Add doc/ dir, move docs there.

1 files changed, 35 insertions, 0 deletions

diff --git a/doc/limitations.txt b/doc/limitations.txt
new file mode 100644
index 0000000..cba6c9f
--- /dev/null
+++ b/doc/limitations.txt
@@ -0,0 +1,35 @@
+Things that won't be implemented:
+
+- Stupidly long nick or channel names. Libera has 24 chars for
+  max nick len, 50 for channels. Any more than that is insane.
+
+- Color. No sane way to do it on the Atari. Color codes will be
+  stripped before messages are displayed. You can *send* color
+  codes (Ctrl-C and numbers) for other clients to see.
+
+- Connecting to multiple IRC servers. While the FujiNet is capable
+  of multiple simultaneous TCP connections, I really don't want to
+  complicate things enough to make this possible.
+
+- Unicode/UTF-8. We're using GR.0 on an Atari, there aren't enough
+  glyphs available to make it worthwhile. Plus, decoding UTF-8
+  would take up a lot of space.
+
+- Support for XEP80 or VBXE, or in general any 80-column hardware.
+  Or software 80 columns, for that matter.
+
+- Extended IRCv3 stuff (tags, capabilities). I *might* try to
+  implement SASL, but it's low on my list of priorities.
+
+- PASS command during registration. If you're an IRCop, you
+  should use a modern full-featured client.
+
+- Log files. SIO bus doesn't have the bandwidth for writing to
+  disk while also doing TCP/IP. I suppose logging to something
+  like a RAMdisk or SIO2IDE (or other non-SIO hard drive) wouldn't
+  be impossible, but doing it right would be a PITA.
+
+- Triggers or scripting. At most, there will be autojoin channels
+  and a (very) few keyboard macros (e.g. you could define a macro
+  that does "/m ChanServ identify <user> <pass>", but you'd have
+  to press a key to send it).