Gearing: Hannibal_Smith: people had a lot of wrong ideas in the 70s
Bacich: Dash: Coq isn’t a programming language.
Gearing: Hannibal_Smith: also you can certainly write web apps in C or C++ but, of course, don’t
Gearing: Havvy: Madsy was saying it was.
Wright: Dash, yet the only response to C or C++ is only today with Rust
Silton: Dash: I wonder. But if we make all code sharing single functions, rather than modules, maybe it’d work. I haven’t really thought much about it so.
Michealson: Dash, personally I don’t like when a people focus too much on the web part of the programming
Alman: Ignoring all the other things
Gearing: Dolby: how do dependent types help with network protocol versioning
Pavia: Dash: you can encode the protocol interactions in your types.
Gearing: Hannibal_Smith: in the grim future of webhammer 40k, there is only http
Gearing: Dolby: and then what?
Shafe: Dash: But HTTP is going away due to HTTPS.
Fossen: Dash: you choose the right one you want to talk to
Gearing: Dolby: how’s that different from what we do now, though
Standfield: Dash: in this case, it would not. Searching for a well-known protocol version is going to be much easier than encoding it in the type. Mostly I was referring to the semantic versioning approach to software libraries, though. Which may not be necessary with dependent types
Doepner: Granted people write the right proofs, which is another problem
Nyswonger: Dolby: Yeah, if we only could prove that our specifications were correct 😛
Kimmes: That’s yet another problem 😛
Kamna: Hi, as #do***entcloud is dead I was wondering if I could post a question backbone/require.js related here
Marrello: Pthreat: you can. If someone knows about it, they might answer. Or they might not. But it might be worth a shot anyway.
Woon: Isn’t require.js pretty much dead though?
Mcginness: No idea, I just have to work with this. So I’ll go ahead and ask, http://pastebin.com/raw.php?i=H62QE4C5
Balcom: At line require’another/script’; the events contained inside that script do not fire
Sarrett: Dolby: what’s the alternative to require.js for a browser though if you’re not yet on ecma 6?
Mezera: Sarrett: CommonJS and Node modules, which is what everyone uses.
Sarrett: Then wire it togther at compile time using browserfy?
Klapperich: Sarrett: or WebPack, or any of the several tools for that.
Sarrett: That’s still on my to learn list, gotta work out how to make that play nice with VS2015 and typescript
Apuzzi: Sarrett: you still need to add the reference comments for typescript
Sarrett: Dolby: even if you compile typescript to CommonJS modules?
Shillinger: Sarrett: unless TypeScript extracts the type information from require’foo’, you need to include the reference information for all types you use
Sarrett: I think it might, I’ve noticed it tends to consider all the scripts in the same director as already included as well
Sarrett: Probably to do with the change towards tsproj.json files
Mottram: I have a function in javascript which uses only the slice function, and calls itself. If the input string is too long it gives an error but for smaller strings it works fine. What could be the problem? Does slice have a limit on how long the string is? Or because the function calls itself it exceeds some maximum time?
Baul: There’s a recursion limit of ~1024 in most engines.
Schoenle: And proper tail calls hasn’t been implemented yet.
Sas: I have some code wrapped like functiona, b{.stuff.}foo, bar;
Barrom: If i put var a = foo, b = bar; in front and remove the wrapper, it breaks
Irias: But i want to be able to fiddle around with the stuff defined inside after i load it