, if so the code would need to be modularised.

Please have a look at the code and check for newbie Rust issues and feedback for me to fix.

arosspope · 2018-03-20T08:14:17Z

You're a legend for working on this, thanks for all your help. I might not get time to review it this week, but I will get around to it when I can.

avastmick · 2018-03-20T12:30:40Z

Okay, great. In the meantime, I'll have a look at the code too, I think can already spot some improvements.

arosspope · 2018-03-29T05:57:52Z

+//! Each character of the message plaintext is encoded as a 5-bit binary,
+//!  these are then "hidden" in a decoy message through the use of font variation.
+//!
+//! This cipher is very easy to crack, once the method of hiding is known, therefore this


I would suggest changing this to:

"
Each character of the plaintext message is encoded as a 5-bit binary character. These characters are then "hidden" in a decoy message through the use of font variation. This cipher is very easy to crack once the method of hiding is known. As such, this implementation includes the option to set whether the substitution is distinct for the whole alphabet, or whether it follows the classical method of treating 'I' and 'J', and 'U' and 'V' as interchangeable characters (as would have been the case in Bacon's time).

Additionally, this implementation allows the user to change the underlying binary character choice. This is traditionally 'a' and 'b', but the user can choose any pair of characters. If no concealing text is given and the boilerplate of "Lorem ipsum..." is used, a plaintext message of up to 50 characters may be hidden.
"

Done. (with amendments...)

arosspope · 2018-03-29T06:01:37Z

+
+/// This struct is created by the `new()` method. See its documentation for more.
+pub struct Baconian {
+    distinct: bool,


Could you please explain to me the purpose of the distinct? This variable name does not seem intuitive to me...

How about non_traditional_dictionary? I'd like to keep the default behaviour as traditional as that is how the original cipher was used.

How about: use_distinct_alphabet?

Okay will sort this. Rammed busy right now, but will get on to it this week.

arosspope · 2018-03-29T06:04:40Z

+///     - as they had equivalent value in Bacon's day
+///     - generally, 'I' and 'V' were more common and used here
+lazy_static! {
+    static ref TRAD_CODES: HashMap<&'static str, &'static str> = hashmap!{


For TRAD_CODES and DISTINCT_CODES hashmap, there seems to be a lot of overlap in the data stored. Surely there is a way we only need to define this information once?

Agreed, I'll change it to that the outcome for the TRADITIONAL mapping is met, but using a DISTINCT codeset.

arosspope · 2018-03-29T06:06:37Z

+    fn new(key: (bool, Option<String>)) -> Result<Baconian, &'static str> {
+        Ok(Baconian {
+            distinct: key.0,
+            decoy_text: key.1.unwrap_or_else(|| String::from(DEFAULT_DECOY)),


Rather using a static default decoy text, perhaps we could randomly generate a string as the default decoy? I know this might diverge from other implementations but I think it would be a cool feature.

I need to look into that. Easy thing in Python, not sure about Rust, maybe there is a crate that could be used. I agree, it would be cool, but I'll park this for now. Maybe make an issue for it..?

It might be not that hard to integrate, see the crate: https://github.com/mgeisler/lipsum. However I understand if you want to leave this out for now, we can create an issue.

Will have a look. Will flag whether it would be better as an issue. It would certainly be interesting, and allow the default always to manage the size of the plaintext

Done.

Used the crate you advised. However, still a good idea to open an issue as the code is not very elegant as it just assigns an arbitrary default decoy text of 160chars long... So, could be better with a bit more time and thought...

arosspope · 2018-03-29T06:10:17Z

+    fn encrypt(&self, message: &str) -> Result<String, &'static str> {
+        let mut non_alphas = 0; // A counter for non_alphas
+
+        for c in self.decoy_text.chars() {


You could tricky with iterators here... something like:

self.decoy_text.chars().for_each(|c| {if !c.is_alphabetic() {non_alphas += 1}});

I don't know what the exact call would look like but it's worth exploring if you haven't done so already.

Okay, I'll look into it. Personally, I'm not a great fan of (pythonic) idioms like this as it makes code readability worse imho. I get it for interpreted languages where concise code gives performance gains, like python and Javascript, but compiled languages, it just makes for hard to maintain code. </rant> 😄

From a learning perspective, though, it's good to use, so when I see it I'll be able to read it (maybe)...

