Skip to content

AliceWonderMiscreations/Groovytar

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Groovytar

A pipe dream privacy centric replacement for Gravatar.

Live Demo: https://trippyid.com/

I am what some people would call anti-social. I am not sure what that term actually means, it seems to have different meanings depending upon the context, but generally I do not socialize. Except online.

I see tremendous value in socializing online, but I also see tremendous danger in the form of tracking.

There is value to a user having their identity "known" across different social social media platforms, but there is also value in that user deciding when it is obvious who they are and when it is not obvious who they are.

Just as an example, many people come from families where they would literally be cut-off if other family members knew they were gay, or knew they did not identify with the gender they were assigned at birth, etc.

For those people, the online community can be a godsend as it can allow them to express who they are and interact with people who accept them for who they are, but they need to be able to do so privately without the constant fear that they will be outed to family and friends who may be abusive towards them as a result.

The purpose of this project is to produce an open source replacement for the Gravatar system of globally recognized avatars that does a few things better.

Conceptually, I actually like Gravatar. I do. I am particularly fond of their ‘MonsterID’ avatars, those are really very well done and to me, they add some fun to commenting on blogs.

However there are two major flaws with Gravatar from a privacy perspective.

  1. The Gravatar is a simple md5() hash of the e-mail address. So for example, if your e-mail address happened to be bufonidaelover@gmail.com then the md5() hash is 5858a087b3f40b3876e804d5640222da. Anyone can write a bot to search the web for blogs where that hash is associated with the avatar on the post comment and know that you almost certainly made that comment. The government almost certainly already does this.
  2. The Gravatar domain uses tracking cookies. I do not know what they are doing with the data they collect, but it makes me feel very uncomfortable.

The Plan

John Doe uses the e-mail address johndoe345@example.org to post at BlogA, BlogB, and BlogC. Instead of using Gravatar where all three of those blogs would use ca84c27b3856cdaa92bcf90fcab687df to fetch an Avatar for John Doe, those blogs use Groovytar and the each have their own set of salts they use to come up with unique hashes that while generated from his e-mail address, they are all different and there simply is no way for anyone to know the hashes are related to each other.

To a bot scanning the web, they look like three different users at the three different blogs.

Those blogs can opt to have an API key with Groovytar that can allow John Doe to link the avatars used to his own image, if he so chooses. My Antifa friends who have WordPress blogs may not even want to have the API key simply because they don't want any of the people commenting on their blogs to be easily identified to an outsider as a real person, but my Animal Rescue friends who have blogs might decide to give their users that option.

When a blog has an API key, they can have a check box when submitting a comment that allows the person commenting to choose to have his avatar used at that blog linked to his or her real avatar.

The blog will take the user's e-mail address, hash it five times with sha512 and hash that result with ripemd160. That final hash will be sent as a unique identifier along with the salted hash in the comments and the blog's API key.

If John Doe does not have an account with Groovytar, nothing changes. Nothing changes because we have not verified that he actually wants any identity linking taking place, he hasn't finished opting in yet.

However if John Does does have an account with Groovytar (or creates one) then an e-mail will be sent to him telling him which Blogs want to link to his custom photo. He has to log in to Groovytar to specifically approve it.

Once he has specifically approved it, then what happens is the Groovytar sends an image of John Doe with requests for the salted hash specific to that site - if one exists - or generic hash associated with John Doe if one does not exist.

I may get fancy and also allow John Doe to specify a Mastodon instance account that the blog can link to with his name, haven't decided yet if I will do that.

So, John Doe will still have three different hashes used at blogs A, B, C but they may all fetch the same image if he has linked them together, allowing people to recognize him.

If in the future, John needs to unlink any of those sites from his real image, he can. For example, if Blog C is a Animal Right blog and John wants to apply for a job working at a sporting good store, John can choose to unlink Blog C and it will no longer serve his image along with the comments he has left at that blog.

That's the plan.

PictoGlyph

This will be the default generated avatar when another is not specified. It will provide a 4 x 4 grid of pictographs. There will be thirty-two different pictographs when finished, currently nine have been created.

