I have always presumed that the sometimes slow response of MO was due to two things (the immense size of their database) and relatively high activity on the site. With the sheer number of observations, I’m sure that the programmers have retuned their access methods and their data structures repeatedly and will always being working on those and related issues.
I have sometimes been unable to do what I wanted to do and thought I ought to be able to do. The “open” nature of the site, the rules (i.e., implementation) of dealing with deprecation of names), and the fact that somethings require extensive effort to accomplish (multiple stages that have to be in a certain order), real concern for data loss due to inexperienced users, etc. are all factors that drive the site’s development and maintenance (I’m almost certain of this although I have nothing to do with the site’s development or maintenance). This makes the unusual or abnormal hard to deal with for users (except perhaps the most experienced, which doesn’t include me).
Since I’m involved in building a large data-driven site myself. I can sympathize with the folks in the backroom at MO. What makes things worse is that if they come up with a new tool, they cannot take the easy way out and let the users test the tool. They have to do it on a separate copy of the system (for fear of zapping things in critical tables of the database that could bring the site to an undesirable condition.
So I can understand the conservatism and the commitment to an “open site.”
Very best,
Rod