Pine's Leaders Comment
Mike Seibel and Terry Gray were asked to comment after last summer's rave review of Pine in Open Computing. As Pine project leader and one of the principal authors of Pine, Mike Seibel oversees its continual evolution. Terry Gray, C&C director of Networks and Distributed Computing and the person responsible for guiding the initial design of Pine, continues to serve as coach and cheerleader for the Pine team.
These questions and answers were supplied via email (using Pine, of course!).
Q: How do you feel about the increasingly widespread "fame" of Pine?
TG: The success of Pine is certainly gratifying, but it is also a two-edged sword. Sometimes I long for the days when there were just a few Pine users, and we could try different design ideas and change them if they didn't work. Now if our early testing does not catch a design error, we may have to live with it forever, since folks at thousands of sites quickly become used to the behavior, and then it is difficult to change anything.
On the other hand, we get messages from people all over the world thanking us for Pine and saying how much they prefer it over whatever they were using before. These messages truly encourage the whole team. When things are particularly stressful, or when we're having trouble tracking down a subtle bug, such messages are especially welcome.
Q: For whom did you originally create Pine?
TG: Pine was originally developed in 1989 for a small group of people who were happily doing email on a UW mainframe computer. The problem was that the mainframe was about to go away, so our task was to find a Unix-based mailer that was just as easy to use.
After looking at the available choices, it was clear that we would need to build something ourselves in order to meet the objective. We started with UCLA's "Ben" mailer for user-interface inspiration, and Dave Taylor's popular "Elm" program for an initial code base.
Although there was considerable skepticism at the time, I had no doubt that if we succeeded in providing a good solution for these folks, Pine would find a much larger audience at UW. Now, even without any advertising budget, Pine has become the mailer of choice for most of UW's over 40,000 Uniform Access computer users, not to mention being used at thousands of other sites in over 40 countries.
Q: Did you ever expect that Pine would become this "big"?
TG: During the early phases of the project I wondered how popular Pine might become, but I never tried to predict. It became clear after its general release that Pine was going to be a "Big Deal." Just how big still remains to be seen, but there is no sign that interest is waning. At least the requests for additional features continue unabated!
Pine's success is due primarily to its ease-of-use, but the fact that we have incorporated support for leading-edge Internet mail protocols (such as MIME and IMAP) has also helped. UW can be very proud of this project and the fact that so many people all over the world have found Pine to be a useful tool.
C&C's Terry Gray and Mike Seibel. Photo by Rick Ells.
Q: How do you contend with the different viewpoints of what Pine should be?
MS: As more people use Pine, the number of suggestions--and sometimes complaints--increases. Comments, ranging from simple status line wordings to extravagant new features, are sent in by the neophyte and old-hand alike. No matter how far afield a comment may seem, we've found it usually points out some part of Pine that could be improved. And a fair percentage do go onto our "to do" list.
Once on the list, several constraints shape a suggestion's priority. We consider how it fits with our ease-of-use design principles, its general applicability, value to our campus, and implementation cost. Of course, bugs--particularly those causing crashes or data loss--take precedence. All in all, the diversity of input has played a large role in Pine's success.
Q: Has Pine been shaped by outside forces? Do you keep Pine in sync with universal email standards?
MS: Pine is built on several underlying technologies that do affect what can be done with its interface. An example would be Pine's use of the Internet Message Access Protocol (IMAP) to access folder and message data over the Internet. We often have to balance our design goals with what IMAP can provide as well as associated network and server performance issues.
As for keeping in sync, we're lucky enough to have IMAP's inventor, Mark Crispin, as part of our development group. So, tracking at least that key standard is as easy as incorporating his latest implementation.
Q: What new features of Pine are you working on now?
MS: The version we're working on is intended to tie up many of the loose ends in version 3.91, though it will have a few new features as well. I don't want to let too many cats out of the bag, but some of the features include performance tuning, improvements to Pine's MIME support (the protocol used to send and receive attachments), and X terminal mouse support.
Q: What is your advice for someone who hasn't used Pine?
MS: Try it! An important goal from the onset was to make Pine as easy to explore as possible. As for experienced users, don't be put off by Pine's docile exterior. You might be surprised by the number of advanced features just waiting to be enabled.
[Editor's note: Other C&C staff members who have made significant contributions to the Pine project include Mark Crispin, Sheryl Erez, Steve Hubert, and David Miller. Much credit is also due to Laurence Lundblade, the "original Pine programmer," who is now at Virginia Tech.]