And yeah, the Apache 1.3 dependency was the first thing we did (it sort of had to be) -- down the road, LJ even used our implementation of it as a reference for their own work. (They got off the 1.3 tree somewhere in 2009, I think.) It was painful.Then there's JQuery-ifying the site, and converting to Template Toolkit, and half the code wasn't even OO, and just, ugh. Ten years of deferred payments -- and that's just on the backend; I'm not even going to get into the mess that half the frontend is in.
Sometimes I look at all of it and get totally overwhelmed by just how much we have left to do, but then I stop and make myself look at how far we've come. We've actually diverged from LJ so much that we can't adopt their patches wholesale anymore; we need to convert them to work on our end. If that's not a sign of how far we've come, I don't know what is!
memcached: man, some of the things Brad came up with were total duds, and his code is occasionally a bunch of sheer wtf, and rolling your eyes and mocking Acts of Brad was a common pastime even back in the LJ office of 2004, but in his own way the guy is a genius, and the things he came up with in order to keep LJ running on a shoestring from 2002-2005 or so, before Six Apart could just throw more hardware and people at the problems, were occasionally works of art. I learned more about architecture from watching him than I ever realized. I really do hope that the historians of the future give him the credit he really deserves.
no subject
And yeah, the Apache 1.3 dependency was the first thing we did (it sort of had to be) -- down the road, LJ even used our implementation of it as a reference for their own work. (They got off the 1.3 tree somewhere in 2009, I think.) It was painful.Then there's JQuery-ifying the site, and converting to Template Toolkit, and half the code wasn't even OO, and just, ugh. Ten years of deferred payments -- and that's just on the backend; I'm not even going to get into the mess that half the frontend is in.
Sometimes I look at all of it and get totally overwhelmed by just how much we have left to do, but then I stop and make myself look at how far we've come. We've actually diverged from LJ so much that we can't adopt their patches wholesale anymore; we need to convert them to work on our end. If that's not a sign of how far we've come, I don't know what is!
memcached: man, some of the things Brad came up with were total duds, and his code is occasionally a bunch of sheer wtf, and rolling your eyes and mocking Acts of Brad was a common pastime even back in the LJ office of 2004, but in his own way the guy is a genius, and the things he came up with in order to keep LJ running on a shoestring from 2002-2005 or so, before Six Apart could just throw more hardware and people at the problems, were occasionally works of art. I learned more about architecture from watching him than I ever realized. I really do hope that the historians of the future give him the credit he really deserves.