Arch Linux

dav7

Member

From: Australia

Registered: 2008-02-08

Posts: 674

Writing a webserver/CMS type thing

So I've had a bunch of different content management ideas in my head for quite a while now, and think that it'd be a fine time to have a go at making them.

A few questions though:

- Since C is fast/tiny/etc, I plan to use the nginx webserver (or some other type of server, if I can find a smaller one that meets my needs) and patch it to have the actual frontend code built into the server itself. That should speed things up by a few million percent (:D). As a possible alternative to nginx, do you know of any easily hackable webservers that are FAST?

- The way my system will work is that there will be a tiny frontend (as I've hinted above) as well as a big, gigantic, enterprisey-ish (:P) compiler thingy that will go through the pages each time they're generated, tokenize them and save them. I don't want the parser to take a week (:rolleyes:) to finish reading each document because it's so slow, but I'm not so sure about using C here since I want my CMS to have a relatively comfortable control language without the code for the language being a headache to maintain... so what should I use here? I don't want to use C++, and I don't really want to use a scripting language either. Hm, redundant question... but you never know any ideas?

- The most important question: This CMS will define a series of formats that can be used for markup (eg, bbCode, Markdown, Mediawiki, and a few ideas of my own) and a second series of formats that will be used for output, such as plaintext, HTML, XML with XSLT/XSL-FO, and so on. The problem is, what intermediary format should I use? I'm considering XML since it's structurally unlimited, but I've heard that it's not the best, and that other formats pwn it. So, should I use XML, or should I switch something else? Considering the position whatever format I use will be in, it would be like switching OS or CPU architecture if I wanted to change it later, so this is one decision I don't want to mess up.

- Does anyone have any abstract ideas for CMS development in general that might help? Since my design idea is so different from all the others language-specific optimization or hack tips probably won't help me, but abstract hints will probably go a long way.

- I need a name. I've thought of stuff like "cms73" (too "web 2.0"), "n7" (nice at first, then I hated it), "{Web,Docu}{Compiler,Collector}" (no, not the name, parse it as bash) - gah, and I just... don't know what to call it. I want a name that sounds "Mac-software-ey", hints that this system isn't written in PHP, Perl or Python like all the others are (>.>), isn't "standard" in setup or usage, makes you think the 400MHz Celeron with 64MB it's running on is a 2.8GHz Xeon with 8GB, and is POSIX-only. In general, it's an odd assortment of C and whatever language the parser gets written in that... and here's the fun part... makes it possible to use any language in the pages that are produced, AND serve up pages in less than a second. Yes, that sounds a bit too good to be true and it is the way I've explained it, but it will work out that way. More or less.

-dav7

Last edited by dav7 (2008-12-23 10:59:27)

---

Windows was made for looking at success from a distance through a wall of oversimplicity. Linux removes the wall, so you can just walk up to success and make it your own.
--
Reinventing the wheel is fun. You get to redefine pi.