Good question Kartik Agaram. It’s not just one paper, it’s when there needs to be a series of papers over years to work out an idea. For example Functional Programming was an academic curiosity for over a decade. More recently CRDTs took a decade to get to their current popularity.

I suspect that a big reason for accessibility being neglected in computing is ignorance. People who can't participate in the Web as it is today are also invisible to those who can and do, so we don't know what exactly their issues are nor what we could do to improve the situation. It's a kind of digital divide.

FWIW, halfway through the last paragraph I was thinking “cool, now we’re done with the front matter, let’s get into this juicy accessibility vs adaptability thing!” – and then the post ended. So take that as a sign that, if you were to write more on this thought (or if you already have?) I am interested to read it.

Thanks Joshua Horowitz! I haven’t really “finished” the thought there, but I have written more about accessibility, all very work focused:

• adhoc.team/2024/02/21/scaling-a11y-va
• eli.li/accessibility-and-the-product-person
• eli.li/accessibility-first

So much wisdom in there! At first sight it might seem as just another regurgitation of what he has been preaching for more than two decades already. But he’s always trying to evolve how he tries to make his points.

I hope he’ll be with us for a few more decades and keeps reminding us of these things. Although I can’t really understand how he is still willing to put up with us, as we have clearly decided to go in a very different direction.

Stefan Lesser In the words of Joseph Heller, "What else is there?" 🙂

Came across this transcript of another one of his talks. This one’s very different, but together they draw a pretty good picture of what he’s trying to do and — more importantly — why. tinlizzie.org/VPRIPapers/m2004002_center.pdf

The problem with AI (and technology in general) in one slide:

“Wisdom requires a lot more context than building.” — Alan Kay

There are a couple of things going on with “context”. As a technologist it’s a bit too easy to get too exited about science and the abstract. But talking all the time about technology and potential can lead to misunderstandings.

Andreas S Please elaborate!

It will be very difficult but I can try. Maybe I should participate in one of those Christopher Alexander Events you linked before. I need to find that link. I can only write from my experience. So first of all a bit about my context about Alan Kay, here is a link to my Main Zettel File about Alans work: andreass.gitlab.io/2022/AlanKay.html . So Although alan has tried to create some form of context for Computer Technology it did not go to places where it may be should have gone.

When I first encountered the STEPS project I was deeply impressed with its technical ambition. But what happend with the final report? There was no coherent source code release of the work. Some researchers did put out some code in some form but it wasn't nothing like the system Alan showed, Why wasn't that put out to the public? Was it because of funding issues that led to troubles with research goals and communicating the whole thing? That still feels very unsettling, on the one hand Alan participated in xerox parc or he talks about Martin luther but then he and his research group fail to think about ways how to actually communicate their technology which could provide paradigm shift potential. You can see some of the strange in this github issue comment thread: github.com/damelang/nile/issues/3

Over the course of many years (!) people asked about ways to understand even this graphic sub system, let alone the bigger systems philosophy of the STEPS project.

So at some point the idea what research means or that you can have a idea on the level of paradigm shift and be able to share it. Some aspects of it you can see here where alan gives a talk but then there happens some actual discussion after that: youtube.com/watch?v=PFc379hu--8 the video is 1:20 , the relevant discussion happens about 1:17. Where you can see how confused alan is about what the students perception is what he is doing at university - could be or should be doing. I think with the years after the STEPS project Alan slowly realized that it is not only difficult nowadays to create something novel and hope or work towards a unfolding of that paradigm shift even talking about becomes difficult such that it enters the teritory of cargo cults (see Alans Conversation with Gardner Campell for context about that) I think this can be seen in this talk for example: youtube.com/watch?v=j9ZGFaIHegE. In this talk from 2019 Alan talks on a conference for circular economy for the alan macarthur foundation. He is basically asked to explain how he as part of xerox research community did achieve that, how did they initiate a paradigm shift ( from pre-personal computing to personal computing). ALan starts of by presenting a list of question and basically asks for a participation on these questions. But after his talk the moderator seems confused that alan did not "simply" deliver "the answer". While all, at least , to some part part alan was trying to was pointing to that even the structures to talk about a problem ( here working with paradigm shifts) matter and that giving a "talk" can not do that, it misses the point. All of this was pre covid and the culture wars have made cultural understanding even more complicatedon so many levels. I think in some sense its almost impossible to work on technology without work on better understanding of ideas and I do not mean like: oh you "just" have to build a better interface which is more intuitevly. Its more work in terms of relationships. See the video from alan called: Portable Portrait: Alan Kay (1990) from my file above to get a feeling about it. I'll stop here for now.

