Allen Gilliland wrote:
i think efficiency is probably the most difficult part of this approach. i really don't know what effect there would be if we were allowing data to be retrieved during the velocity rendering process. i would guess it's fine, but we would need to play with it. from a conceptual standpoint i would like to think that all the relvant data could be retrieved through the "weblog" object. i.e. something like (in velocity) ... weblog.getName() weblog.getHandle() weblog.getEntries() weblog.getCategories() weblog.getFolders() weblog.getBookmarks() and on and on.
I think this is what they call the "Open Session in View" pattern in the Spring/Hibernate world.
Perhaps a template could have an optimization hint or a pre-fetch setting that could be set for efficiency in the built-in templates. New templates could default to the pre-fetch nothing setting (or the pre-fetch everything setting, I suppose, because either could have a negative impact on performance), and could later be configured/optimized if performance became an issue.
It certainly was handy (for me) to be able to use the Flavor servlet to develop the RSD template, even though it was later moved to the optimized RSDServlet that you wrote.
yeah, i am pretty familiar with MT and i do like the way it does templates a little better. unfortunately we are bound by velocity, which can't do the custom loops/conditionals.
See my message on the developer list for some ideas on this... -- Sean
