@Dr. Daniel Dizdarevic It isn't just because of compression, nor is it because I scale my images down from my original 2100x1400 renderings to 800x533.

As I've said: I don't describe the image with the things in it. I describe the things. Not as they appear in the image, but as they are in-world where I can walk closer to them or move the camera closer to them. It's like an image with a near-infinite resolution.

For example, if there's a light grey blob in the image, four pixels wide, three pixels high, I describe it as what it is in-world, a white sign with three lines of black writing on it. I transcribe the text on the sign 100% verbatim including all spelling mistakes, I translate it afterwards if it isn't in English, I may even explain the text if someone out there needs an explanation, and I may go as far as naming and describing the typeface.

Or if there are two by two pixels on different levels between red and white, I describe them as what they are in-world, a strawberry cocktail in a conical glass, somewhat like a Martini glass. And I slap an "alcohol" content warning on the whole post. Nowadays, I'd even flag the image sensitive just because of these four pixels.

I used to go as far as describing images within my image and even images within images within my image at higher levels of detail than anyone else would describe their own images. I used to describe things that weren't even visible in-world in the place shown in the image. Pictures of places that I would have to walk or even teleport to to be able to describe them. Textures that I would have to make visible otherwise to be able to see all details.

The last time I've described an image in an image with details not visible in the place shown in my image was in this post. I used almost 5,000 characters to describe a poster on the info board. I had to walk to the place displayed in the image on the poster to be able to describe it. The description of the image within the image got so lengthy that, when I was done, I had to remind the reader that I'm returning to describing "my" image. And I actually "cheated" by adjusting the camera in such a way that one of the three posters on the info board is entirely concealed behind a tree trunk because it would have been painfully difficult to describe.

I stopped going that deep when I wrote the image description for what will probably remain my last image post on this channel. The long description was already growing absolutely humongous, and it's my longest one to date with over 60,000 characters. I had actually thought this scene would be easy to describe.

The problem I encountered was that there were simply too many images within images within my image. There's one teleporter near the left-hand edge with a preview image that made me reconsider. In-world, no matter how close I move the camera to the preview image, it mostly shows a square area that appear to be tan all over except for something dark and unidentifiable in the middle.

Actually, however, the place shown in the preview image has hundreds of single-destination teleporters. Several dozen of them are activated and have one preview image each of their destination. I teleported there to take closer looks at everything. I was actually about to write a description of that "teleport station" when I realised that I also had to describe every single one of these preview images, at least those that face the camera in the preview image on the teleporter in the place that I was originally describing. And some of these preview images had images in them in turn.

I would have had to describe probably over a hundred images. In dozens of images. On teleporters which are shown in yet another image on a sub-pixel level. In an image description which was already going out of hand length-wise. On the second day that I was working on that image description. I would have had to teleport at least three times from the place shown in my image to be able to describe these sub-sub-subimages.

That was when I decided to sacrifice details for convenience and only describe what's visible in-world within the borders of the image, excluding both objects that are entirely obstructed by something else and surfaces that entirely face away from the point of view. I do fully transcribe any text that's partially obstructed, though, although I'm considering two transcripts of such texts, namely one transcript of what's visible and one full transcript for better understanding.

#Long #LongPost #CWLong #CWLongPost #Metaverse #VirtualWorlds #ImageDescription #ImageDescriptions #ImageDescriptionMeta #CWImageDescriptionMeta #CW #CWs #CWMeta #ContentWarning #ContentWarnings #ContentWarningMeta

@damon It doesn't help that Mastodon itself is largely a bubble.

Some 70% of all Fediverse users are on Mastodon. But it seems like that within Mastodon itself, at least 95% of all posts originate from Mastodon. Maybe even more.

There are several reasons for this.

First of all, other projects don't federate with Mastodon that much.

Misskey is huge in East Asia, especially Japan. And Japanese Misskey users who hardly know English or not at all won't be interested in connecting with Western Mastodon users, so a large chunk of the second-biggest free project in the Fediverse is out of the equation.

Lemmy is the third-biggest, but Lemmy federates with Mastodon only barely so, also because Lemmy is all about discussion groups and enclosed conversations, both of which Mastodon simply doesn't support. Lemmy users can't follow Mastodon users because Lemmy users can't follow users, full stop. And Mastodon users have to wrap their minds around how to federate with Lemmy. It isn't as straight-forward as communication within Mastodon. And so they simply don't.

Other examples include Hubzilla and (streams) channels having ActivityPub off on purpose to keep ignorant and obnoxious Mastodon users out.

But this goes the other way as well. Mastodon can be outright hostile to non-Mastodon users. Why? Because they don't behave like what Mastodon users are used to from Mastodon and, by extent, partly also Twitter. And they have joined the Fediverse in expectation of something that's one big distributed but homogenous Twitter clone. Anything that deviates from that may be disturbing.

There are Mastodon users who, upon seeing a post with over 500 characters, and be it in the federated timeline, block the poster. This alone cuts into the reach of everything that isn't Mastodon. Not few wish for a switch with which they can permanently filter out all posts with over 500 characters.

Others may block everyone who uses text formatting. Either it simply goes on their nerves. Or they can't imagine that it's even possible to format text in the Fediverse because they can't do that on Mastodon, so they think it's all some Unicode trickery. And as this Unicode trickery is not accessible and inclusive because it irritates screen readers, they deem whoever uses text formatting ableist and therefore blockworthy.

Then there's the issue of content warnings. They must be provided the Mastodon way, or you risk being blocked. However, not everything out there provides a) the right text field with b) the right label on it. Non-Mastodon projects may still label the summary field a summary field instead of a CW field like Mastodon does.

Friendica, for example, has done away with that text field entirely and users BBcode tags instead. Hubzilla doesn't provide any means of adding a summary/a Mastodon CW to a reply. And both have had their own way of adding CWs since long before there was Mastodon which their own users consider vastly superior to Mastodon's way.

In general, boosts are very important on Mastodon. I'd say that most activity on Mastodon is boosts because they're so easy to do on a phone without a hardware keyboard. Your reach on Mastodon depends on boosts.

But if you don't play exactly along Mastodon's written and unwritten rules, and if you don't adhere to the "Fediquette" which is entirely defined by only Mastodon users and geared towards only Mastodon's features (or lack thereof), you're boosted far less.

If you post more than 500 characters at once, it takes a lot for your post to get boosted.

If you post an image without alt-text, the post will be boosted dramatically less because not exactly few Mastodon users refuse to boost image posts without alt-text. You may even be muted or blocked for not providing alt-text. But alt-text only is a thing on Mastodon, and hardly anyone provides it outside Mastodon.

In general, anything that deviates from the standards defined by vanilla Mastodon will cut into your visibility on Mastodon deeply.

CC: @Hiker

#Long #LongPost #CWLong #CWLongPost #FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #Fediverse #Mastodon #NotOnlyMastodon #AltText #AltTextMeta #CWAltTextMeta #CW #CWs #CWMeta #ContentWarning #ContentWarnings #ContentWarningMeta #Fediquette