For what it is worth, here is my advice as someone who has lost my sight over time and who have nevertheless continued to do a good deal of coding. Firstly, my

For what it is worth, here is my advice as someone who has lost my sight over time and who have nevertheless continued to do a good deal of coding. Firstly, my impression is that people tend to overestimate the practical challenges associated with vision loss and to underestimate the psychological side of it. These things are of course linked – the more quickly someone accepts the fact of their vision loss, the more open they become to using various forms of access technology. That said, coming to terms with vision loss is not something you can rush. It took me a long time. Secondly, I don’t think it is necessary to try and learn everything before you actually need it. By all means, encourage your friend to install NVDA on a Windows computer and mess around with it when he feels like it, but for the most part I’d say use your eyesight while you have it. I don’t think he needs to worry about getting stuck if he doesn’t learn everything ahead of time. He might also be more motivated to learn new tools once he really needs them. Two critical things to learn will be to touch-type, if he can’t already, and to use a screen reader like NVDA or JAWS. These are underlying skills that will make everything else easier. By learning a screen reader I mean learning its basic usage, but also learning a whole set of new screen-reader-specific short-cut keys. People who don’t use screen readers often make the mistake as thinking of it mainly as a text-to-speech engine, while in fact the essence of a screen reader is the multiplicity of ways it allows you to interact with the computer in non-linear ways. To be an effective blind/low vision programmer, it helps a lot to be a super user of whatever screen reader you use. The related thing that should happen naturally is that he will set the TTS voice to speak faster and faster – which brings us to another point. One of the most popular TTS engines is Eloquence, which in fact is one of the least natural-sounding TTS voices around – but that has the advantages of working very well when speeded up and having very high latency (at least that is my perception). In time he will find that things like voice latency really matters as your brain gets used to this new way of working. In any case, Eloquence may not sound like a first choice when you first hear it, but for the above reasons it is my recommended TTS for new users. When it comes to actual programming, I have found that most things can be done, although one often needs work-arounds, and there almost always is a workaround. For example, for data analysis I often use the R package RMarkdown, which makes it easy to e.g. render data tables as screen reader-friendly HTML (while R Studio has improved somewhat, it still isn’t fully accessible). I mainly use WSL and either VS Code or Notepad, depending on what I want to do – unfortunately many IDEs are only partially accessible – on the whole I have had to learn to do more from the terminal than before, including debugging. In any case, I get along just fine working on c++ projects and I’ve found few obstacles when programming in Python, Julia or R (although Jupyter notebooks are a problem). Incidentally, for a long time I avoided Python because I thought the indentation would be a problem given my lack of eyesight – but when I finally tried it I found that having the screen reader announce the indentation level actually works much better than I thought it would and these days I am perfectly comfortable in Python. As with many things, simply trying it, rather than being limited by my preconceptions about what is possible, was the key. Finally, there are online communities of blind programmers who will have already figured out many of the work-arounds your friend might need and who are generally very willing to help. An R-specific group has helped me a lot over the years – but there are many others. Good luck.