search

dsw / proquint

Proquints: Identifiers that are Readable, Spellable, and Pronounceable.
http://github.com/dsw/proquint
Other
289 stars 22 forks source link

Repeated consonants in multi-word proquints are hard to enunciate #6

Open DanielHeath opened 7 years ago

DanielHeath commented 7 years ago

From https://arxiv.org/html/0901.4016 : 

127.0.0.1       lusab-babad
63.84.220.193   gutih-tugad
63.118.7.35     gutuk-bisog
...
216.68.232.21   todah-vobij
198.81.129.136  sinid-makam
12.110.110.204  budov-kuras

The first entry in this list is much more difficult to say than the others because the b consonant is repeated.

How badly would the bitrate be affected by selecting the first consonant from a shortened list?

dsw commented 7 years ago

The design is public so you can check this yourself. Frankly I think it is a fools errand. How many more properties do you want?

On Sun, Dec 18, 2016 at 8:59 PM, Daniel Heath notifications@github.com wrote:

From https://arxiv.org/html/0901.4016 :

127.0.0.1 lusab-babad 63.84.220.193 gutih-tugad 63.118.7.35 gutuk-bisog ... 216.68.232.21 todah-vobij 198.81.129.136 sinid-makam 12.110.110.204 budov-kuras

The first entry in this list is much more difficult to say than the others because the b consonant is repeated.

How badly would the bitrate be affected by selecting the first consonant from a shortened list?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/dsw/proquint/issues/6, or mute the thread https://github.com/notifications/unsubscribe-auth/AAC_XF468auyGGYgCkwFhkWOO-RP_Cgdks5rJg8ggaJpZM4LQX4K .

DanielHeath commented 7 years ago

I apologize; clearly I've bothered you, which couldn't be further from my intent.

This is such a cool piece of work that it's easy to get excited about further developments & seeing how far it could be taken.

I think some simple modelling of which adjacent consonants are difficult to pronounce might result in a small loss of bitrate for much clearer enounciation.

dsw commented 7 years ago

You have not really annoyed me, it's just that one of the properties of the system is that the encoding is very simply reductionistic: (1) you chop the bits into pieces and (2) map them to letters. You get both (a) a decent bit rate and (b) something guaranteed to be pronounceable. I find the combination of these properties to be rather remarkable.

You could wish for more, but the encoding needed to get it will immediately be much more complex to the point where I think you will end up with basically an entirely new encoding. That will be an entirely new project, rather than a version 1.1 of this one. Let me know if you manage to pull that off.

Daniel

On Thu, Dec 22, 2016 at 4:15 PM, Daniel Heath notifications@github.com wrote:

I apologize; clearly I've bothered you, which couldn't be further from my intent.

This is such a cool piece of work that it's easy to get excited about further developments & seeing how far it could be taken.

I think some simple modelling of which adjacent consonants are difficult to pronounce might result in a small loss of bitrate for much clearer enounciation.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/dsw/proquint/issues/6#issuecomment-268919019, or mute the thread https://github.com/notifications/unsubscribe-auth/AAC_XN_diwyIt7m0DiPauykGGGsbZWDAks5rKxK6gaJpZM4LQX4K .