ConTeXt is a good software. I was impressed with a manual typesetted with it and decided to give it a try. After some small documents, despite my ridicously small knowledge of it I decided to use ConTeXt for my thesis. Bad move. ConTeXt is beautiful because it's very configurable, for the most uses doesn't need any external module (unlike LaTeX), and gives me more power on the presentation of the document. But in a few days of work I've come to realize that it's not suited for me and this project. I've had some unpleasant surprises: - Bibliography doesn't work the way it should on MKIV. Unlike MKII. - MKIV, unlike MKII, doesn't setup any background color: ---- \setupcolors[state=start] \setupbackground[background=screen] \setupbackground[state=start] \def\quotebox#1#2 {\blank \midaligned{\startbackground \quotation{\em #1} \crlf --#2 \stopbackground} \blank} ---- - the above \quotebox command, in some cases, sends pdftex and luatex (MKII and MKIV) to the moon with an infinite loop (100% CPU). One reason for this is probably my very limited knowledge and experience with ConTeXt. It's extremely configurable, and this is a plus. But on the other hand if you don't know how to move, what to do, the system internals and how any configuration affect the typesetting, a lot of thing won't work. And worse, you won't know where to bump your head. I don't have time to read the reference manual (I've already read the excursion) because the deadline is too near, so I have to call defeat and go to the LaTeX camp. I've wasted days of work, now I can't afford more of that. It's a learning experience: don't use an "experimental" (new for me) tool for an important job. Stick to the tried&tested ones, and use new stuff only in a safe context (without a near hard deadline). Maybe, in some future, I'll take again ConTeXt and try to learn it the proper way. But for now, it's "fired". For your patience, time and help: thanks to all of you! A little wish: I hope that when I'll return here there'll be good documentation (user-oriented) and a single (througly tested, see TDD/BDD approaches) reference implementation. -- Manuel P.