ok, well, is a JPEG of a page of text an image?

ok well, if you print out a JPEG of a page of text, surely that is still an image then?

hm

what if i take a Polaroid photograph of a page of text? is that an image?

@Lady Which I guess the first one is saying some of that, so I'll vote that one, but it's missing crucial parts of the reason (e.g. a movie is not an image).

@Lady Oh, also gifs are intended to be used as either static or animated images, so it wouldn't be entirely accurate to use a format type that implied the sole intent of the format was for animation.

@unspeakablehorror why do all gifs have to have the same format type?

@Lady Seems more of a hassle to split them up than just have one single format. Also seems excessive to have a whole type that would be devoted almost entirely if not exclusively to one file format.

@unspeakablehorror on mastodon gifs are actually served as webm format. do you think that is fundamentally a different type of thing just because it isn't image/gif?

@Lady Yes, because the format itself allows audio. To me it's not about any given usecase for a format, but the format's capabilities as a whole

@unspeakablehorror what is your definition of “format”?

@Lady for example gif, png, mp3, mp4, avi, webm. Each of these is a type of computerized encoding that has a certain set of capabilities that implies a certain intent they are particularly suited for. Collectively they can fit into larger subtypes which the mimetype is used to indicate.

@unspeakablehorror what about physical resources? is a polaroid not be an image, because it has no computerized encoding?

@Lady Sure, it's an image. I guess the question is, is the intent of this category to be used as a formalized labelling (eg. a mimetype) or informally? Because formally, I'd consider gifs images but informally I'm happy to distinguish between an animated gif and a static one.

@unspeakablehorror personally i would say a polaroid is formally an image, not just informally :P

@Lady True. The question of the intent of the categorization still stands though. Is the purpose to be part of a dictionary definition, a formal specification like a mimetype, as use in casual conversation, or something else entirely? The context is relevant for determining what categorization(s) I would use.

@unspeakablehorror i think it is a formal specification, but one that encompasses physical resources in addition to digital ones

@Lady So not a mimetype, then. I'd need more details about the purpose of the specification to know how I'd categorize it, then. Is it for putting in a dictionary to formally define words for definitional purposes, or is there some other different or additional purpose to the categorization?

@unspeakablehorror let’s say it’s so that when users filter their search to only show images, they get back a list which only shows images

@Lady In that case, I'd say there's more than one fairly reasonable choice. Depends on how much work you're willing to do for it. If the goal is a low effort implementation that exhaustively shows all images, then yes, I think considering animated gifs as images is the way to go. If the amount of effort required for the implementation isn't a consideration, then no, I'd give them a separate category like 'short animation' and reserve 'image' for static images.

@unspeakablehorror do you think pre-talkie motion pictures are not videos, then?

@Lady No, they're videos. It's about the ability of the format as a whole, not about the potential for silent videos. Modern videos can also be silent. But by modern standards the format itself needs to have the capability for sound as well as moving images.

@Lady its an accessibility nightmare is what it is

@Satsuma hey it could have alt text

@Lady no.

@Lady It's a treacherous image, but an image.

@aschmitz @Lady ces ne pas une APNG

@Lady Only if it's artefacted enough

@gaditb i was originally going to say PNG but then i thought some might have this take

@Lady Less snarkily that's not the EXCLUSIVE condition.

If the content has fundamental pieces that are inextricably (statically) visual, it's an image.

So like:
https://vnsmatrix.net/wordpress/wp-content/uploads/cyberfeminist-manifesto-vns-matrix-english-sphere-final-bleed-artwork.jpg
(Shoutout the the https://obn.org Old Boys Network for this, even if nowadays I can't currently find it on their website.)

That's an image in JPG, that'd still be an image in PNG, and if you recreated it in SVG with <text> nodes that can be highlighted and manipulated as text while still doing that picturewise, it'd still be an image.
Same as if you recreated it in DOCX.

That should answer your next two to three questions, if I'm correctly guessing where this is going.

@Lady ... dang, maybe I got lucky but the two outgoing links from https://obn.org/obn/reading_room/fs_read.html that I've tried point to this as the most stable collection of web links I've seen this side of like IETF-and-co. stuff.

@Lady you should have included a “… well now i’ve backed myself into a corner” option on these polls

@Lady indeed, a jpeg of a page of text is an image, but a printout of that jpeg is text. My reasoning: On a computer, plaintext can be copy-pasted and have assistive tools used on it, therefore an image of text is not the same as text. But you can't do that with text on paper, so both a printout of plaintext and a printout of a jpeg of a page of text is text.

@unspeakablehorror wouldn’t the fact that you can’t copy-paste text on paper imply that it isn’t text, not that it is, if that is part of your definition of “text”?

@Lady No, because my definition of text depends on the context. On a computer, that is part of my definition of text, but in a non-computer environment it is not.

