Read the complete article in eWeek:
Some of the more interesting quotes from the article (inline commentary by Jon Stahl, from The Plone Blog):
Open-source didn't have a starring role at UNC until Plone came along:
"We really never had a large open-source solution before. UNC-Chapel Hill widely uses proprietary software solutions. Education has been slower to adopt open source," Hitlin said. Therefore, open source was not a clear path two years ago when the OIS began evaluating its options, he said.
Plone blew away both commercial and open-source competition:
[UNC] considered both commercial and open-source content management software. The committee evaluated Vignette's Vignette and Oracle's Oracle Portal, which was attractive since the SOM [School of Medicine] already used an Oracle database as its back end....
Plone could be adapted to fit UNC's specific business requirements:
All first- and second-year students attend the same courses at the same time. Each course typically is taught by a number of different professors and clinicians, rather than one instructor. According to Hitlin, if customization was going to be a large part of the project, the CMS committee reasoned, why not go with an open-source version so that, at least, the code would be free? That way, the SOM could devote its limited resources to development and implementation.
Plone's active local user communities were a huge win:
The Chapel Hill area has an active Plone user group, which was an advantage. Since the SOM was in the process of building up its developer staff after a few cutbacks, the committee believed it would not be difficult to find people with Plone skills in the area.
Some problems with a highly proprietary software/hardware combination cropped up:
[T]he project team got an unpleasant surprise when it installed the application on one of the SOM's Sun Microsystems' Solaris SPARC boxes.
"As soon as we installed the application locally, we saw significant latency — 10 seconds to load a page, where it had been 1 second in the development environment. It turned out the SPARC hardware doesn't run Python efficiently," Hitlin said.
Cignex advised the SOM to run the open Plone application on inexpensive Linux boxes, which would run a reasonable $2,000 to $3,000 each. But, according to Hitlin, the SOM's system support group was accustomed to supporting the Sun Solaris architecture and was reluctant to add a new platform to the mix.
But in the end, Plone's usability saved the day:
During June and July of 2005, the team dealt with last-minute bug fixes and user training, which turned out to be trivial, according to Hitlin and Thangavelu. The faculty members intuitively understood the user interface and took charge of updating their course materials with ease.