Guillermo Rauch: MyDB

all right thanks for the introduction that was kind of embarrassing but it's all true so i'm at route g on twitter and everywhere else soon i i actually migrated to rawajtv.com i'm kind of obsessed with being uh rauchi everywhere now i'm a gishe on github grouchy on twitter sunrise.com so uh the cool thing about socket io is

that it kind of goes back a very long time like the reason that i got interested in real-time collaboration uh happens to be that i wanted to enable real-time document collaboration in the classroom so we thought one of the best applications for sending data back and forth really fast is actually helping teachers communicate with students and we came this idea we came up with this idea for something called rosepath

uh very similar to uh etherpad and google wave they were all in the same like um i guess uh fundamental idea and that's soccer it actually became the transport for that it was it was kind of like uh going on a tangent back then and we ended up changing our idea a bunch of times a lot of open source software was created that explains why everything is under the learn boost uh github handle um and then we actually came up with an

idea for something even simpler a very easy way of sharing anything uh that could be basically rendered on a web browser so you would give us a file and we would process it on the backend and then we would send real-time events about how we transform those files uh to the browser in this case we find ourselves sort of back at where we started because we're working a lot on media management media processing and real-time text collaboration

and what i wanted to talk about today is precisely how rich text collaboration is a very hard problem with very uh specific decision trees that you have to make in terms of how you implement it that have an impact on everything so the way that you implement your wysiwyg editor has a very uh it has very direct implications on how you're going to collaborate in real time so that's basically what i'm going to talk about today you guys probably

remember projects like fck editor or tinymce so you probably already know that uh real time rich text editing is really hard and then real time rich sex editing just like adds a huge other layer of complexity or you could you could look it as simplicity too uh bridge text editing in real time is a very very hard problem and the best way to tackle it is by introducing uh constraints into the

system so um before i go into the details of how i would implement real-time rich sex collaboration i want to basically set some foundations about our use case and what features and what capabilities we expect out of our rich text editor so we we started with the fundamental idea that what we create ideally is mobile first but essentially the same code base should be able to run on different viewports with

different modes of interaction namely touch and keyboard input considering again that the ui might vary but we want to enable this real-time richtex capabilities on every device some of the ui capabilities we want are pretty straightforward and common like canvas and color links text that could be in different layouts like tables or lists something we call the arbitrary blocks which could be images or galleries