[marginalia_nu]

I think the elephant in the room is that in many cases, the only way to improve accessibility beyond a fairly low point is to lower the utility.

A good example of this is tables. Tables are amazing at conveying large amounts of complex data in a very intuitive way. Not only that, it's great at manipulating data as well. It is not without reason Excel is powering a significant portion of the business world, and try as many startups have, it's very hard to pry Excel out of the hands of people who are using it. Excel is extremely versatile and very good at what it does, owing largely to the table metaphor.

Problem is that tables are also not very accessible. Among other things, they all but require sightedness. They also don't really work on mobile in any practical sense. At the same time, any accessible alternative to large tables are a strict utility downgrade for people who are able to partake in tabular data.

Starting with an accessibility first design principle, it would not be possible to produce something like Excel. We'd be stuck with a hell consisting of a million mobile apps, one for each conceivable workflow and task, rather than having one tool that can be made to perform any data manipulation task.

[extra88]

1. "Accessibility" means more than just "blind people can use it."

2. "Tables" are primarily for static tabular data (but not exclusively). Excel is a "spreadsheet," a grid of interactive cells.

3. While accessibility is best thought of as an aspect of usability, the practical limitations of tables or spreadsheets on phones are usability problems for everyone, not just for people with disabilities. It's more about the touch interface than what you can see on the screen; a modern phone can display a lot more than the monitors used for the first spreadsheet programs.

4. While having vision can make it more efficient to take in and understand a lot of information in a table or spreadsheet, a screen reader user can navigate them in two dimensions if they're made correctly. Excel's core functionality is useable with a screen reader though not every single feature is.

https://support.microsoft.com/en-us/office/screen-reader-sup...

[marginalia_nu]

I'll try to reiterate my point:

One of the most useful desktop applications uses a metaphor that is among other things, fundamentally incompatible with mobile devices, and also all but useless for blind people.

It's also important to note that Excel's value is not primarily as a presentation format, but a data transformation tool. Its strength is that it permits the user to define ad-hoc tables and then proceed to transform that data in completely arbitrary ways.

Adopting an accessibility first mindset, you can design software that does not use such a metaphor, enabling it to be accessible on mobile and for people who aren't sighted, but such a design would exclude a lot of the utility of Excel, since almost all of its utility comes from the grid metaphor.

> While accessibility is best thought of as an aspect of usability, the practical limitations of tables or spreadsheets on phones are usability problems for everyone, not just for people with disabilities. It's more about the touch interface than what you can see on the screen; a modern phone can display a lot more than the monitors used for the first spreadsheet programs.

This isn't actually correct. Modern phones have very high resolution, but also very high DPI. The amount of (readable) text is very low, even compared to an old 15" CRT on which you could read a 8x4 pixel font without much trouble.

No matter how high the resolution is, the fact of the matter is that a mobile phone is generously half the size of a page in a paperback. Even an old, small computer screen was the size of full two pages, if not more.

[Frummy]

Maybe in some cases. I think it's a creativity issue. In an ideal world the utility can be preserved or increased, with the use of an innovation that improves significantly the life of the functionally varied, with secondary effect of reducing for example cognitive strain on the original user. There may challenges in mindset, funding, time, organisation, lack of understanding or empathy, and so on, for this to be reality. But I would argue that there is a flow of unforeseen benefit from the user that requires a feature, to users that don't require a feature but reduces time and energy spent on getting something done just because the path is shorter, easier, friendlier. Personally I have used many accessibility features as part of small productivity hacks. I also think the study of for example UI accessibility doesn't only apply to the functionally varied but also makes life easier for any user.

[marginalia_nu]

If you ignore all the cases where a statement is contradicted, any statement appears true.

Ultimately the visual medium permits random access 2D representations of information that just aren't viable in the sequential-access 1D medium of speech / small screens.

Sometimes accessibility improvements are to general benefit, but this isn't universally true.

