11/30/2004

Princeton Portal Project

On September 12, 2002, a bunch of Princeton computer geeks met in room 412 Brown Hall to discuss the future of "Sleep", the popular network search system run by Finn Calabro '02. There were fears that since Finn had recently graduated the new proprieters at Quadrangle Club were going to charge for the service. The people in the room consisted of myself, Matt Stack '04, the infamous Dan Peng '05, Niraj Bhatt '03 (my former MAA and 2-year hallmate), Joe Barillari '04 (who was slashdotted for his paper on the RIAA suit against Peng) and a few other computer geeks. Of course I was present to represent the "Mac people" since I was the Apple Campus Rep and president of the Mac user group at the time.





The discussion turned to other ideas for campus web services. I brought up the idea of creating a campus portal with student-specific information. The first couple of ideas I had were to display the events going on at "the street" that night and student-submitted campus jokes (although that seemed to brush too close to the Nass's territory, although in my opinion Verbatim is no longer as funny as it used to be).





We all left the meeting with projects to work on (the most pressing of which seemed to be a lobby to convince OIT to give all users PHP on university-hosted web sites, which Barillari spearheaded but it ultimately never happened and for good reason), the portal project being a long term goal that seemed like we would all collaborate on. Yet within a week Stack, who didn't have a specific responsibility since two other search sites were being started by others (Gank and Wake), registered the find.princeton.edu dns name with the Princeton hostmaster and began work on a site which he called the "Princeton Portal Project." The site first simply had links to the three search services. Stack put up flyers around campus with the text "Search the Princeton Network/find.princeton.edu".





Stack's site quickly became popular when he added street information, which showed what parties are going on and where (my idea). Other features included an advanced directory search which allows more advanced searching of the Princeton people directory than the university had provided on their site (although the Princeton site now offers most of these features) and a calendar of student events information. Stack requested help from a number of people, including myself, on November 1, 2002, showing that he wanted this to be somewhat of a community effort. In my own laziness I didn't contribute too much.





In March I submitted an event to the site directly to Matt via email (as opposed to using the web form). I was about to submit the event, a non-alcoholic Colonial Club party, via the web form when I noticed a credits page on the site. The page had a description of the site and the process by which Stack created it. He had a picture of himself in front of all his computers in his room and at the bottom of the page he listed some people who had helped him, all of which seemed to have very minor roles. So I emailed Stack with the event:




On Wed, 2003-03-26 at 22:59, Daniel P. Semaya wrote:

Dude. Didn't I give you all the big ideas for the Portal site? You could at least show me some appreciation on the thank you page.



Also, add that tomorrow night at colonial that there's DJ Bob and FOOD from 1-3 AM.



Dan







Stack's response ten minutes later was:




On Wednesday, March 26, 2003, at 11:19 PM, Matthew Stack wrote:



HA HA HA HA HA



Oh shit :) Of course ....



Thanks,





-stack







After that I got a small credit line on the bottom of the page that mentioned something to the effect of "general idea and concept" if I remember correctly. That satisfied me since I was too lazy to do any work at the time.





Fast forward to the following fall, senior year, Stack and I are in AST 205 together and did homework together on Wednesdays after lunch at Colonial. At one point we discussed the idea of a campus calendar. I had worked on a project over the summer for the Last Week web site which syndicated the band's tour calendar in iCal format. It allowed users to do such nifty things as syncing the tour calendar with their PDAs and emailing and modifying tour dates for their own personal use. Stack again organized a meeting with myself, and a couple of USG people who were working to create an official online calendar for the school. Stack didn't show up, so the meeting never took place. The university apparently went on to purchase some calendar software. I was told that they didn't like the idea of Stack's web site showing street events and wanted to be the official controller of the public calendar for the entire school.





I later brought the idea up to Matt about doing a spring independent work project with him to create a "real portal" where users would log in with their netIDs to get personalized information. It would show the user's class schedule, and their favorite social events. weather, customized news feeds, a discussion board, etc. I explained how the only way a site like this could succeed is if it was separate from the school. We had seen how the most popular calendar that students cared about was the street calendar and I had just the idea on how to create such an independent system. Stack was concerned that only the university had (and still maintains) the server power to run such a system and to maintain it indefinitely (after we graduated), however P-SMUG, the student Mac user group, which I ran, had an iMac running OS X 10.2 as a web server, with PHP, MySQL, and all the other tools necessary to host the site. The server sat, and still resides, in the third floor networking closet of Colonial Club, and it will remain there indefinitely, as the host database subscription is generously paid for by OIT Student Computing Services. My idea was that P-SMUG and PUG (the student unix group, which was mainly an email list at the time) could run and operate the site, therefore maintaining the interests of the students, which were largely different than the interests of the university.





