... the standard category of a programmer who doesn't work well with non-programmers and sucks at writing specs/documentation.
That is an extremely rude way to characterize a volunteer who has single-handedly saved volunteer-centuries of time and then taken a principled, non-zero sum stand to offer agreement to the FOSS release requirement on Wikimedia Labs in return for the resources necessary to build new systems. Without the ability to experiment in an unfettered manner, "Labs" should be renamed "Unpaid Intern Server Farm."
the resources necessary to do it. (probably upwards of 25-30k dollars)
Could we please see an itemization? Is Dell's surveillance-compatible premium hardware still being purchased by the Foundation? What is the current state of the art at the Foundation for redundant arrays of supposedly inexpensive disks comprising network attached storage at present, and how much does the Foundation spend for it?
many organizations share our mission, some (like IA) quite eager to push the boundaries of the law in ways that improve global access to information
For values of "push the boundaries" equivalent to "conform as closely as possible to both the letter and the spirit and quickly respond to legal requests for abridgement when requested." Oblique suggestions that IA has any more legal risk than the Foundation are not productive. See _Field vs Google,_ US District Court, District of Nevada, CV-S-04-0413-RCJ-LRL. The Foundation has repeatedly failed to support IA, webcitation.org, and the like, even though our volunteers depend heavily on them for conscientious improvements.
... insistence on a more formal proposal, and a deliberative process of some kind to evaluate it, is absolutely correct.
The insistence was on hierarchical "supervision" and a demand for examination of the finished product in advance. Where was the request for a formal proposal?