With 32 pictographs to choose from on a 4 x 4 grid, that makes a total of 32^16 different possible unique combinations - which numerically is:

1,208,925,820,000,000,000,000,000

That's approximately the same number of stars as there are estimated to be in the known universe.

Additionally there will be 128 different high contrast color combinations. However I don't count them in the entropy for possible combinations because they may not provide the same level of visual distinguishability for all people.

The basic nutshell idea is that when looking at a post, the color combination in concert with the presented glyphs will be enough to act as a sort of digital signature we can visually identify as associated with a particular person.

Many identicon schemes exist. Most just translate a hash string into simple geometic shapes, for example:

Gravitar Identicon

I believe that creates confusion in the human mind rather than identity. It might as well just be an image of the hash itself.

By using a grid of pictographs that represent human ideas, I believe it triggers familiarity within the human brain making it easier to identify the avatar with a person, the same way a picture of the person would - or their signature on a piece of paper.

Currently the following pictographs exist:

  • Triquetra Knot
    Irish / Keltic Origin. Symbolizes earth, air, and water or life, death, and rebirth. Also symbolizes the Triple Goddess - Maiden, Mother, Crone.

  • Fertility Rune
    Celtic Fertility Rune.

  • Yin and Yang
    Opposite and often seemingly contrary forces are often actually complimentary to each other.

  • Asase Ye Duru
    West African symbol for the divinity of Mother Earth.

  • Elvin Star
    Also called a Faerie Star. Basically an equilateral seven pointed star. It is found in many different historic and some modern cultures, Seven seems to be significant number to a lot of people groups.

  • Hawaiian Turtle
    Symbolizes good luck and the importance of ocean life.

  • Boa Me Na Me Mmoa Wo
    West African (Adinkra), literally means "Help me and let me help you" - it symbolizes cooperation and interdependence, which is at the very core of how I personally define socialism.

  • Taino Coqui Frog
    I have been able to find what this glyph meant to the Taino people, but it is very common in their historic art. It clearly meant something to them. And it is a very beautiful reminder to me of the importance of ecological diversity. The Coqui frog is not yet extinct, but it is endangered like so many frogs. Hopefully it can continue on its evolutionary path.

  • Native North American Indian Sun Glyph
    I am not sure what tribe or tribes used this glyph, but it is a simplistic yet very beautiful depiction of the Sun, from which all life draws energy.

  • Noongar Waugal
    In Noongar Dreamtime Mythology, it was a Waugal that created the major rivers in their part of Australia.

  • Neo-Druid Awen
    Awen is a Celtic symbol showing three rays. The Neo-Druid version of this symbol often has three circles on the outside containing three dots and three rays, and is frequently used to represent masculine energy, feminine energy, and the balance between them.

  • Zuni Bear Fetish
    The Bear fetish is the Guardian of the West and has the power to heal and transform human passions into true wisdom. Bear reminds us that one of the great powers we have is the power of turning to solitude and introspection through which we integrate new experience and change.

There will eventually be 32 glyphs.

Small Size Variant

A four by four grid of pictoglyphs really needs at least 120x120 to display well. Not yet implemented, but planned is a variant than is three by three instead of four by four that will work better for social media sites that only want really small avatars.

That only will provide 36^9 (just over 100,000,000,000,000) different unique combinations, but it still makes collisions extremely unlikely.

Lots To Do

The Confeti identicon generator is working.

The Pictoglyph identicon generator has some of the planned glyphs but still needs more glyphs created as well as more known good color combinations added.

I need a few identicons too, and I will need funding for hosting this project.

In addition to the Confeti generator, I want one that generates stylish amphibians based upon the hash. Frogs, Toads, Newts, Salamanders, maybe even Caecilians.

For websites migrating from Gravatar, these would be used in place of the Wavatar generated avatars.

I also would like a replacement for the MonsterID identicons.

It also might be fun to have one that generates Gingerbread Cookie identicons.

And some others.


EOF

About

A pipe dream privacy centric replacement for Gravatar

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages