-- Leo's gemini proxy
-- Connecting to thrig.me:1965...
-- Connected
-- Sending request
-- Meta line: 20 text/gemini
The campfire (that is, the #gemini IRC channel on tilde, which serves a similar function to gather round, minus smoke-in-eyes) brought up the notion of Gemini applications, which would presumably be little todo apps or similar treats. Doubt was expressed. Links were proffered.
gemini://zaibatsu.circumlunar.space/~solderpunk/gemlog/ssh-apps-and-accessible-gemini-publishing.gmi
The off the cuff doubt concerned portability and limited utility. This can be expanded on.
Portability here is "how easily can the gemini application for X be installed and run?" There will be various amounts of friction here. Maybe you can do something like
$ doas pkg_add amfora
but, probably not, especially if the yet another tiny gemini application hasn't been added to the various ports and package systems, or does not compile due to toolchain issues, or needs additional work to lock it down with pledge and unveil (like I did for amfora). I've gotten rid of useful tools because having to pull in the whole Glasgow Haskell Compiler to compile that one little application was too much bother. Shell code is portable, right? Not really.
$ head -c1 /dev/zero
head: unknown option -- c
usage: head [-count | -n count] [file ...]
$ which bash
which: bash: Command not found.
And the security model of random sh code ain't great.
If portability means "the Android store" or "some yet new Linux thing" or "the web, jangled with JavaScript" that's not very portable. Maybe it's portable enough for you? But then you're competing with lots of other apps, existing frameworks, user expectations--they want WordPress, even if there's a framerate because the CPU is struggling to keep up. Where's the advantage of gemini? Where are all the apps?
A problem might be that gemini clients are "good enough" for reading blog posts and whatnot and the benefit minimal for anything more specialized. Sure you could imagine a gemini weather app tool thing, or you could just follow
Works fine, good enough. This is a similar problem to where a text editor is good enough, and since one probably spends a lot of time in that editor one gets to know it pretty good. One might imagine a better tool for some task, say for writing poetry in lojban, a tool that automatically tracks beats per line, rhythms, highlights alliteration, warns if a phrase has been re-used in a renga conga line, stuff like that. This could be useful. But someone would need to write and maintain that tool, and would it really be better than wrangling words in vi or scribbling in a notebook?
On the other hand, I was having trouble finding that comic, so I wrote a little curses application just now to search and show XKCD comic metadata. With more work, the code might even be publishable. Sometimes you can use a search engine to find XKCDs, sometimes not.
> Search is getting worse. You've noticed it. I've noticed it. Everyone can see it plain as day. It's not just slowly getting worse either. I feel like the 2nd, 3rd, &c. derivatives here are all negative.
> You've noticed search getting less useful, but it's hard to tell for how long. Google's top results are full of garbagey sites; you're adding +reddit to every query to get what seems like the last non-astroturfed source of opinions online, and who knows how long it'll be before that's gone as well.
> Google's results have been getting worse and worse these last few years, and we're at a point that it's becoming unbearable. Even ignoring the numerous ads that Google stacks on top of your search results, the actual organic results are in many cases plain garbage.
This might be seen as application bitrot, where the application is no longer a good fit (or never was, and you only now noticed) for the problem space, kind of like how rogue shows its age compared to modern roguelikes. This rot is more abstract than the kind where a capsule goes poof! but is still a problem. Of course the applications would need to be written for them to have a chance at rotting...
Of course small applications are of limited utility. That's the point! Duh! However, we again run into the "why not just point a gemini client at..." problem, where a link is good enough, and if it bitrots, well, there are other links to be had and hoarded, until some sword-bearing Dane does rudely interrupt your slumber. An application has the combined problems of:
server bitrot, where the server-side portion goes away even though there are still clients happily clienting. This happens I hear for games, though some developers now have a "if it's not profitable, we'll open source it" policy. Now there's some capitalist realism!
code bitrot, where the software or toolchain to build it no longer compiles on modern systems. This can take a while, and depends on the software involved, and whether anyone cares to modernize. Abandonware is a common fate.
abstract bitrot, where the application no longer fits the use it was written for. This can also happens when a corporation buys something and ruins it, or some new shiny comes along wielding the +7 sword of Distract Nerd.
Little gemini applications may be quick and easy to write, but they could also be quick and easy to abandon if the tuits aren't there, or the burn-out is, or life happens. Of course, again, they'd have to be written for that to happen. So where are the applications?
* thrig wields stick of hornet nest poking
< Rynn> It's super effective!
(A valley or a pot might be more useful from the empty space it contains, not one full of applications.)
-- Response ended
-- Page fetched on Tue May 21 09:58:10 2024