From 1e290deb5f670f8a65ee20fe8e095fe5786d1237 Mon Sep 17 00:00:00 2001 From: Marshall Lochbaum Date: Tue, 28 Sep 2021 20:16:31 -0400 Subject: Links to the TopAnswers language changes page --- commentary/stability.md | 2 +- docs/commentary/stability.html | 2 +- docs/spec/index.html | 2 +- spec/README.md | 2 +- 4 files changed, 4 insertions(+), 4 deletions(-) diff --git a/commentary/stability.md b/commentary/stability.md index daccc1e5..8d433370 100644 --- a/commentary/stability.md +++ b/commentary/stability.md @@ -2,7 +2,7 @@ # Is BQN stable? -The short answer is that code running online or in CBQN is unlikely to break. In rare cases we add experimental system values (the `•` things) before we're ready to commit to a particular design; read further or check on the forums if you'd like to know the status of a particular system function. +The short answer is that code running online or in CBQN is unlikely to break. In rare cases we add experimental system values (the `•` things) before we're ready to commit to a particular design; read further or check on the forums if you'd like to know the status of a particular system function. And we are still making some backwards-compatible additions of moderate importance to core BQN: see [this page](https://topanswers.xyz/apl?q=1888). So it's possible that that extra knowledge will be needed to keep up with BQN in the future, but only a small amount. I have thousands of lines of running BQN code including the self-hosted sources, website generator, and Singeli compiler. There are also now many BQN examples and REPL links [spread](../community/README.md) across the web which would be harder to change. So there is a strong reason to maintain compatibility for common or even moderately used features. Because BQN's been designed in a conservative way, avoiding fiddly decisions by keeping things simple, it seems that few compatibility breaks will be required. The history so far seems to bear this out. diff --git a/docs/commentary/stability.html b/docs/commentary/stability.html index 9e12ef34..6c23680f 100644 --- a/docs/commentary/stability.html +++ b/docs/commentary/stability.html @@ -5,7 +5,7 @@
(github) / BQN / commentary

Is BQN stable?

-

The short answer is that code running online or in CBQN is unlikely to break. In rare cases we add experimental system values (the • things) before we're ready to commit to a particular design; read further or check on the forums if you'd like to know the status of a particular system function.

+

The short answer is that code running online or in CBQN is unlikely to break. In rare cases we add experimental system values (the • things) before we're ready to commit to a particular design; read further or check on the forums if you'd like to know the status of a particular system function. And we are still making some backwards-compatible additions of moderate importance to core BQN: see this page. So it's possible that that extra knowledge will be needed to keep up with BQN in the future, but only a small amount.

I have thousands of lines of running BQN code including the self-hosted sources, website generator, and Singeli compiler. There are also now many BQN examples and REPL links spread across the web which would be harder to change. So there is a strong reason to maintain compatibility for common or even moderately used features. Because BQN's been designed in a conservative way, avoiding fiddly decisions by keeping things simple, it seems that few compatibility breaks will be required. The history so far seems to bear this out.

Various edge cases were fixed when I first ran the primitive specifications through unit tests(!) in February 2021. Since then there has been a single compatibility break, that is, change from one intentional (i.e. excluding bugs) non-error behavior to a different behavior.