I had a useful conversation with Ken Thomas ’93 about the future of the Williams Conversation. The context is that Ken is building websites for the reunion classes of 1988 (mine) and 1993 (his) using Drupal. But before diving in to those details, I want to talk again about some of the aspects of the Williams Conversation and how we need to organize them.

First, we have the blogs. These include blogs by any Eph, whether a student, professor, staff, parent, alumni or local resident. We also include news sites (like greylocknews) or relevant web sites (like the Williams forum on College Confidential). Anything with a dedicated feed belongs in this category. Ideally, this would include the posts from the Ephs that participate in group blogs, e.g., Dan Blatt ’85 at Gay Patriot. We want to know about every post by dan, but we don’t care about the posts from his co-blogger. This is easy to do if there are author specific feeds at the group blog, but more difficult otherwise. We currently try to aggregate this information at Eph Planet, which is now fed off of our Eph Blogroll. At our old site, this worked fairly well because we used Planet software. (Thanks to Eric Smith for setting that up.) Things aren’t working nearly as well now because WordPress is not as cool as I thought it was. (Ask in the comments if you want more details.) But, with luck, Ken will solve this problem the “right” way.

What would the right way be? First, we need an interface for the Eph Blogroll “wrangler,” the person who would actually keep track of this. (I volunteer for this and everything else here that does not require actual coding.) That interface would just require three key items for each blog: Name, class and blog url. (Class is left blank for non-alums.) Given this, the software should look for the correct feed. If it finds it, fine. If it can’t the wrangler would need to poke around. Any other data that folks wanted to see (e.g., the blog’s title) should be available from the feed. Given this data, the site can automatically generate a blogroll and an Eph Planet-like collection of individual posts, each with a link back to the original blog.

The wrangler would also have the option of adding “tags” to a fourth (optional) field. These would be things like “finance” or “local” or “faculty” that provided additional information about the source or content of the blog. These tags (along with class information) are used to provide “views” into the large stream of all posts. For example, instead of views all posts, someone might want to just see those from her classmates or those with something to do with finance. Recall our various discussions about Eph Communities of Interest: EphCOI. Big picture: We need to first gather all possible posts of interest and then make it easy for Ephs to see either all of them or just those that they are interested in.

Second, we have news. Unlike the blogs, this sort of information will not usually come to us via an already-created-by-someone-else feed. Instead, we need to make the feed. We have talked about this as Eph News in the past, and that same wish list, slightly edited, still applies.

Eph News would present a feed of news stories (or blog posts or whatever) which mention specific topics. Duplicate items would be deleted or aggregated in some way. There should be a pretty interface for users, along with an RSS feed. Google News is a nice example. Consider a search for “Williams College.”

Although the UI for Google News is fine, the substance is not. For example, there seem to be many important sources that Google News does not (or, more likely, can not) aggregate. A current example is the New York Times. Note the articles which mention “Williams College” but which do not appear in the Google News feed. So, Eph News needs to search more than one source and then collect the resulting articles/posts together.

An interface for the maintainer which allows him (read: me) to easily maintain a list of text strings that we want stories about. Google News is fine, but it is not easy to simultaneously search for stories which mention either “Williams College” or “Morton Schapiro” or “Bethany McLean” or “James MacGregor Burns” or several dozen more strings. We also need to have simple boolean searches like “Chris Murphy” & “Congress”. We don’t want every story that mentions Chris Murphy since it is such a common name, but any story with both Chris Murphy and Congress is probably about the Eph Congressman Chris Murphy.

Although the interface is mainly just one line for each search string (boolean or otherwise), we also want an option field for a list of one or more tags. For example, the line for Bethany McLean would include “class of 1992″ as a tag. The one for “Chris Murphy” & “Congress” might have “politics” as a tag. As with blog posts, the goal is to allow for readers to easily see just the news stories that interest them. Tags will make that (relatively) easy. As always, there is much more that one can do, including users to add their own tags, but the above is good enough to get started.

Third, a decent on-line forum. We can quibble with the details about how to thread the conversations, how to moderate them, whether to require registration or forbid anonymity. I am flexible about all of these issues. But we need something good, perhaps even at the Slashdot level of technical sophistication. Why? Consider the several hundred posts at WSO (handy links from Will Slack ’11) in multiple threads on the topic of the Social Honor Code. We need a forum that brings these conversations together, that allows sub-conversations to easily form, that allows quoting and linking to individual posts. And so on. Most importantly, the forum must allow participation from all Ephs. Drupal may come with this capability built in.

Little of this, of course, is that different from what most users want from an on-line forum. But we still need someone professional to select, maintain and improve that “virtual log.” And, once again, the concept of “views” will become valuable in the second of third iteration. Not only, as at Slashdot, will we want to make higher quality comments more visible. We will also want to allow users to see just those conversations and/or comments that they are interested in. Having moderators who tag conversations as well as tags associated with different commentators will be useful. Yet all that is some distance down the road.

Is there more to be done? Of course! Once we have these three elements up and running, we will want to integrate them. How? I don’t know. Yet I am sure that the journey will be fun.

Getting back to the reunion websites, my view is that, over the next few months, we can think of three stages to this process. First, we need a website that is fairly similar to EphBlog. It should have a central area in which new information, say posts, are presented. Off to one side would be permanent pages. In the case of the class of 1988, these permanent pages or links for give details about the reunion, contact information for class officers and so on. The recent comments would also just refer to related posts, not to posts from anywhere on EphBlog. In the past, I have thought of these as “views” on to EphBlog. It should be able to go to here, for example, and see just material related to the class of 1988. As you can see, this currently uses the “category” feature of WordPress to show you just the posts tagged as “Class of 1988″ but does not change the boxes on the right. It was on our list to add this functionality, but, with luck, Ken will be building a better world for us. So, as a first pass, one could just take EphBlog’s feed (and note that you can get a separate feed for each category), plop it into the middle of a Drupal created page, add some permanent pages in the upper right and, Voila!, instant class of 1988 reunion webpage. In the first stage, these web pages could just use the appropriate feed from EphBlog. Ideally (and to make our friends in the reunion office happy) this will be available by the end of February.

Second, once we are aggregating all Eph-related blogs posts and news articles, as described above, we will be able to substitute much more class-specific into the central news feed at each site. At this point, we might have three different sites up and running: class of 1988, class of 1993, and the proto EphBlog replacement (PER). Each site would have an Eph News and Eph Planet button that provided the appropriate view into the stream of all Eph blog posts and news items. Each would feature a subset of this information in its main page, probably selected by a volunteer from each class. (Or, by default, the main page could take some nominated article from a central location.) The PER could have the main page with a direct feed from EphBlog, but then have the full stream for the Eph Planet and Eph News buttons. With luck, this is could be up, even in a rough form, by the end of March. There would also be discussion forums and book collaboration available.

Third, conceptually, there is only one Williams Conversation, one collection of all blog posts and all news items. Indeed, blog posts and news items are just “things,” much more alike than they are different, perhaps mapped as “nodes” in Drupal. The same applies to items like Record articles and class mailings, scanned pages from old Guls and pictures from Homecoming, campus podcasts and YouTube videos, College Council minutes and football play-by-play broadcasts. They are all a part of the Williams Conversation, all items to be collected and organized, linked to and discussed. If you build it, they will come.

Fourth, and no due date for this, we kill EphBlog and replace it with what Ken and his team have built, a new site whose motto is “All Things Eph,” a site that “encourages, organizes and supports the Williams Conversation.” We select a new Board (without me), and I retire to essay writing, just another blogger who contributes his thoughts to our 200 year-old conversation on the log.

Make it so.

Print  •  Email