(Gu)Estimates

This post trigged something in me. I like to estimate tasks and projects but in the world of software I think it is near (damn right) impossible to estimate "correctly". An estimate is ofcourse still a Good Thing(tm) as you have no other way of controlling the (alleged) progress, but it is important to remember an estimate is still - an estimate. As a follow-up on my previous post I'd like for managers to remember - we (developers, architects and code smurfs) estimate tasks as we are the best one's to do so based upon our experience (in this area) BUT this doesn't mean management doesn't have a part of the responsibility... Well - I digress

JSR-170 and JSR-283

I once in a while browse the net to see if there are some news regarding the above two mentioned JSR's. I don't mean new as in updates for the specs but new implementations based upon the specs but there does not seem to be a lot.. I've started to write an implementation myself but it is still FAR from being finised and since I just became a dad for the second time just two months ago, I doubt if I'll ever finish off the project. I've come quite a long way with the basic blocks such as Session, Workspace, Node and Property but opposed to the other implementations I've found, my will be a client/server much like using JDBC except its JCR. Well - anyways I was just wondering why there only seems to be like three (read: 3) implementations available for now (atleast I know of):

Jackrabbit
Day's CRX (which I guess is build atop some Jackrabbit code (I might be wrong))
Exo platforms (Damn - I tried to figure out if its a standalone, but couldn't)

Why aren't the web scattered with implementations? I know, some says it is too little, too much but I find it rather interesting and good.. Hmm mayB it's just me... (back to the dipers)