Seems like the dominant sentiment is that the maintainer is being defensive... I'd rather like to think he's being up-front and clear about his intentions for the project.
Indeed, and the polish is what makes Clef awesome and a pleasure to use. Their Wordpress plugin is a great way to drive adoption, but I find myself wishing more sites supported it.
This (incorrect assumptions leading to frustration) is definitely the most troubling thing to me about automagic solutions. I'm currently working on a project that incorporates this idea of sorting email into contexts (http://solnovus.com, not much info up as of yet), so we've been thinking a lot about this problem specifically.
We are working in user-defined contexts from day one (not all users fit into the 5-context model), but I think incorrect classification is the most important issue to address to minimize user frustration. We are approaching it by treating topic and context membership fuzzily and using graph visualization to aid in traversal of inexact queries.
I couldn't agree with you more: there are serious problems with binary membership and "automatic" filtering.
I agree, you can construct a pretty flexible search system with the existing Gmail implementation. (disclosure: I work at a small startup trying to address email search and archive organization). How many rules do you have set up to achieve your (arguably very reasonable) workflow?
However, to perform a successful search with specific constraints, you must either:
1) Remember a keyword
2) Have a label for the particular constraint you want to use
Our little startup, SolNovus (http://solnovus.com), has been investigating how to make a couple power-user functions like this more accessible. Our priorities are:
1) Fuzzy topic search: make keywords more forgiving
2) Analysis and visualization of email graph: sift through related threads and see the impact of search constraints, represent scale and organization with more than an inbox counter
They made a pretty striking case that the police exhibited gross negligence, I suppose. Although they didn't do it in a way that has the weight of law behind it, they did succeed in getting the case reopened.
If Anons truly received a flood of testimony after posting on some forums, law enforcement must have pretty much actively avoided hearing testimony.
The photo/video-sharing aspect of this case is particularly interesting; not analyzing available data from the internet (or simply not even knowing photos/video pertaining to the case are being widely distributed) is likely going to be a focal point, but heightened law enforcement use of evidence from the digital domain is probably touchy for many of the same Anons that contributed to this.
I'm glad this has gotten the attention it deserves; I purchased the beta when embarking on my first Django project, and it was a bit over my head at the time. Now, however, 2SoD fills the important role of an up-to-date best practices guide, which I find is often hard to compile in largish community projects with fragmented documentation.
Not that the official docs aren't good, because they are; the 2SoD authors just get to be a little more narrow and opinionated.