Sometimes I get a bit overwhelmed because many of my creative projects are things that I can’t ever “finish.” Like detailing a fantasy world, or mapping the real world, or anything of importance, probably, projects are forever unfinished. It’s us humans that decide to add “Fin” to the end of track, but we all know the work continues. It is perpetually work in progress. Sometimes we just decide to no longer work on it.
Face Generator generates faces from elements like eyes, ears, noses, mouths, hair, and so on. I could add new collections, or try and find artists who wanted to contribute, or I could add to my own collection. I sometimes load gallery of random faces and wonder. I’d like to add more elements to generate recognizably black men and women; I would like to add more good looking hair styles for women; I would like to make elves better in some way; I’d like to add more intelligent beasts. The spider faces are particularly cool, but there’s probably other types that would make more sense. On the technical side, I also know that there’s a problem: the code tries to figure out if an element has transparent pixels or if “white” is actually transparent. For my images, that is the case: white is transparent. Therefore, there’s no way to add “white hair” – because underlying lines show through (head silhouettes, ears). So I could also spend some time drawing “white hair” with actual white pixels and real transparency, and make sure the code knows how to mix these two modes. At the same time, I feel there’s the problem of diminishing returns, here. I always wonder: is it worth the extra effort? Perhaps the most enjoyable job would be to just add a new type of images and start drawing again, right from the beginning.
Face Generator: random faces gallery
Text Mapper generates maps for tabletop role-playing games. I’m happy with the Alpine algorithm. It just works. I’d love it if it could generate steeper valleys, but I’m not quite sure what I’m looking for, so I’m happy enough as it is. I’m happy with Erin D. Smale’s algorithm but it doesn’t have rivers; so one could add them. That would require an altitude model, which I currently don’t have. I’m happy with the Gridmapper algorithm, but I’d also love to generate two or three levels for a dungeon so that the stairs line up. I have code that does it, but it isn’t quite done. It was surprisingly tricky and I’ve lost steam. The Apocalypse algorithm is simple and cool. Perhaps I could think of more details to add? The Traveller algorithm generates beautiful curved routes. The Island algorithm is new and interesting but I ran out of steam. What now? What next? The Archipelage algorithm was supposed to offer something new, but I forgot what it was and now I just see an unfinished project.
Hex Describe is the worst. It contains many hundreds of random tables, big and small. It takes a Text Mapper map and uses the random tables to generate a mini setting of tabletop role-playing games. When it works well, reading it makes me want to run games in these worlds. I love that. I love the Alex Schroeder tables together with the Alpine maps. I should get back to these tables and add more details. A while ago, I wanted to detail all sorts of extra-planar explorations but I ran out of steam. The Traveller tables with Traveller maps combo works well enough, but I’m thinking the level of detail is uneven. The crews of a type M subsidized liner have a bit too much detail, for example. I’m not sure about the skill selections. Many ship types are missing. I’m not sure about ship type distribution. It needs more missions. In actual play, I’m finding some of the missions to be too boring; I never use them. But I also don’t delete them from the generator, so what’s the deal? I don’t know.
Gridmapper is an SVG+XHTML+JavaScript single-file application. It could use an overhaul. Compressing and base64 encoding the maps. Organizing the saved dungeon maps on the wiki in some way. Rewriting the code in a style that I like better. This is still my first and only JavaScript project but I know that I don’t like how the code is organized.
Traveller Subsector Generator is a separate web app that also generates a Traveller subsector or sector map. That’s right. The algorithm from Text Mapper, rewritten. I guess I’m happy it’s on CPAN, at least. I’m not sure what to think of the traveller name generation, of the “realms” and the color codes. Was it really worth it?
Character Sheet Generator is a web application to fill SVG files with variables. In a way, better than fillable PDFs because a computer can easily churn out a ton of character sheets, and I do it for Halberds & Helmets. The promise was, however, to do this for more rule sets, to have more templates, and I never went back to it; and I never made it easy for other people to contribute their rules, either.
Halberds & Helmets podcast is aways in need of new episodes, I guess. But contemplating how many episodes it will take for the podcast to be “finished?” Impossible.
Sigh! Unfinished business wherever I look. I mean, it’s cool. Forever projects take forever. It’s like practicing any art, you just play and paint and write and do … and then you die, I guess. But it’s probably also why we like to take slices of our art and label them “finished”. There are songs, albums, paintings, essays, short stories, books, blog posts, and they’re “done” at some point.
Even this post will be “done,” promised!
Perhaps I’m setting myself up for these things. I can always go back to wiki pages and edit them. Leave comments on them. Perhaps that’s why I like the blog-like structure of this wiki. I like the idea of a wiki page, but a wiki page is “forever in progress.” A blog post is more or less “finished.” Sure, I can add comments, and I do, but those too are “finished.”
Anyway. Thoughts on projects and writing and art.
#Programming
(Please contact me if you want to remove your comment.)
⁂
It would’ve been two seconds of work for me to change tsp to white if I knew you wanted that; I was trying to more or less match random.png
– Sandra Snan 2021-07-30 06:42 UTC
---
The problem is rather that I need all the PNGs with transparent background and the opaque parts filled in with white pixels. None of the existing assets have that. Currently all are either blue and white, or blue and transparent.
– Alex 2021-07-30 07:05 UTC
---
I reckon you’ll probably need real transparency for some other case, so it’s probably better to implement it now rather than building hacks on top of hacks. Then again, if it’s just a toy project, just do whatever’s more interesting instead \/(ツ)\/
– Anonymous 2021-07-30 07:27 UTC
---
Yes. Technically, it’s all there. I just have hundreds of tiny PNG files where I would have to add either add the opaque pixels or make the surroundings transparent. It’s not a programming task. It’s the not automatable task of fiddling with hair assets, and I’m not motivated to do it. So everybody has dark hair.
– Alex 2021-07-30 09:22 UTC