Jeff,
First, I realize my previous message was terse. I am sorry for that. I should have waited until I could give it my full attention, but alas, I was heading out the door and just had to opine! =)
1. a tax-exempt nonprofit *is* at risk of Unrelated Business Income Tax (UBIT) if it engages in "substantial" for-profit activity. IRS determination is largely on a case by case basis ie. subject to irs agent interpretation. The nonprofit is responsible for UBIT.
2. An IRS auditer may not be knowledgeable about open source, thereby viewing nonprofit's behavior as UB if the nonprofit agents include commercial entities and/or their agents.
3. UBIT can be much more than for-profit corporations pay in taxes without putting out the effort to maintain tax-exempt status, because nonprofits cannot declare expenses that are not directly attributable to the UB activity.
4. Cost of fighting a UBIT determination could be expensive.
5. the statement "There are things that we (the community) want to do that would not be permissible within a non-profit but we are not going to try to do them within a non-profit." is consistent with "..a US non-profit is not suitable for [all] openacs endeavors".
6. "There are clearly quite a few things that we can do with a non-profit we could not otherwise do" is also consistent.
7. Setting up a foundation does not change the legal status of the code.
8. You write:
"it's quite telling that it is hard to find a well established open source project without an affiliated foundation".
Not really.. aolserver is opensource code managed by for-profit AOL. tcl/tk is managed by forprofit activestate (with some copyright held by a nonprofit university). I believe postgresql, inc holds the tm and manages postgres via the The PostgreSQL Global Development Group (project). Let's not forget arsdigita's ACS is now Redhat's CMS.
9. You write:
"a foundation.. *is* a corporation and can mitigate the kinds of risk you allude to with your comment about 'attack[s] by for-profit companies'"
I'm not discussing personal risk. I mean risk to the resources of the foundation not being disbursed as intended. A nonexistant foundation does not create risk. Noone wants to see others' hardwork and dontations in jeopardy. A nonprofit does not have the horizontal market flexibility to compete against something like WebCT without risking UBIT activity. Also, UBIT investigations can be triggered for instance, when an agent from a for-profit company contacts the IRS about a potential UB activity by the foundation..
Consider also that developers who find themselves working under the foundation's umbrella might be restricted from contracting work from commercial entities because of the potential IRS perception that the agent's work might be UB.
Why not leave the openacs project as an informal collaborative venture (of multiple targets), and create a foundation that becomes a collaborative entity among many others?