Hacker News Reader

Comment by seanmcdirmid | Hacker News Reader

You are being a bit harsh. These were ambitious undertakings with high levels of risk going into them, which, I bet, were well known. I've struggled in this field for 10+ years now and there are lots of dark alley ways that end in walls. Then everyone constantly tells you this has already been done when, no, it really hasn't (at best, the technology is there but so piss poor designed that it isn't useful).

If you reign in your ambition, well, isn't that why innovation in the programming experience field is so stagnant (basically stuck at the Smalltalk level) in the first place? We should be free from disdain to take risks, possibly lose, with a chance of hitting it big.

View on HN

As a side note, and as someone else working in this space, you do yourself a disservice by telling yourself work in this space is stagnant since SmallTalk.

It’s not that programming ergonomics are stagnant, it’s that all of the gains have been made by professionals. And the tooling is nearly impossible to leverage for a beginner environment because all of the assumptions of a pro developer are built in to their design.

The innovations are there, Heroku, Git, CSS, Markdown... these are all triumphs of programming ergonomics. They’re just all inevitably coopted by professional “Foundations” and amended to the maximum level a Pro developer can handle. JavaScript is almost useless for beginners now because the tooling is so complex, but full time front end devs can crank out HTTP packets like no ones business.

So you will need to watch those developments and lean on their ideas and some of the low level tooling if you ever hope to build a beginner environment. But you can’t use the tools themselves.

Still, if you start with SmallTalk you will fail. If Chris Granger and Bret Victor started their and failed, you will fail too, because those guys are rockstars.

You need to take the ideas being tested in the Pro tools and use them, without adopting the Pro implementations or even the interfaces. It’s or easy.

None of those are interesting experiences beyond the command line. They represent exactly the stagnant thinking that we are fighting against. Especially GIT, it could have come out in the 70s with the interface it relies on. It makes me sad that people see this as progress.

It doesn’t take a rockstar to make progress here, and anyways, I’ve been relatively successful on the academic side (e.g. just got a 10 year influential paper award for my 2007 live programming paper). I don’t think Bret Victor failed so much as lost interest, and he accomplished a great deal in getting people interested in this area again. Chris hasn’t really failed either, he is still young. And let’s not forget all the ex-HARCers...

Also, not all of us see this as a make programming more accessible problem; e.g. that’s not my thing, I really want to improve programming en masse. But we all agree the existing way is a dead end and we need to experiment ei5h radically different approaches.

Interesting experiences beyond the command line are game development tools; Each iteration try to make the barrier between artists/content creators and programmers thinner.

https://www.youtube.com/watch?v=hXSKGYQvZsI

Indeed. Also interesting is a look at games themselves, especially those that show a long trajectory from beginner to experienced player with significantly different interaction patterns between the two groups.

It would seem that practically no programmer's tool has ever received the attention to interface detail that is common in the best games.

How is Markdown innovative exactly?

It took ideas from the 70s, dropped the interesting parts, and was hailed as a revolutionary approach to marking up documents. Ie, the past 30 years of computing have been about narrowing the interface between programmer and computer to the equivalent of a straw (everything as text!) and then try to build an entire system around that.

+1 - Watching from the sidelines I admire the work both from you and Chris. It is incredibly hard to move away from the current local maximum because of network effects.

Not only network effects, but path dependence due to the hegemony of our current hardware/software stack.

This is (in principle, at least) why we have public funding of basic research.

Which right now is mostly going to the deep learning train :(. There are some academics working in this area right now, Ravi Chugh, Phillip Guo, to name a couple.

Isn't generalized AI and even more ambitious long term goal though? Why program anything, just tell the computer what you want and he'll make it for you.

That's funny, since Engelbart's work in the mother of all demos was seen as irrelevant by much of the CS community at the time because they thought generalized AI was just around the corner! After a 2 or 3 AI winters, they changed their minds.

Generalized AI is like practical fusion, always 20 years away. I mean, it will come eventually, but until then...there is still value in making humans better.

> You are being a bit harsh

I know, and I rewrote my post several times before submitting it, trying to minimize the harshness. But the pattern with Granger is pretty obvious, isn't it? Grandiose ambitions, over-commitment, and a resulting failure to ship.

There have been plenty of people out there with grand visions who tried and failed the first few times. Then either they give up and tone down their visions, or they keep trying and perhaps eventually succeed.

LightTable and Eve are still very influential and inspiring even if they aren't successful. I am personally glad they existed, as they provide experience and lessons for future efforts.

> There have been plenty of people out there with grand visions who tried and failed the first few times.

Most of them a con men, selling something they know will never work, especially on kickstarter and the like.

This was nothing unique, it's something people have been trying to do since computers were invented and it has failed every single time. COBOL was conceived with the idea that people in other disciplines would be able to do their own programming.

When there is every reason to think you will fail and you ask people for money to try then you start to look a lot like those con men.

The problem I have with your statement is that you are attributing something to somebody about which you know nothing, based on almost no information.

A less asshole-y phrasing would be "I wonder if he had reduced his ambitions somewhat, it might have had more chance to succeed".

And what's the problem with that? I think he (and everybody else) is free to have whatever dream they want, since they did not try to deceive others.

It’s called research.

Reply on news.ycombinator.com