On 29/08/2011 17:25, Arno Trautmann wrote: > And, maybe even more important, at least for > me personally, a kind-of-a-roadmap where the development is going. That > may be in three parts, short-term, middle-term and long-term. Maybe it > is my personal fault, but I have no idea of where L3 is supposed to go, > in which steps, etc. If it is the policy of the team to keep this > non-public, it is sad, but ok. If not, it should be made more public – > that doesn't have to be much text, nor must it be too concrete or > ultimate decisions. The L3 news are a good step in that direction, but > they mostly “only” report on what has happened. Roadmap-wise, things are rather diffuse. I guess that my (personal) view is something like Short-term ---------- 1) Potential re-implementation of l3fp in an expandable form. Bruno is looking at this (see l3trial on the SVN/GitHub for details). 2) Create some basic driver code such that l3coffins works in format mode (using a 'hacked up' output routine). 3) Look again at box affine operations, again to support coffins beyond LaTeX2e. 4) Take a look back over the 'niggles' list (the issues on GitHub), and try to deal with some of them. Medium-term ----------- 5) Get some form of decisions made on the galley: does it work, are we going the right way, etc. 6) Sort out issue of expandable versus protected definitions of document commands (in xparse). 7) Font selection (mainly a port of NFSS) and loading mechanism (fontspec-based plus something for pdfTeX). Long-term --------- 10) Output routine: get xor 'up to scratch'. ... Some of these are ongoing: for example, work on the galley has actually ongoing. To some extent, the divisions here depend on how significant the stuff is, as well as the likely time-frame. Once we get (5), (7) and (10) done, we'll be in a position to build a format which can typeset something. My view there is that this will start *really* basic, and probably also somewhat 'flexible', and that we'll then add 'stuff' as testing takes place. -- Joseph Wright