💾 Archived View for midnight.pub › posts › 1842 captured on 2024-05-26 at 14:39:49. Gemini links have been rewritten to link to archived content
➡️ Next capture (2024-06-16)
-=-=-=-=-=-=-
Last straw for html
I recently became interested in have an infinitesimally small /dev/null presence (i.e. a http/html blog), and at first messed with it being minimally html by putting content mostly between <pre> tags.
But then I thought, oh, you silly <gender>, just go the whole distance!
So I somewhat did, leveraging <head>, <title>, <link>, <style>, <body>, <h*>, <ul>, <li>, <p> - and even <div>!
The latter was only to facilitate supplying an itty bit of resize-related style in a separate .css file, making that easy to change across the site.
Well, it went pretty much as usual. It started out feeling "cool". But soon all that tag overhead obscured the writing itself - something I feel gemtext solved rather nicely.
But there was one nagging problem that I absolutely couldn't solve, namely that one blog article didn't display the same as the others from a size/resize point of view, despite using exactly the same <link> referring to the same separate .css file in order to style the one <div> container in each article file.
I mean, I stared... I removed everything but tags to do diff's.. I was sure to refresh to avoid page cache biting me in the previous-version-of-a-file ass.
Absolutely nothing worked.
All the articles looked the same in Chrome on the Chromebook. The difference for that one file was in Chrome on my Android phone, where the text displayed tiny, and whose text didn't reflow at all, meaning I'd have to resize that page than then scroll left/right in order to read it.
Absolutely no "what" or "how" discovered.
So after too many hours of my life wasted on something that ought to Just Work by now, I gave up. I'm officially motherfucking done with motherfucking html!
Well... except for one tag. In order to have a index page that would display when going to the root site address, I'd apparently need an index.html file, which led to employing a single <pre> wrapper with style="white-space: pre-wrap;", which somehow allows text reflow within a <pre>.
(Looks like I could have a default index file of "index.txt" if I self-hosted and put The Right Stuff in the server's .htaccess file. But self-hosting seems like boiling the ocean for what experience screams won't get much visibility anyway....)
All the rest of the files are .txt, which are somehow treated by Chrome, elinks, and lynx the way I'd hope (although it's important to remember to put text you want to reflow/wrap together on the same line).
Anyway, here's how it looks: https://tilde.club/~oldernow/
check minim.blog, he has stuff on an index.txt (type) page that works. gh repo too.
try:
There are a lot of advantages - you really get an old-school website that will
work in old browsers, via HTTP
you will be able to edit the site from the admin panel (even under DOS).
you can add your domain painlessly
if you are not satisfied with third-level domains
your site is located on a subdomain and not in a folder with a tilde
More than 40 templates are already ready for use, many of them are proprietary, work under older browsers, and are backward compatible.
Widgets are available by simply entering a code
blog widget, guest widget, feedback, chat, and so on
same, i just use a text file now :) gets the job done
https://jordanreger.com
I think I know why your page didn't look right on Android. I think you omitted the following:
<meta name="viewport" content="width=device-width, initial-scale=1">
This meta tag is irrelevant on desktops and laptops, but it does all kinds of dark magic on smartphones and tablets.
Also you don't need <head> or <body> in HTML5; that stuff's required for older versions of HTML (as well as XHTML, but nobody ever cared about that).
I’m of two minds about this. I’m definitely in favor of simplicity, but that’s not the same in my view as *nothing*. It’s important not to throw out the baby with the bathwater, in other words.
For example, on a 4K screen (like I’m using now), I generally find un-formatted text to render too small. It also adds extra work to follow links that aren’t rendered as such (even if it’s not very much). I say this to say that the more minimal the site, the more of the onus you put on your user/reader to fix things. A small bit of CSS can aid readability significantly, and also respect things like the user’s choice for dark vs. light. I also don’t know how well screen readers do with that level of minimalism, so accessibility can be an issue.
The text size/display issue you mention says to me that you hadn’t quite found the best rule(s) to use, up to and including what units of measurement you were using.
I don’t want to sound hyper-critical or anything; obviously you do you for your own space.
http://bettermotherfuckingwebsite.com/
I read your several posts about html and I sympathize. I am also using a chromebook, isn't it annoying in chrome, when you start typing an url, it automatically autocomplete to the most specific, convoluted, and always the least useful version of the url that you're typing? If I type foo. it completes to foo.bar.baz/res/1231654124/index.html FUCK!!! What's the point of having an autocomplete if the only thing it's good for is to either delete most of it or keep typing the whole url I actually mean to open!?
Anyway, I use the chromebook to use the shitty parts of the internet, you know, just for piracy and to sometimes to peek into Big sites, which I don't much use anymore anyway.
On my tablet I have, on purpose, removed all google stuff (of course, I know this is somewhat pointless because they can keep profiling me if they want to), and keep the factory version of chrome, which doesn't even let me open cloudflare apps or even recaptcha. I think sites that use cloudflare are just ddosing themselves lmao.
Anyway, I have been thinking of dusting off my old little lappy and just install plan9 (I've talked about this already) and write my own gemini client and web-scraping scripts to use only the sites I like, usually textboards and sites mostly without javascript. For a while I've wanted to use a mostly-plaintext version of the web, and a very plaintext approach to computing in general, the old Unix way (not that I'm a unix fanboy but I am a sort of minimalist because I can't be bothered to be installing all sorts of programs written in all sorts of languages each of which needs it's own tooling and... you get the idea.)
~bartender, is it too early to have some bourbon? I see, bring me some regular black tea, then.