Read correctly; I said extraordinary claims failed V. But the other reading doesn't seem to paint a pretty picture anyway. Every point made in the original article is a reasonable expectation by outsiders, and many of them are still unsatisfactory ("failed claims" if you like).
What you should do now is to decide what to do with those points instead of arguing. If the decision is WONTFIX (okay to do, not everything can be made into the language) then the advertisement should be updated (ProTip: you should really have done this years ago). If the decision is to do something with that then the advertisement should be still updated, hopefully with a link to the tracking issue. If you are already doing something about that then you still should have a link to the tracking issue. If the point is "misleading", then you should write out clearly why it's misleading, how the author could have concluded in that way (i.e. assume no malice and debug instead), and how to verify your updated claim. With no strings attached.
Honestly though these points have been iterated and reiterated years ago. I had a hope that you have learned (hard) from that past experience; my hope seems not justified.
Look, you are doing what I've just said not to do. I have nothing to say because the OP did most of claims already, you should directly respond to the OP.
And you are slightly altering the very claim you've already said; the OP specificially tests the claim that "V compiles [...] ~1 million (x86 and tcc backends) lines of code per second per CPU core", which I can easily verify on my machine (1m_helloworld.v took 25.5 and 15.5 seconds to compile under the same settings). To be fair these test files are edge cases you can easily dismiss (ProTip: you can make your advertisements more accurate), but edge cases show the weakness of your design and you should not confront them.
Read correctly; I said extraordinary claims failed V. But the other reading doesn't seem to paint a pretty picture anyway. Every point made in the original article is a reasonable expectation by outsiders, and many of them are still unsatisfactory ("failed claims" if you like).
What you should do now is to decide what to do with those points instead of arguing. If the decision is WONTFIX (okay to do, not everything can be made into the language) then the advertisement should be updated (ProTip: you should really have done this years ago). If the decision is to do something with that then the advertisement should be still updated, hopefully with a link to the tracking issue. If you are already doing something about that then you still should have a link to the tracking issue. If the point is "misleading", then you should write out clearly why it's misleading, how the author could have concluded in that way (i.e. assume no malice and debug instead), and how to verify your updated claim. With no strings attached.
Honestly though these points have been iterated and reiterated years ago. I had a hope that you have learned (hard) from that past experience; my hope seems not justified.
Can you elaborate how language features failed V? And what claims were extraordinary?
Like the language compiling itself in <1s? It's true and you can verify it yourself:
https://www.youtube.com/watch?v=pvP6wmcl_Sc
Look, you are doing what I've just said not to do. I have nothing to say because the OP did most of claims already, you should directly respond to the OP.
And you are slightly altering the very claim you've already said; the OP specificially tests the claim that "V compiles [...] ~1 million (x86 and tcc backends) lines of code per second per CPU core", which I can easily verify on my machine (1m_helloworld.v took 25.5 and 15.5 seconds to compile under the same settings). To be fair these test files are edge cases you can easily dismiss (ProTip: you can make your advertisements more accurate), but edge cases show the weakness of your design and you should not confront them.
12 replies →