I don't know about the author, but I think the whole web application stack of today is overly complex.
Auth can be done in request headers, POST-data or even as a part of the URI in a GET request, if you want to.
HTTP state management... You know, GET requests were supposed to not have any effect on state, yet cookies can be rendered invalid at any point during an HTTP session. As a human, it's pretty easy to know the difference because the webapp tells you most of the time, but managing state as a computer can be quite a complex task. And sometimes, it's all server-side! No cookies, or a "dumb" cookie only used to associate the user agent with a certain server-side state.
The DOM (not HTTP, but while we're at it) is a PITA too. To load a resource and build its DOM is far from trivial if you want full support. And don't forget to support VBScript as an alternative to ECMAScript...
URI's are fun too. Even the regex in one of the RFCs doesn't cover all the edge cases. Oh hey, look. An IPv6 link local URI... whoopie...
I'm not a hater most of the time though. HTTP and by extension the web is quite organic. It's like a rain forest. There's a lot of things in there, and a lot of it is redundant and some things serves no real purpose, but at the same time the bio-diversity is wonderful.
//Cato
Skickat från min Samsung Mobil
-------- Originalmeddelande --------
Från: Szymon Olewniczak <szymon.olewniczak_AT_rid.pl>
Datum:
Till: dev_AT_suckless.org
Rubrik: [dev] Why HTTP is so bad?
On the
http://harmful.cat-v.org/software/ I've found very interesting
statment about HTTP: "or best of all: don't use HTTP". Can someone
explain me why the author claims that? What is bad in http?
Regards,
Szymon
Received on Wed May 22 2013 - 20:37:47 CEST