Andreas S Thanks for sharing that! Looks like you’ve been following him for quite a while. Impressive. I thought I had watched pretty much every video with him, but you easily pulled several videos out of your hat I hadn’t seen before.

I’d be delighted to have you join the discussion about simplicity next Thursday.

It sounds like your criticism is primarily based on not making research publicly available, or at least not doing that properly, like in the case of STEPS. Is that a fair conclusion?

I am as frustrated about that as you are. I don’t know the exact details of what happened to STEPS, I’m more familiar with the funding issues around Bret Victor’s Dynamicland, which also doesn’t exactly share much of what’s going on there in the form of public artifacts. However, if you were lucky enough to visit and talk to Bret and all the other researchers, they couldn’t be more open about what they do, how they do it, and what they are trying to achieve (at least in 2017, when I got to visit a few times). There seemed to be similar dynamics at play in both projects.

I’m not sure I understand your later point. Is Alan not doing something that he should be doing? Or is he doing something he shouldn't be? Are there expectations that he doesn’t meet? Or is this more a big picture point about culture that you’re making?

Yes funding is a major part of the problem. Like described in the nile issue github.com/damelang/nile/issues/3 It comes basically to the point where dan amelang says he would need more then 4k per month to work on the project, weather that means only coding or also explaining things he leaves it open. It was arround that time that I became aware of the whole crypto bubble. Humongous amounts of money got "spent" there. Ethereum also had for some time the label - programmable money. It seems to me as a real tragedy that a few thousands of dollars could not be gathered. Is it a marketing problem? Is it a problem of norms and values? Is society unable to fund its own sustainable progress or computing culture? One of the projects that did things in a similar scope and that actually succeeded is urbit. 15 years of evolution : youtube.com/watch?v=DFeN-TKZor0

or this resource urbit.org/blog/hoon-4-lispers They apparently found funding but also managed local user groups which maintained a kind of local "hacking" culture.

Interesting that crypto is your prime example for misspending of money. As I understand him, one of Alan’s main points is that governments/academia messed up funding after ARPA, and then companies messed it up after PARC, because they began demanding at least rough ideas of how such “investments” will be recouped. And now “humongous amounts” of money are “misspent” in the form of government grants, academic funding, and venture capital that makes all the money ever converted into crypto still look like pocket change.

Isn’t Urbit also somehow connected to Peter Thiel? That would probably explain why it worked for them but that probably means there are some skeletons hiding in the closet…

Clearly, money shouldn’t really be the problem to fund good research. The problem is that all the people who have that kind of money want even more of it, and as close to a good story you can come up with of how you will make them more money if they lend you theirs. And that is incompatible with what researchers need to do good work. And that is essentially the main story I get out of pretty much every Alan Kay presentation in the last two decades. So I think he would agree that, yes, as a society we are unable to fund this. For pretty stupid reasons.

(Christopher Alexander tells an analogous story about contemporary architecture, which also has chosen to follow different norms and values and to no longer care about what Alexander thinks should be the priority — designing spaces that serve the communities that live in them. Instead, architects care about how their skyscraper helps define the skyline of a city and makes them and their architecture firm famous and prosperous.)

I just read through the HN, some people in the thread claim to have a STEPS system running:

news.ycombinator.com/item?id=39636074 it’s almost tempting to try to bring this back to a bigger audience.

Yes, that would be great! And it should be possible. I read a lot of the VPRI papers and even if they are more abstract and there’s little source code available, the concepts are pretty solid and there are enough clever people out there who could certainly “fill in the blanks”.

What I would like to see even more is that thinking that motivated STEPS, the economy of creating something incredibly useful with orders of magnitude less complexity, and not being afraid to reinvent some of the foundations, spread to more projects. Instead of just taking what we are given, in the form of existing platforms, libraries and frameworks, and exclusively trying to innovate on top of them.

Alan Kay answers questions on Quora. 638 so far. This one is pretty related to this thread: quora.com/What-does-Alan-Kay-think-independent-researchers-in-computing-should-do-for-money-if-commercialization-was-part-of-the-problem-What-if-good-funding-for-research-isnt-available-but-your-ideas-are-worth-pursuing

I watched huge number of Alan Kay talks and this one is just repetition of what he said before…

The problem with AI (and technology in general) in one slide:

are there ways you see solvespace carrying this torch more than other CAD systems?

Nope, I'm not familiar with the space! Just saw the demo on the site.

Joshua Horowitz I'd love to hear about other alternatives you like or use. I'm just starting to get interested in this space.