@unspeakablehorror but is not(computer‐text) not still not(computer)‐text? or do you think there is something fundamentally different between printing to an e‐ink screen and printing to paper?

@Lady Yes and no. They both have a partial overlap of definition in terms of being symbols with the ability to record and impart meaning to people, but the functional capabilities are part of the definition of text to me, and those change depending on whether text is on the computer or not. I'm viewing this in the context of a more formalized version of coercion (https://socialsci.libretexts.org/Bookshelves/Linguistics/Analyzing_Meaning_-_An_Introduction_to_Semantics_and_Pragmatics_(Kroeger)/05%3A_Word_Senses/5.04%3A_Context-dependent_extensions_of_meaning) in that I think of the formal definition as literally changing depending on the context.

@unspeakablehorror sure, but it seems like a lot of privileging of “computer” to me. what makes computers so special?

@unspeakablehorror (also, for the record, an iphone can select and copy text from a printed document (or a JPEG (or a rendering of a JPEG on a screen)), so i question your context delineation. people aren’t without computers just because they are dealing with physical entities. we live in an age of cyborgs)

@Lady Yes, computers that aren't iphones can sometimes do that as well, but image-to-text doesn't have perfect accuracy anymore than text-to-speech does. So from an accessibility context, that is absolutely not functionally the same. That action also cannot be considered the same as copy-pasting from plaintext to plaintext in terms of fidelity. So on both counts, it means the source document is definitionally not a text document with respect to the destination.

@Lady I don't see this as privileging computers at all. In fact, I would see it as privileging a context if the definition was instead static and only addressed one context or the other. I think this should apply in non-computer contexts as well. For example, on Earth, the definition of up is defined with respect to Earth's gravity, but in space I define it differently as Earth's gravity is not as relevant in that context.

@unspeakablehorror but “on computer” vs “off computer” is a significant part of your definition of context. how is that not privileging “computer” as a category?

@unspeakablehorror there are websites which swap out letters for other letters and use special fonts as a form of DRM to prevent people from copy‐pasting from them. there are websites which don’t let people select text at all. there are PDFs whose underlying text that you can select and copy is OCR’d and buggy. there are PDFs with no underlying text at all.

i don’t have any problem classifying all of these things as text documents, because my definition of “text” doesn’t require any particular technological affordance or computer interface. it doesn’t require that a thing be on a computer, or off a computer. but you are defining “text”, in a computer context, based on what computers are able to do with a thing. this is a definite privileging of computers, and technological affordances, as the basis of your ontology. and it is the reason why you need to then come up with other definitions for other contexts where those technological affordances do not apply.

@Lady Sure, there are websites like that and pdfs like that.I don't consider those kinds of websites and pdfs to contain text.They are 'documents', but not text.

I think words should have different definitions depending on the context.Sometimes dictionary writers even think this, as they give multiple definitions to some words.

I understand that your definition of text differs from mine, but I would argue that you are simply privileging different things from me in constructing that definition

@Lady If your definition of privileging simply means 'affecting the definition of words', then sure, I am privileging computers in this context. But if by privileging you mean 'treating in a way I don't treat other differing contexts', then I disagree.

@unspeakablehorror by “privileging” i mean, kind of, both: you are, for example, treating the computer affordance of copying and pasting as distinctive, but not stating, for example (to my knowledge), that text carved into stone or written on a whiteboard is not text, because you cannot cut it out and paste it onto a piece of paper (the way you can a newspaper clipping). nor do you merit things as text which a human can manually trace in a paint program, even tho that is a form of copying.

and certainly, you privelege assistive technologies being able to read a thing over humans being able to read it in your definition, because you find the former to be distinctive, and the latter not to be. you assign weight to computer processes which you do not consider significant as human processes. that is what i mean by “privileging”.

@Lady "and certainly, you privilege assistive technologies being able to read a thing over humans being able to read it"

Is the purpose of the assistive technology to help computers read text, or to help humans read it? If I were privileging the computers themselves over that of humans in that context, I would be saying things like 'Python is a language when on a computer but English isn't'. But that's not what I'm saying. What I'm privileging are significant functional differences for humans.

@Lady "but not stating, for example (to my knowledge), that text carved into stone or written on a whiteboard is not text, because you cannot cut it out and paste it onto a piece of paper (the way you can a newspaper clipping)."

You are correct that I am not making these kinds of distinctions, but I would not necessarily consider it invalid for someone to do so. Perhaps someone considers text on a whiteboard text only if it can be erased and added with ease. I think that's valid.

@Lady "nor do you merit things as text which a human can manually trace in a paint program, even tho that is a form of copying."

Yes, my definition considers this distinct from copy-pasting ability because of the order of magnitude of difference in difficulty involved in doing this.

@Lady Merriam-Webster agrees with me (though I would've phrased this definition differently):

5b. matter chiefly in the form of words or symbols that is treated as data for processing by computerized equipment
| text-editing software

https://www.merriam-webster.com/dictionary/text