tag:blogger.com,1999:blog-882821950355432699.post4289874705370640524..comments2009-08-30T22:01:55.109+10:00Paulhttp://www.blogger.com/profile/18224234643439645641paul.batum@gmail.comBlogger2125tag:blogger.com,1999:blog-882821950355432699.post-15930911294850582902009-08-30T22:01:55.109+10:002009-08-30T22:01:55.109+10:00Hi Anthony, thanks for stopping by! It was interesting for me to reread this post just now as FNH, as you may be aware, has just gone 1.0. I guess things turned out pretty well. The code base now has the necessary underlying structure; what little nastiness remains can slowly be refactored out. It was quite a learning process for me to see my rewrite successfully transplanted into the existing code base and I am sure FNH still has many more learning opportunities instore for me.Paulhttp://www.blogger.com/profile/18224234643439645641noreply@blogger.comtag:blogger.com,1999:blog-882821950355432699.post-2290254177800746502009-08-30T16:18:09.006+10:002009-08-30T16:18:09.006+10:00Developers, by nature, often find it very difficult to admit even the most minor mistakes. We are always able to justify "imperfect" choices in design, implementation, or process. I give you strong kudos for the very objective way in which you reviewed your experiences with your Fluent NHibernate prototype/branch. I very much agree with your ideas related to using the hbm object graph. I think that would be a very worthwhile modification. While you are correct that rewrites are often "fools' errands", I would also mention that the longer one waits, the more difficult it is to correct deficiencies in an application. This is due, of course, to the increasing number of dependencies that often exist upon a section of code. If you feel there are issues, it is probably best to sound the warning bell now and loudly and correct those items before it becomes impractical. Keep up the great work!Anthonyhttp://www.blogger.com/profile/15953156541631599636noreply@blogger.com