Table of Contents
News for April 2005April 30Website updateI've been messing around on the website again.
April 20Calculation completeThere we have it; L3DT 2.3 can generate a 256-megapixel heightfield in just under 4 hours and 10 minutes using an AMD XP2600+ with 512Mb RAM. This includes all the standard overlays such as peaks, erosion, terracing and smoothing. Erosion in particular was very slow, but that's hardly a surprise. Now, how to render it? I may have to get back to you on that one… April 19Calculating, please wait...As I write this I am running a test on L3DT's dynamic caching performance. My poor PC is about two and a half hours into generating a 16,384×16,384 pixel heightfield, with at least another hour of erosion to go. The caching seems to be working properly, with the RAM allocation stable at 65MB and around two gigabytes of heightfield stashed to the HDD. I still have to tweak the cache settings, but I'd say the architecture is now all OK. I will post some benchmarking figures when I've completed a reasonable set of tests. More last-minute additions include:
Still to do:
Slowly but surely, we are approaching release 2.3. April 14Ramblings (continued)A big hello to the folks from the the World Machine Community, the Ashundar Terragen Community, and the Middle Earth DEM Project. There are some very exciting things happening out there, I see. As far as L3DT 2.3 is going, the development plan looks pretty well done. However, given the magnitude of changes between L3DT 2.2c and 2.3, I want to take a long time debugging before the release. Lets say May/June, shall we? April 1RamblingsWhile I feel strangely compelled to post some elaborate hoax message today, I will eschew such light-hearted frivolity due to a complete lack of creativity and a severe case of sleepiness. Anyhoo, I spent a bit of time today reading Joel Spolsky's Joel on Software articles. The excellent chapters on user-interface design set me thinking about improving the L3DT user interface and, in particular, switching to a design model that is based on activities, rather than features. The addition of an activity-centric 'wizard' could really flatten-out the learning curve for many common but unnecessarily complicated user tasks (eg 'I want to import a heightfield, then make a texture for it' - currently 13 mouse clicks). I'm planning on packing a wizard into release 2.3 in a somewhat minimalist form, with extensions to follow in subsequent minor revisions. What about a paper clip? Or a really cute kitten? Except where otherwise noted, content on this wiki is licensed under the following license:CC Attribution-Share Alike 3.0 Unported
|