![](https://secure.gravatar.com/avatar/e748cd2d3e4c2ee4c34925fcc1e1ebc7.jpg?s=120&d=mm&r=g)
Hi all, recently I finished typesetting a block of inset illustrations for a journal on history, which I first implemented in LaTeX and then moved to ConTeXt (I am using a version from TeXLive 2013 packaged with openSUSE 13.1). Basically I am greatly satisfied with the result, as the code looks much clearer now. However, I still see some things which can be done in LaTeX by a more natural way. First of all, is there a dedicated method to rotate an entire combination or floatcombination by 90 degrees? I see floats have a special option to control this, but float combinations don't (am I right here?). And if I enclose a float combination into a \rotate command, then the height of the rotated block seems to be incorrectly calculated, so that about a half of the block goes outside of the upper page boundary. Second, not all image combinations can be described by a simple matrix, and the documentation doesn't explain what to do in such more complex cases (nor the Wiki does). For example, in my document some pages have three images ordered in two columns: one relatively tall image side-by-side with two smaller ones. To make the things worse such combinations are usually placed on landscape pages. The obvious solution here seems to be to somewhow group two smaller images and then put them into a float combination. Unfortunately, this doesn't work, as floatcombination seems to ignore all enclosed boxing commands/environments and even nested float combinations. Finally I managed to achieve the desired layout by defining a set of parallel paragraphs, which fortunately worked even inside a \rotate command (columns and columnsets with embedded floats didn't). Still it would be nice to have a special interface allowing to combine floats and float combinations into larger blocks. -- Regards, Alexey Kryukov <anagnost at yandex dot ru> Moscow State University Faculty of History