Lunchtime look: your snippet is sound and does the same as original. But, actually not gaining anything there - rustfmt will reformat over the same number of lines... and don't actually make any savings.

From a learner perspective, I can see what you are thinking. I'm sure I (we) can find a better example to learn that syntax.

Sure. I don't feel to strongly either way and I haven't really used streams extensively in this project, so we can keep it as you have for now.

arosspope · 2018-03-29T06:10:58Z

+        for c in decoy_slice.chars() {
+            if c.is_alphabetic() {
+                match secret.remove(0) {
+                    'B' => {


Some comments would be nice to explain what is happening here

arosspope · 2018-03-29T06:12:33Z

+#[macro_use]
+extern crate lazy_static;
+#[macro_use]
+extern crate maplit;


Are you still using this crate? If so, where?

Yes, I use both these in the initialisation of the static Hashmap constants for the code look ups.

arosspope

Hey @avastmick, I've added some questions and suggestions throughout that I would like to discuss before merging the PR. Overall this is a pretty cool cipher.

avastmick · 2018-03-30T01:19:54Z

Thanks @arosspope, I'll look at this and comment / make changes after Easter. Some good thinking.

avastmick · 2018-04-02T05:11:44Z

Had to change .travis.yml:
No longer require nightly to run rustfmt - note that upstream in nightly rustfmt the formatting rules have changed 🙄 so the current code will have large-scale changes when that is merged into stable.

avastmick · 2018-04-02T05:15:04Z

That's me for the week, I'm afraid. Anything else, shout out and I'll look at them early next week. In particular, the autogen of text - as you said fun, but I only had a couple of hours today...

PS: @arosspope Not sure what editor you prefer, but integrating in rustfmt on save is easy. I'm old (not school, just old) so I still use vim. It was easier than integrating racer that's for sure. 😉 🍷

arosspope · 2018-04-09T08:48:11Z

+    //  for traditional usage.
+    let mut key_upper = key.to_uppercase();
+    if !distinct {
+        match key_upper.as_str() {


Perfect, this is what I had in mind 👏

arosspope · 2018-04-09T08:56:18Z

@avastmick Almost ready to merge, just two points that I would like to discuss before that happens. Let me know what you think.

In terms of text editor, I usually use Atom. I'm pretty sure it can integrate rustfmt... I should check that out, thanks. 😄

avastmick · 2018-04-16T02:36:43Z

I covered you two changes and raised an issue to improve the usage of the lipsum crate.

I may drop off radar for a bit as I have a heap of work changes coming up.

Cheers, Mick

avastmick added 3 commits March 15, 2018 10:10

Set up for baconian cipher coding

c1e6fee

Logic smash in, horrible code to clean up

e2ce31f

Added Baconian cipher - initial for review

7e381e6

Added test to handle issues with traditional encoding

82bb38c

arosspope reviewed Mar 29, 2018

View reviewed changes

arosspope requested changes Mar 29, 2018

View reviewed changes

avastmick added 3 commits April 2, 2018 12:20

Stash initial fixes

71de3d2

Clean code for PR, add test for trad vs dist

114fa22

Changed rustfmt to stable using rustup

97a2e0d

arosspope reviewed Apr 9, 2018

View reviewed changes

glarsen mentioned this pull request Apr 10, 2018

Implement the ciphers outlined in the README #2

Open

Using the lipsum crate to autogenerate the default decoy text

ed750bf

avastmick mentioned this pull request Apr 16, 2018

Inelegant use of lipsum text generator in Baconian cipher #37

Open

arosspope merged commit 83d0852 into arosspope:master Apr 18, 2018

avastmick deleted the baconian branch April 18, 2018 01:06

Conversation

avastmick commented Mar 20, 2018

Uh oh!

arosspope commented Mar 20, 2018

Uh oh!

avastmick commented Mar 20, 2018

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

arosspope Mar 29, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

arosspope left a comment

Choose a reason for hiding this comment

Uh oh!

avastmick commented Mar 30, 2018

Uh oh!

avastmick commented Apr 2, 2018

Uh oh!

avastmick commented Apr 2, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

arosspope commented Apr 9, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

avastmick commented Apr 16, 2018

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

arosspope Mar 29, 2018 •

edited

Loading

avastmick commented Apr 2, 2018 •

edited

Loading

arosspope commented Apr 9, 2018 •

edited

Loading