Gemini vs the Pile of Specs

Before I had heard of Gemini I read this post:

The reckless, infinite scope of web browsers

I loved it, fantastic text; I got into Gemini because of it.

But what I found quickly was that Gemini was just even more specs to learn and implement on top of all the old stuff. It didn’t take away any of the old scope.

I know people like to trot out xkcd standards a lot but yeah. It’s just a good strip that fits many cirumstances.

Now, Gemini is fantastic. I love it. But. It doesn’t really make the web as a whole any easier. Instead it just makes the web as a whole more complicated. We now need special browsers, special servers, special pages, special certs. And I’m not really ready to miss out on the old stuff any time soon.

I can only hope xkcd 927 is wrong. That sometimes something new can be worth it.

The gem spec text itself says that it wasn’t the intent to replace the web (or to replace Gopher) that but a lot of people (not necessarily the inventors themselves) have turned to it as sort of a… “the web is complicated, now here’s Gemini, jolly good.”

And the end result in practice so far has been more stuff to implement.

People like to invent and make and have Not Invented Here. That’s how we get things like json-feed (which hasn’t rid the world of Atom and RSS) and XML (which hasn’t rid the world of SGML) and Gemini.

And arguably there are two kinds of 927 standards; those trying to engulf and encompass its predecessors (like Matrix) and those trying to be a simple thing on the side. There is a qualitative difference between them.

It’s not that I’m not onboard with Gemini. I am. It’s that so far, it has meant more time spent coding and fiddling, not less. I think a lot of people like doing that and that’s why they don’t see the problem. It’s so pleasant that they don’t notice the hours.

But to me… imagine you’re trying to write a letter. Do you wanna spend time making the envelope, making paper for the envelope, or do you wanna spend the time writing the actual text?

Gemini can replace web pages (blogs, news posts etc) but not apps.

We’re either gonna need a separate solution for replacing apps (SSH kiosks fail in the a11y dept, but maybe if there’s also an API) or we’re gonna have to stick with the web.

One way to work on a simpler web would’ve been to make a brutally strict validator/linter for a modern subset.

I mean, Gemini has already been created now. Undoing the creation of Gemini is not on the table. Might as well adopt it. (And I have.)

I’m just emphasizing that it has not made my dev life any simpler. There is absolutely nothing that has been made simpler in practice because of Gemini. It’s just more. More to support, more to hack on. The dirty secret is that a lot of people like that, they like implementing things. It’s satisfying to them. It’s a hobby almost, like model trains.

Which is fine (I don’t necessarily feel the same way) but that doesn’t support claims that it’s in service of decreasing the scope of the web.

The joy of tinkering makes the argument to Gemini as a solution to the whole reckless scope thing not really legit. The scope is reckless because people like to mess and implement and NIH and grow and improve and startover and the mess, the pile of specs, is full of such projects.

Now, all of the ↑↑above↑↑ is just idle musing.

I hope I am wrong. I hope we’ll solve the web’s problem and Gemini is as good a shot as anything else.

I’m just been so burned so many times by “if you buy such-and-such gadget you can live a more minimalist life” and then you end up not really being able to get rid of any of the old gadgets because the new thing doesn’t quite replace the old ones. Sometimes it does (you need some discipline) but even when I went to dumbphone it was because I had no phone, my phone was broken, and I thought I might as well go to dumbphone as anything else.

I’ve been on an “I hate the web” kick for a while, which is why peeps pointed me in the direction of both Drew’s essay and to Gemini. I’ve been using scrapers, atom feeds, emacs modes etc to do less actual web. So it’s not that I’m not sympathetic. I am. I’m just a pessimistic cynic who hope’s she’s wrong about this.

Let me phrase it another way.

In a world (trailer voice) where cars and roads and infrastructure is a complex and dangerous engineering problem and peole are drowning in ordinances, specs, plans and drafts…

The sudden invention of bicycles doesn’t make that problem any easier.

Bicycles are still good, cars are still bad. Happy bikes exists. Do not get me wrong on that part (that’s why this is such a good analogy).

It just means that we can’t say: “look how easy we made life for city-planners, road-builders, and legislators!” Not really. The opposite thing happened. Because now you have bikes running into pedestrians, and cars running into bikes.

So hearing about how Gemini made your life better, you love being on there and reading and writing etc, that’s great! Makes me happy, keep it up!

The argument that it solved the problem of the web’s complexity, though? Jury’s still out on that and so far it’s leaning towards the opposite being true. I’m still finding and fixing issues in my own capsule setup.

Radical changes can and have happened. Java, Perl and XSLT are on the wane, while json and react or whatever have been on top for a while. Aeons might shift once more.

xkcd standards

I went to dumbphone