I worked for a time on the Comprehensive $\TeX$ Archive Network (CTAN). If you write $\TeX$-related software that you want to offer to the community, then you send it to CTAN where users can find it and where the standard distributions MiKTeX and TeX Live (as well as commercial vendors) can also find it. I saw there the advantage of having an authoritative single source — think of the waste if every $\LaTeX$ user had to do their own search for the latest version of the package that makes hyperlinks.
There are sites that offer lists of open math books (the American Institute of Mathematics (AIM) comes to mind but there are others such as the Open Textbook Library). These sites are great but have limitations. For instance, AIM requires that a work has been used by someone other than an author, which rules out some in-progress projects. Other sites are restricted to one subfield or another, or are simply restricted to what the site's owner has run across. I'd like to see a comprehensive archive with a mission of helping the community, including helping it grow.
- One stop shopping for users: Instructors and students alike can easily see the choices. This also includes one-stop shopping for aggregators.
- People could search by topic, level, license, etc.
- Opens the possibility of offering and collecting reviews, keeping a count of adoptions, as well as allowing upvotes and comments.
- Gives new initiatives the chance to get noticed. Established projects appear on the top of search engine results, naturally. New works could use a boost.
- Offers authoritative versions. I have a text that has been online since 1996-ish and when a person enters my information into a search engine, years-old versions are prominent.
- Links to other resources, such as mailing lists or events. For instance, the site could have been used to announce the Baltimore session on open-source texts.
As I mentioned, I have some experience in this area. The network traffic is not that great. When I ran the CTAN site there was traffic (particularly in late August), but it was not a major issue.
In addition there are now software systems well-adapted to writing such sites. I like Django (free, open-source), but there are a number of others. Such systems reduce the burden of development.
Such a site would be, I would think, a boost to the community. But in some ways the technicalities are not the hard part — the hard part is getting people to adopt the site. I wonder; are folks interested?
Saint Michael's College