Show HN: E-- – A language you dial between English and Python

I had the same exact feeling. Python is so natural that I also feel it is hard to beat human reading-wise. When I first learned python, I remember thinking that the professor I was watching online was writing in pseudocode just to explain some sorting concept. Little did I know it was runnable Python. It was a laughable moment at my own experience at least. It was so easy to understand I genuinely thought runnable Python was a made up on the fly programming simplification. That's why I recommend to non-dev technical people to learn Python if they want to learn how to program. Now it's different I guess the Claude/ChatGPT/Grok/Copilot/Gemini does everything but still maybe you want to learn at least one programming language so you can officially be a bit more powerful than a vibe coder. That something can be Python.

I've been experimenting with Cucumber lately and I think it starts to make more sense now because you can write the .feature files first and have an LLM write the unit test code that interprets them. So you're actually writing English language specifications, with just enough mandatory syntax to enable some useful tooling like unit test reports and IDE navigation, but not so much that it's hard to learn.

But with respect to SMEs reading/writing them, I think we need to be realistic here. That concept of an SME died out a long time ago. How many projects still have business analysts on them? The idea that you have SMEs writing or even reading requirements, handing them off to an architect who hands a spec off to an army of programmers is an IBM era idea.

What we've seen over time as coding got easier is a blurring and compression of all these different roles into one super role: the Developer. Modern "full stack" developers are expected to be subject matter experts, coders, architects, UI designers, testers, and perhaps also know how to deploy their software and run it in production. Heck, a significant number are also expected to learn how to run a business and do fund raising from investors!

Historically, all these things were totally different roles and people. But even predating LLMs we were designing better developer tools which took away much of the manpower requirements, freeing up the time of people who were once full time programmers to take over the other roles.

And this coincided with a cultural shift where most people decided they did not want to be in charge of a software project at almost any cost. Software projects fail in mysterious ways that to non-technical people seem random and uncontrollable. Much better to outsource it or use a SaaS even if you lose flexibility, than experience the career death of being an executive responsible for a failed software project.

With the exit of SMEs and BAs stage right, developers had to step in to fill the gaps by both taking on the commercial risks themselves (bootstrapped SaaS) and also learning the subject they're trying to automate to the same depth as a genuine SME.

With LLMs making all of this stuff far easier still, I don't see any reversal of this trend. The successful SaaS solopreneur is the archetypical meta-competent man well known from the Victorian era, someone who has mastered many different domains, AI assisted or not. He will use languages at whatever level of abstraction make sense. Sometimes that will be English, sometimes code, sometimes reading and sometimes writing. But all the tools will be designed for that one type of person, not specialised roles.

Was confused about this as well. The whole thing sounds as if someone confused LLMs writing code and inspecting or executing that code - and then ran with it.

Still, the idea of mostly formal code with dedicated sections that the LLM fills in doesn't sound so bad - but also a bit outdated. It would have been useful a year ago, but not when coding agents generate the entire project in a coffee break.

Edit: The author explains it a bit more in another post: https://news.ycombinator.com/item?id=48863898

If the audience are people who aren't necessarily able to inspect the generated code, but still want more structure and determinism than pure "vibe coding", this makes more sense.

E-- != E

Also https://en.wikipedia.org/wiki/Amiga_E

It's like English... But with worse sigils than Python...

Context: 码 means "code" in Chinese and prominently appears in words like, e.g. 源代码 (source code).