2023 was a fairly successful year for R18.dev. At the start of the year, the website didn't even have a homepage. Now there is not only a homepage, but also database dumps, social embeds, as well as a new colour scheme. A lot of bugs have been fixed, and processes have been automated. Some components are now available as free / open source software, with more to come. Usage of the website has also grown, with over 100,000 visits just last month! Oh, and there was even a sponsorship deal which helps cover some of the server costs (which is much appreciated).
At the same time, I'd be lying if I said that everything is perfect. I do feel that development of the site has slowed somewhat. Now, that's partly because I've been busy with other things, but also because there's quite a bit of technical debt makes developing R18.dev a bit of a pain.
There's also the fact that some scrapers and other projects have integrated with the R18.dev JSON API (which isn't really for public use). Now, I don't have a huge problem with people doing that, but the fact that projects are relying on R18.dev to behave a certain way (despite the fact that it's labelled as alpha) makes me anxious to change behaviour and break other people's projects.
At the same time, I have to look to the future and think about the long term plan for R18.dev. Although the project was born from the idea of wanting to preserve the information that was on R18.com, the idea has evolved somewhat. Really, the aim for R18.dev is now to provide high quality JAV data.
To some degree, this is already achieved with the database dumps, but there are limitations with the system as it stands. Currently, all the data is imported from other sources, which is okay, especially since R18.dev has functionality to normalise the data. However, there are known errors and omissions with the data, and this cannot be addressed without having the ability to make manual changes to the data.
That is why my next big focus for R18.dev is to enable manual editing of data. Now, I did consider directly implementing this feature into the existing solution, however there is inherent complexity in combining manual edits with automated imports, and I think it would be best to start with a fresh canvas.
So here's the plan: I will begin building a brand new web application, tentatively titled R18.dev Wiki. One of the first pieces of functionality will be the creation and editing of data relating to people. Aside from data about videos themselves, data about the cast and crew are probably the next most important, so I think that this is a good place to start. Currently, R18.dev gives people IDs based on their role: actress, histrion, actor, director or author. This means that if a person has been both an actress and a director, they get two different IDs. Furthermore, a person may have multiple IDs of the same type, especially where they've been credited under different names. I will aim to rationalise that by giving each distinct person a single ID. This can then be associated with role-specific IDs or even IDs in other datasets (e.g. TMDB, IMDb, Wikidata, etc.).
To start with, R18.dev Wiki won't have any automated imports. All the data will be manually curated. I think it is important to get this right first of all. Once that is working well, I can look to introduce automated imports that will work together with the manual edits. Once all the automation is implemented, the current R18.dev essentially becomes obsolete and can be decommissioned.
To be clear: the existing R18.dev will not be going anywhere anytime soon. It will continue to function whilst R18.dev Wiki is being built. And although I probably won't be focusing on it too much, it will likely receive bugfixes and possibly some quality-of-life improvements where necessary.
So that's my big idea. It will take some time before my vision for R18.dev is anywhere near close to being achieved, but the aim really isn't to build just another JAVLibrary (which could be done much quicker) but something rather different. I appreciate everyone's patience and support in the meantime.