I’ve been going back and forth on whether or not to add all the blogs that sign up for Blaugust this year (assuming we have Blaugust this year) to the Daily Blogroll.
On the one hand, yes, of course, the Daily Blogroll was written specifically to keep track of Blaugust bloggers in a way I didn’t find myself able to do without it. I probably have a good fraction of the bloggers in the D.B. already.
On the other hand, the DB is basically just a fancy RSS feed reader. Not every blogger provides an RSS feed. Some are video only. Right there, some people are shut out. Then there’s the blog thumbnails; I rely upon blog branding to make them, and a lot of people don’t brand their blogs. And why should they? So then I have to come up with something. Which isn’t a terrible burden, just takes a few minutes for each one.
The biggest issue, I think, is that I only show twelve blogs, max, on the DB. Partially because that is how many easily fits in my browser window (and on the iPad I am using to write this right now). Partially because I just send everything to GPT 5.2 in a clump, and twelve blogs pretty much fills the response buffer. Much more and the DB crashes.
Of course, I could fix that — probably split things into sub jobs and call it a couple of times to get everything. And then raise the limit — this is just an easy switch I can set. But then I wonder if the DB will just become another source of stress, too much to read, something to just be ignored.
The DB eventually will get to everyone’s posts at least once. I’m hoping readers who find bloggers they love will bookmark them and visit them directly, without going through the DB. (Again, I don’t make any money from it, there are no ads or trackers or rel no follow links or anything). People who do blog every day may find some posts are skipped. I choose the blogs that appear on the DB randomly from among all blogs that have updated since the last time they were featured, and then it takes the most recent.
Aside from Tofutush. All of her posts show. Dunno how she has her RSS, but we’re still working through her backlog. I don’t even get that.
Anyway, I did want to talk about the recent DB rewrite. I’d been dumping all the blog rolls in one folder, and it was getting a little crowded. I refactored it like WordPress keeps their contents, in folders by year and month. This meant I couldn’t use relative paths to get the CSS, icons and images, because the index.html at the top level didn’t match the needs of the daily HTML files at the leaf level.
It just occurred to me that all the leaf level pages are at the same relative level to the CSS et al, it’s just the index.html that needs special handling. Maybe I just make a special Jinja template for the root page. Hmm. This would allow me to verify the page locally before I upload it.
I noticed, during the rewrite, that the RSS page had broken. I make the RSS page by reading the index.html and looking for images and summary text there according to specific CSS classes, and those had broken when I went from a grid to an article format. Fixed now.
So I guess I do have a “TO DO” list:
- Refactor leaf pages to use relative paths to their dependencies
- Remove the response size limit from GPT to allow more blogs on the DB
- … and then don’t allow more for “reasons”, but it’s a good idea not to have known crashes around.
- Add new Jinja template for the index page
- Keep looking for new blogs to add
WordPress wanted me to make an AI header image so I told it to make a silly one.
All the text in this post was mine because, really, you could tell if it wasn’t. First time a chatbot can mimic my style, I’m gone, baby.
Leave a Reply