While I like writing, I always feel capped by the blog engins.

The following keys are what a Blog engine should emphasis on (YMMV):

• be editor agnostic I do think everyone has his/her favorit editor, and therefore should not be forced to use some strange webinterface to write texts. (unless one is dead happy with that)
• support Math Some egines do support math by using dvipng or equivalent LaTeX to image converter, and that is probably the best and most practical soltuion so far as TeX is the de facto standard for writing scientific text
• support multiple Languages I do feel home in English. But I do speak German and bits of Vietnamese and Chinese. Why not support this? We do have technologies to lookup the prefered language from the Useragent, and obtain geographical information to some extent, why not utilize this? Django does it to a certain extent.
• support RSS/ATOM filter. Say I want to only read blog posts given certain tags, or tag combinations in certain languages? E.g. I can’t read Spanish (yet!). So why can I not exclude all posts in Spanish from the feed I read? And why should I read something concerning agriculture if I’m only interested in physics? Time is sparse, information plentyful

On another side, Spawning looks promising, I’ve yet to read about memory consumption, but maybe I can cut them even further on this machine.

So I guess the only valid way is to put together a new blogging engine that focuses on mature blog backends, (i.e. MT), support tagging, categorization, and a language property for every post. Have a filterable syndication framework, so everyone can syndicate his/her very own mixture of news.

In the end we want to provide information to those who are interested, and at least for me, I’m interested as long as the (to me) uninteresting posts do not outnumber the posts that I’m interested in.