[Haskell-cafe] Type system madness

[Andrew Coppin]

Bryan O'Sullivan wrote:
> Andrew Coppin wrote:
>
>> Oh well, the problem is easily fixed... *sigh*
>
> I doubt that anybody minds having you talk about Haskell.  You've been 
> responsible for spawning a lot of interesting threads.

[And that one about compression that's still going on somewhere... lol!]

> All I would suggest is that you take your cue from the other people 
> who post to the list, and try a few tactics before you post:
>
> - If you have a chatty one-line comment, do 2,000 other people need to 
> see it?

Mmm. This is why I prefer NNTP. (If a thread becomes tangent, everybody 
just marks it "ignore" and they don't have to waste time downloading it 
or reading it.) But yeah, point taken...

> - If you have a question to ask, try to spend 2 minutes with Google or 
> the Haskell wiki to find the anwer.

I continue to be surprised at the things that don't seem to be on the 
Wiki... Google is typically no help at all with anything 
Haskell-related, because Haskell is so completely obscure. The various 
haddoc documentation is also frustratingly sparse in places. (E.g., 
Control.Concurrent.STM.TVar contains *nothing* but terse type 
signatures. And concurrent programming is already a tricky thing to get 
right.) Some things seem to be "well known" yet not actually written 
down anywhere - e.g., the finer points of using "seq" to make stuff go 
faster.

But sure, I do like to try to puzzle a thing out first before posting 
here. (If something else, I kind of enjoy a challenge...)

> - Join us on #haskell on IRC.  It's extremely chatty, and you'll be 
> welcome.

Not in my experience, no.

(Maybe I ask the wrong way... but almost everybody seems to simply 
ignore me. Actually, usually when I go there absolutely nobody is 
speaking at all. What time zone do these people live in?)