My ultimate goal for the site was to charge nominal fees to campus groups for advertising, since the way most groups would advertise would be with 8.5"x11" flyers posted on lamp posts around campus. Creating and posting these flyers was a long time consuming process (I know this all too well from personal experience) and Joe Barillari even created a guide to creating such flyers. Yet the flyers would soon be covered up by other flyers and students would often ignore them. I thought that if advertising on the portal site was cheap and effective enough, student groups would turn to it over other advertising forms. Some student groups even went to great lengths to have large color posters created for their events. Portal advertising could have been faster, cheaper, and more effective. Also, groups such the Alcohol Initiative (or whatever group posts the large anti-alcohol flyers around campus), which have large budgets, could run public service ads (as an example of this in the real world, the Ad Council has adopted internet advertising as a major avenue for public service announcements). I thought that through the revenue the portal project could eventually spin off from P-SMUG and PUG and be a student agency, employing students to do the work necessary to maintain and upgrade the site for the future. I never intended for the site to be a profitable business, but instead would be required to break even in order to pay its employees and in the future move up to a dedicated web server (hosted in OIT's server room at 87 Prospect). In contrast some student agencies, such as the more recent incarnation of the Tiger Computer Agency, were clearly businesses run by students out to make a buck (or many bucks, but more on that in another blog entry).





Unfortunately none of this ever happened. I wrote an email about my ideas to the group, which included Ryan Walsh '06, who responded to my email, largely agreeing with my ideas. However Stack and I never did the project and Find remained largely the same for the year. I discovered a few weeks ago that the site was not dead. It was being run by Ryan, who picked up the find dns name. It appeared that Stack handed over the site's code to Walsh. The site now (apparently no longer up) had customization features that required a netID login, many of which are features that I had brought up with Matt months (and years) earlier. On the site's info page Walsh has a description of his motivations for creating the site and thanks Stack for pioneering the site and passing it on to him. The site is not complete so I will refrain from any criticisms at this point.





Enter the USG's new site "The Point." In today's Prince there is an article about the site, which is a slick-looking school-sponsored portal site for students, created by Clay Bavor '05, partly for his COS 333 project last year. The article states that the USG paid Stack $3000 for Find. It seems like that purchase didn't really do much since it appears that none of the code was used. The school payed Clay the same amount for Point, plus some more for a dedicated server (which apparently isn't up right now since the site is running off Clay's Mac in his room).





So two people each got paid $3000 for ideas that I had first, but failed to capitalize on. So why am I blogging this? I certainly have no one to blame for this but myself but I thought I would at least tell the world (or at least the 4 people who read my blog) about the role that I played in the creation of Find/Point/etc. No one really "stole" my ideas, since they weren't all that original. But I certainly had the ideas and the knowhow to create a great Princeton portal years ago and didn't do it.





This isn't the first time that I've done (or not done) something like this. Last summer I had the idea for a social networking site based around Princeton. I got the idea while planning personalization features for the Colonial Club member web site. Six months later Mark Zuckerberg at Harvard rolled out such a site as the thefacebook.com. This site has become phenomenally successful at over 100 schools. Yet another example of my inability to take an idea and do something with it. I often have ideas like this, talk about them, write about them, but never CREATE them. Now that I'm out in the real world I think it's about time that I do some creation. And maybe I'll be a bit more secretive with my ideas until I do so.... just in case ;-)





Also, I want to clear up a few things mentioned in the article. The last line of the article is:





"I had a seven-computer cluster in my room to run all the sites I was managing, but I didn't receive any University resources," Stack explained. "I didn't care about my personal funding. I cared about resources."





Stack's "seven-computer cluster" was a bit of a joke. He was abusing his (and his roommates') dormnet subscriptions by hosting sites using each of their host database entries and masquerading multiple computers behind each subscription. He was running all of these sites from his room in his dorm, on a 10Mbit network, all behind hubs and routers. That is no way to run a server room (I'm sounding like an IT manager here). It was clear that he didn't really plan ahead with these projects because he would have had very little problem getting the USG to give him one dedicated server in OIT's server room (all he would need is one). Instead he had a room full of old frankenstein PCs running various Linux distros. In contrast I got my student organization a dedicated computer, with a host databse subscription ($11/month) and found a proper location to host it in. I'm not criticizing Stack's work on those projects, since he did do some good work, but the IT guy in me has to point out that he didn't demonstrate a proper understanding of how to host sites.





Considering the big budget-ness of the Point, you'd think that they could afford an SSL certificate for the site. I do give them credit for actually using SSL, but self-signing (not paying for a certificate) throws up a warning in most web browsers and makes your site look pretty ghetto. Decent certificates cost $100/year. I think that the USG could afford this and the people running the site right now should not even have considered self-signing for such a high-profile site. I won't even go into the fact that everyone on campus is sending their passwords directly into Clay's computer in his dorm room.





I'm sure that there's more I could find fault with but considering that I didn't even have the brains to see any of my ideas to completion I should probably shut up now.