Sounds like good ideas.
Sounds like good ideas.
While searching for examples, came across a template for browser extensions. It's the only one that I tried that worked with no errors, and has all the features I want. Hopefully it will save me a lot of time: https://github.com/davidnguyen179/web-extension-svelte-boilerplate
I was able to get it compile out of the box and run in both chromium and firefox.
It's using a framework called svelte which seems pretty cool. I took a look at it and it's a lot easier to use than react and vue and should be faster and less bloated because it generates javascript at compile time rather than runtime like react/vue does.
So today, I went through the tutorial on svelte, and I think this is what I'll use for the extension frontend if nothing goes wrong. I think by learning this framework will be the fastest way for me to create a prototype.
I again didn't get to work too much again, unfortunately, due to having to do some errands around the house.
I was thinking about adding the ability to know how many comments were posted for a website even before clicking on the extension, by making the total show up in the extension icon.
One problem is that for privacy purposes, you don't want the extension looking up every url the user goes to against the server. So, only when the click the extension button do you want the client to ask the server about an individual url.
In order to solve this, I came up with an idea. First the extensio would hash the url using a secure hash algorithm, such as SHA-256, then send, lets say, the first 2 bytes of the hash to the server. This would reduce the number of possible urls that match the hash on average to 1/(256*256) = 1/65536. Then, if the resulting list is small enough, the server could just send all the matching urls along with a total comment count for each. If it's not small enough, then that server would just ask for another byte. By doing it this way, there is no way to determine what url the user went to with only 2 bytes of a sha-256 hash. I mean, literally the client would be sending a string such as "3f" to the server, and there is no way to guess what url the user went to from this.
I was also thinking of how to integrate with websites that want to sign up easily. Using the plugin, I can open up any url, including the cookies that are active. So all the website would have to do put a page in the logged-in area of the website that signs the brower-extensions pub key for the user, along with the username and maybe an icon or something. If the user is not logged in, the extension can redirect them to a login page, or open it in a popup window or something and try again afterwards. Then the backend servers for captaindirgo would be able to verify that the user had signed into the site.
Although, I need to check if there are more standard ways to do this, because if there are prebuilt plugins that work with the website, it would make the website owner and my life a whole lot easier.
For the backend servers, I think I want to make them completely separate from the frontend servers. So if you run a website, you can allow your userbase to use the captain dirgo plugin by only building out the signing page. The storage of comments would happen automatically. And if you want to help out by running a backend server, you can do that, too, by running a completely different application, that connects to a p2p network for the captain dirgo backend.
Additionally, anyone could run a backend application without running a website if they wanted to, too.
But that's just my thinking at the moment and could change. My first priority is to get a prototype done, so I can get some feedback on the UI.
While searching for examples, came across a template for browser extensions. It's the only one that I tried that worked with no errors, and has all the features I want. Hopefully it will save me a lot of time: https://github.com/davidnguyen179/web-extension-svelte-boilerplate
I was able to get it compile out of the box and run in both chromium and firefox.
It's using a framework called svelte which seems pretty cool. I took a look at it and it's a lot easier to use than react and vue and should be faster and less bloated because it generates javascript at compile time rather than runtime like react/vue does.
So today, I went through the tutorial on svelte, and I think this is what I'll use for the extension frontend if nothing goes wrong. I think by learning this framework will be the fastest way for me to create a prototype.
I again didn't get to work too much again, unfortunately, due to having to do some errands around the house.
I was thinking about adding the ability to know how many comments were posted for a website even before clicking on the extension, by making the total show up in the extension icon.
One problem is that for privacy purposes, you don't want the extension looking up every url the user goes to against the server. So, only when the click the extension button do you want the client to ask the server about an individual url.
In order to solve this, I came up with an idea. First the extensio would hash the url using a secure hash algorithm, such as SHA-256, then send, lets say, the first 2 bytes of the hash to the server. This would reduce the number of possible urls that match the hash on average to 1/(256*256) = 1/65536. Then, if the resulting list is small enough, the server could just send all the matching urls along with a total comment count for each. If it's not small enough, then that server would just ask for another byte. By doing it this way, there is no way to determine what url the user went to with only 2 bytes of a sha-256 hash. I mean, literally the client would be sending a string such as "3f" to the server, and there is no way to guess what url the user went to from this.
I was also thinking of how to integrate with websites that want to sign up easily. Using the plugin, I can open up any url, including the cookies that are active. So all the website would have to do put a page in the logged-in area of the website that signs the brower-extensions pub key for the user, along with the username and maybe an icon or something. If the user is not logged in, the extension can redirect them to a login page, or open it in a popup window or something and try again afterwards. Then the backend servers for captaindirgo would be able to verify that the user had signed into the site.
Although, I need to check if there are more standard ways to do this, because if there are prebuilt plugins that work with the website, it would make the website owner and my life a whole lot easier.
For the backend servers, I think I want to make them completely separate from the frontend servers. So if you run a website, you can allow your userbase to use the captain dirgo plugin by only building out the signing page. The storage of comments would happen automatically. And if you want to help out by running a backend server, you can do that, too, by running a completely different application, that connects to a p2p network for the captain dirgo backend.
Additionally, anyone could run a backend application without running a website if they wanted to, too.
But that's just my thinking at the moment and could change. My first priority is to get a prototype done, so I can get some feedback on the UI.
(post is archived)