I've long had an interest in the idea of a community built around a particular machine or infrastructure. Not just resources hosted on a particular machine, but resources established in a particular way. SDF and Nyx are good examples of public use UNIX machines, and most universities have similar systems established for students to use, even if e-mail and some other aspects of the computing environment have moved away from the UNIX shell.
I think part of my interest is I've long wanted to witness a community with people who have an interest in a particular type of system and the technical ability to connect to it and use it to do various tasks. I've also long wanted to run a complete Windows infrastructure. I previously had a planning document referring to the idea as the PAWC, or Public Access Windows Computer. Under that plan, TECT was going to be a virtualization server, ultimately running different virtual machines for remote desktop, SharePoint, Exchange, Lync, etc.
The technical ability thing is important, I think. Web forums are interesting, but at this point they are the lowest common denominator of communication on the Internet. Anybody can sign up for them, usually with really low effort, and then later question how things like posting image attachments work. I also suspect part of it is when somebody connects to a resource using a more complicated method (much like Hotline back in the day, and to a certain extent IRC even today) they are perceived as an insider to a certain type of knowledge. I'm typically not in favor of excluding people, but I'll admit there are times when I'm pleased that there are communities that place more technical demands on users.
The other thing for me is it forces me to put effort into setting up a system correctly or as close to correctly as possible and maintain it for other peoples' use. TECT is getting close to this status. I have a few users of it, not only of the web site data, but I have two or three e-mail users, and for that reason, I am really consistent with the backups on the volume that houses TECT's system, e-mail and SharePoint data.
Of course, the fact that there are users (myself and others) is much of what makes configuration changes to TECT so complicated. It's one of the factors behind my present stalling with TECT.
One other aspect of running such a system is that getting any more use out of it than my "users" already do really does require a lot of network throughput. One of the things I'd ultimately like to add to the setup is a terminal server and official capacity for people to use VPN to access resources or get to file shares, and that requires a lot of network throughput.
The network throughput and need for low latency involved in using a remote graphical environment is, I imagine, one of the reasons why "public access" systems tend to be based on UNIX most often. If I had TECT running FreeBSD or Linux (or another machine, such as nachibes, which I wanted to open to the world) the fact that I'm on such a slow DSL connection would matter less, because people using it remotely would be less sensitive to the latency of the connection, and would be using less throughput in total. I bet my line could handily support ten to twenty or so active remote users if I wasn't doing something else on the line such as streaming music.
Although, and this is where the post was going the whole time, I simply don't know if there is room for any more "community machines" (like SDF) at this point. Everybody who wants one can reasonably afford their own computer on which to run FreeBSD or Solaris, and everybody who wants access to a lot of remote storage can find something faster than copying files to a remote server stored in somebody's house.
The other challenge with a Windows based system is that the "free" licenses that are sometimes available (such as for students via Dreamspark) are not licensed for other people to use them, even in a non-commercial environment. I had previously waxed philosophic about buying a single Microsoft Multipoint machine [PDF] and buying licenses for that and Small Business Server 2011 Standard as users showed up.
I suppose in the end it's a form of scene heroism and isn't necessarily something I want to do completely as a public service, even if it could be used as a public service, by owners of Windows RT devices and Macs. (In fact, I like the idea of having a volunteer or commercial service providing remote access to a Windows desktop for legacy applications for just such a situation, but that kind of thing tends to get shot down by Microsoft.) If such a thing were technically possible for me to do (TECT had enough resources and I had a fast enough network connection to support it) then it would remain prohibitively expensive to bring it out of the "proof of concept" stage.