Re: [ciphertool-devel] next steps

[Wart <wa...@ko...>]

On Wed, 2004-03-31 at 20:12, Alex Griffing wrote:
> Hi Mike,
> 
> To address the problem of redundant keys possibly generating conflicts, 
> maybe there could be a canonical key which doesn't contain any redundant 
> key information and then an auxiliary key that contains the other 
> information.  For example, for an aristocrat, it could be something like:
> 
> cipher1.type=aristocrat
> cipher1.author=alex
> cipher1.title=foo
> cipher1.ciphertext="..."
> cipher1.plaintext="..."
> cipher1.key="..." "..."
> cipher1.auxkey.subtype=k4
> cipher1.auxkey.keyword1="blah"
> cipher1.auxkey.keyword2="word"
> 
> I left out the ciphertext, plaintext, and key since I was too lazy.  
> Anyway, the only key information you need to encipher/decipher would be 
> in cipher1.key, but any additional information could be in 
> cipher1.auxkey.  Also I guess it would be good to be clear which values 
> get quotes around them and which don't.

This sounds good.  Let's use the "key" attribute to be the canonical key
attribute for all ciphers then.

Here is some clarification on the use of quotes in the key:

"Quotation marks must be used to delimit any part of a multi-part key
that contains spaces.  The quotation marks are optional if the
multi-part key part does not contain any spaces.  Quotation marks are
also optional for single part keys, such as the trifid, even if they
contain spaces."

I'll add the save file description along with the list of cipher names,
key descriptions, and any type-specific attributes to the ciphertool
documentation.

ciphertool's parser for the new format has already been written.  I'm in
the process of adding unit tests for the various corner cases now.

--Mike