The scope of their language is very narrow. That is basically a configuration language for their SDK. One does not need to be an expert in language design to do that. To me it looks much cleaner than their TypeScript configuration, but comes with drawbacks they listed in the blog post.
For what they wanted to do (basically auto-generating boilerplate of modern web-dev), inventing a language probably was not neccessary. Naming their thing "wasp-lang" was really a misnomer since that is not their actual value proposition, if I understand correctly.
That's true, although I remember when they first started publicising this project (before the VC/YC push), and the idea was pretty much always "this is a programming language specifically designed to do web dev". In fact, I think the earliest versions had none of the resource management, it was just a combination ORM/webserver/templating engine wrapped up in custom syntax.
It sounds like they themselves only figured out the value proposition relatively late on, after they'd gone down the "web dev language" route, and then pivoted, first with the language, and then later by taking the language away and focusing on the configuration aspect.
Yeah, I wondered if their marketing problem (not their IDE maintenance problem) would have been solved by just renaming it to wasp-dsl. That kinda implies that the custom syntax is used only for a narrow purpose, and not everything you'd want to do.
I disagree. There's obvious marketing aspects to the post (the frequent interjections that Wasp works best with AI), but by and large its a good and through walk through the hidden complexity that actual production languages have beyond the design. If at all, I find it a bit rambly.
I also don't think it can be held against people with a CS background (where basics of programming languages are taught) that they started designing a new language. That is one of the task that most people do once in their lifetime, at most twice. There's not apprenticeship in language design. A lot of successful languages nowadays are their designers first shot.
Not knowing that 5mio is by far not enough of a budget for bringing a language to the world is also not to be held against them: there's very few disclosed budgets of language development costs over all aspects (I would even say "none", but I'm not sure if I missed something).
To be clear, I meant they marketed the product to investors with no relevant backround. I don't mean this post is good marketing or that this is a marketing post.
There's not apprenticeship in language design
Not sure what you mean by this. Many of us use many languages, then design many small/toy languages, study PLT as a formal discipline, contribute to languages/compilers, work/intern with the language teams (in FAANG or others), ...
Not knowing that 5mio is by far not enough of a budget for bringing a language to the world is also not to be held against them
Anyone who contributed more than two commits to any real programming language would tell you how many engineer/hours it takes to ship even the smallest features.
osa1 | a day ago
2 people with no experience and credentials in designing and implementing languages decide to invent a new language
Then they convince a bunch of people with money to invest millions
One of those stories that show that it's 99% marketing. Nothing else matters, practically speaking.
the-lazy-guy | a day ago
The scope of their language is very narrow. That is basically a configuration language for their SDK. One does not need to be an expert in language design to do that. To me it looks much cleaner than their TypeScript configuration, but comes with drawbacks they listed in the blog post.
For what they wanted to do (basically auto-generating boilerplate of modern web-dev), inventing a language probably was not neccessary. Naming their thing "wasp-lang" was really a misnomer since that is not their actual value proposition, if I understand correctly.
Johz | a day ago
That's true, although I remember when they first started publicising this project (before the VC/YC push), and the idea was pretty much always "this is a programming language specifically designed to do web dev". In fact, I think the earliest versions had none of the resource management, it was just a combination ORM/webserver/templating engine wrapped up in custom syntax.
It sounds like they themselves only figured out the value proposition relatively late on, after they'd gone down the "web dev language" route, and then pivoted, first with the language, and then later by taking the language away and focusing on the configuration aspect.
strugee | 19 hours ago
Yeah, I wondered if their marketing problem (not their IDE maintenance problem) would have been solved by just renaming it to
wasp-dsl. That kinda implies that the custom syntax is used only for a narrow purpose, and not everything you'd want to do.skade | a day ago
I disagree. There's obvious marketing aspects to the post (the frequent interjections that Wasp works best with AI), but by and large its a good and through walk through the hidden complexity that actual production languages have beyond the design. If at all, I find it a bit rambly.
I also don't think it can be held against people with a CS background (where basics of programming languages are taught) that they started designing a new language. That is one of the task that most people do once in their lifetime, at most twice. There's not apprenticeship in language design. A lot of successful languages nowadays are their designers first shot.
Not knowing that 5mio is by far not enough of a budget for bringing a language to the world is also not to be held against them: there's very few disclosed budgets of language development costs over all aspects (I would even say "none", but I'm not sure if I missed something).
osa1 | a day ago
To be clear, I meant they marketed the product to investors with no relevant backround. I don't mean this post is good marketing or that this is a marketing post.
Not sure what you mean by this. Many of us use many languages, then design many small/toy languages, study PLT as a formal discipline, contribute to languages/compilers, work/intern with the language teams (in FAANG or others), ...
Anyone who contributed more than two commits to any real programming language would tell you how many engineer/hours it takes to ship even the smallest features.