[Frummy]

You claim utility first and accessibility as first principle removes utility. Removes utility for who? Accessibility is all about utility for the groups who need it the most. You can have normal excel and an excel that uses sound for blind people on the side. It's not one or the other, and the other in this case also puts utility first for the people it applies to. And what you learn from that can make normal excel better too.

[marginalia_nu]

> Accessibility is all about utility for the groups who need it the most

Do some groups really need utility more than others? This seems to contradict the idea of accessibility in general.

> You can have normal excel and an excel that uses sound for blind people on the side. It's not one or the other, and the other in this case also puts utility first for the people it applies to.

It really is one or the other. You cannot represent an effectively infinite 2 dimensional grid in speech. This is impossible because speech is one dimensional. This 2d grid is not accidental to excel, it's at the core of its utility.

[Frummy]

Some people need help and it's a positive sum game to help. You have misconstrued me.

>Do some groups need really utility more than others?

Some people have a lot of utility at hand, and some people who have no utility at all, we can do some stuff so they also can use a tool.

>It really is one or the other. You cannot represent..

The 2d grid is stored in computer memory. You don't need speech to store it. Speech is just a communication medium. A lot of accessibility is just altering the communication medium. It has columns, rows, things that are described with words. You can index into the grid. It is true that sight is a good thing, but your total refusal to provide features to those with alternate needs is ableist. It doesn't matter if the visual component is the original source and I agree excel couldn't have been created without sight. Clay tablet accounting couldn't be created without sight. Doesn't mean you can't use other senses when something has been created. Why the refusal to share? It's pointless not to share. An ending note: Leonhard Euler was rather productive while blind.

[marginalia_nu]

> Some people have a lot of utility at hand, and some people who have no utility at all, we can do some stuff so they also can use a tool.

Right, but the point is that we all need the same utility, right? Some people may presently have less utility available, but surely the goal is for the utility to be the same?

> The 2d grid is stored in computer memory. You don't need speech to store it. Speech is just a communication medium. A lot of accessibility is just altering the communication medium. It has columns, rows, things that are described with words. You can index into the grid. It is true that sight is a good thing, but your total refusal to provide features to those with alternate needs is ableist. It doesn't matter if the visual component is the original source and I agree excel couldn't have been created without sight. Clay tablet accounting couldn't be created without sight. Doesn't mean you can't use other senses when something has been created. Why the refusal to share? It's pointless not to share. An ending note: Leonhard Euler was rather productive while blind.

Enough allusions to a nebulous solution that must surely somehow exist. How would you represent Excel to a user that is not sighted? A spreadsheet is not just a table, it's tables within tables. The data is scattered in tables around the sheet and inconsistently surrounded by areas of empty cells, the data is ad-hoc and user created and can not be assumed to have any standard headings or type annotations as many spreadsheets just don't. Unrelated tables may also be adjacent to each other within the sheet. A row of empty cells may be semantically meaningful, or just the space between two different tables. There may or may not be headings or column labels. Any cell may be represented both as a formula and a value. Formulas may refer to any cell both within the sheet and within other sheets in the workbook. It's as truest digital equivalent of unstructured writing with pencil on a blank sheet of paper.

Many of the things that make this a complete accessibility nightmare are the same things that makes it an extremely versatile and high-utility tool for abled people. The complete lack of structure and freedom to invent completely ad-hoc work flows that may or may not resemble structured data is exactly what makes it so incredibly powerful.

My entire point I've been trying to make is that there is an inherent conflict here. Pretending like it doesn't exist doesn't make it go away.

[hulitu]

> I think the elephant in the room is that in many cases, the only way to improve accessibility beyond a fairly low point is to lower the utility.

They did this on Windows and on Android. It didn't work. Try resizing a window on 4k screen in win 10. Or try to get rid of an error message in Android 10 (hint: there is no Ok or Cancel button and the error message obscures other UI elements).