I'm not sure how many people come back and say how their long term experience has been.
As I said in the story, I wrote my own after Slash whipped me. More accurately, it ran like a pig on stilts. My solution ran for some time and worked fairly well considering it was a flat file database (not just one flat file: one for each story and for each batch of comments on said story).
Anyway, on a whim I installed scoop and I was surprised at how well it went on and worked. Also, I was able to move my ancient flat files data into the database fairly easy, which isn't a software feature, it is a design feature. Being that scoop is made of Perl goo, I was able to bend it to suit my needs.
Things I've changed:
I've pulled subjects except on my 'change the title' thread
mod'd some of the .pm files to change bits of the design of the site
wrote some boxes to add new toys to the site
mod'd the conf file so it wouldn't slurp the whole url into scoop (so I can have non-scoop pages like /phill)
Things I've not liked:
speed. I wish it was faster and the more hardware I throw at it the less it seems to improve. I've gone from a 166 Pentium to a 933 Duron and I've seen very little improvement.
lack of documentation. I'm pretty good at reading the code now, but I hate that I have to go through about four modules till I find the information I am looking for.
Things I've liked:
out of the box, it worked and worked well
site design changes (for the most part) are quite simple
general mod'ability is great due to the ability to toss a box of perl on any page
To be truthful, I tried to replace it recently. I thought I had found a tool that was fairly speedy and customizable. To my dismay, the database structure made a conversion more of a headache than needed. There were 4 tables just for comments! I went back to scoop, obviously, and I think I'll stick with that and anxiously await the Apache 2 version. But for the most part, I'd say that I really like scoop and I won't be replacing it anytime soon.
-p