I like what I see and of course I had to optimize it for my very peculiar taste:
html
head
title My web page
meta name=this content=that
link rel=stylesheet href=style.css
script src=jquery.js
body
header
logo
img src=logo.jpg
main
content .center
h1 Welcome to my site
p
This is my first post
Feel free to comment at the end.
pre #hello
a=1+2+3
print a
footer
p no copyrights at all
Assume every first word is a known tag or convert it to div with id. Use significant white space. Content tags like 'p' or 'pre' are closed by a blank line.
It may work for 90% of my own projects and it is by far the simplest markup I have seen, evar.
No no that's just a quick brain dump just to see if it could be optimized a bit further. I love simplicity and always try to simplify stuff just for the fun of it.
If you do get the time, check out the code and see if you can tweak things to your liking.
I'll be doing a cleanup/refactoring later this week to make things better.
However, the core of Nemos' parsing is handled by pyparsing, so if you can read and write eBNF, then you can tweak Nemos' syntax to your liking pretty easily.
I should probably mention. Nemo is built on top of Mako, so there's a tiny speed bump for Nemo processing then its as fast as Mako once the templates are compiled and cached.
Mako is faster than Django templates--at least by a factor of 4--and in the neighborhood of speed with Cheetah. Only Spitfire templates are faster.
This is important for any app that is CPU bound due to rendering speed.
A while back, I wrote a very small Python HTML rendering library that gives some of the benefits of this notation while still being Python source code. You _may_ find it interesting if this is the sort of thing you find interesting...
To me, it seems that everyone is striving to find an easier way to programmatically create HTML in Python. The current range of templating languages just don't cut it.
I think there's no definitive solution partially because everyone is copying Haml, and Hamls' syntax and trade-offs aren't really appropriate for Python users.
I've been contributing to another HAML derivitive for python called HamlPy (https://github.com/jessemiller/HamlPy). It compiles down to Django templates, which made it very easy for me to convert my existing project. It's under pretty active development right now.
You might want to update your readme file so everyone knows that it compiles down to Mako just at a glance.
Right now its confusing.
But if I'd known about this project a few days ago, you'd have definitely been added to my list of Muses.
Also, does your fork support the full range of Mako control structures? Looking at the nodes.py it seems that its still targeted towards a Django-templates backend.
Yeah, I quickly hacked it together, I will update the readme soon. It does support the mako control structures. You see them rendered around line 200 in nodes.py.
You can actually use Nemo alongside Django-templates.
It registers a Mako template loader with Django, that by default points to your /templates directory in your folder.
However you can override it by modifying MAKO_TEMPLATE_DIRS in your settings.py. MAKO_TEMPLATE_DIRS is just a tuple containing the directories to look under for Mako/Nemo files.
My full project runs under Mako, but 3rd party apps, and the django admin run with django-templates.
I don't want to talk badly about it--because to be honest it has some really cool features like chained-method-calls that allow you to quickly compose HTML attributes via inheritance----but Mint has always struck me as more Perl than Python.
% p .'title' title='${doc.title}' || ${ doc.title }
I don't know why Nemo looks better in my eyes but I think it has to do with explicit separation, and string substitution instead of a method-call syntax.
Mint feels like a programming language that happens to produce HTML as a side-effect.
It may work for 90% of my own projects and it is by far the simplest markup I have seen, evar.