Interview with David Harris: I've never been a businessman
What the first mail sent with the program was about, no one knows anymore. But in December 31 years ago David Harris ran early tests with his application - then named ComNet Mail. A few weeks later it was ready to use. To give some perspective: 1989 was the year Tim Berners Lee started working on the World Wide Web. None of the digital amenities and nuisances of todays life existed for people outside a very small circle of researchers and students.
Thus David Harris can be called quite a pioneer and his decision to publish a free mail program life-changing - for a growing user base after 1990, but also for him. We took the anniversary as an opportunity to talk to him. We found out what happened in 2007 - when the end of Pegasus Mail was announced(öffnet im neuen Fenster) . David Harris keeps on working though and he hopes, that it evolves into a program that serves people relying on handling large amounts of e-mail efficiently and quickly, with a strong emphasis on privacy, encryption and data security. The following is the unaltered script. A german translation can be found here(öffnet im neuen Fenster) .
Golem.de: How did Pegasus Mail come to life? What was the reason for you to start working on it and how did the initial development happen?
David Harris: I was responsible for the installation of the first Novell NetWare network at the University of Otago, where I was working in 1989; we fitted out a laboratory for use by the Commerce Faculty - quite state-of-the-art at the time, but it wasn't until after I'd finished the installation that I realized that NetWare didn't come with a mail client. I was pretty surprised by that - I was used to systems like VMS and Unix, where mail was just part of the landscape, so finding that the industry's premiere PC Network didn't have mail was completely unexpected. The options that were available were quite expensive, and we'd pretty much used up our budget, so I went home one weekend and wrote a mail system.
That possibly sounds boastful, but it's really how it happened. I'd been a developer for a long time and had good DOS-based programming tools, so it didn't take much effort to create a basic mailer. I put it in the labs, and after a couple of weeks, wondered why so much disk space was in use - it turned out the students had discovered it and were using it more than they were using their classwork programs.
From there, I put it up on a friend's site in Hawaii (at the College of Business Administration) and was surprised to find that it got heavily downloaded almost at once. From there, it just began cascading: I never expected anything like the popularity it had, but I suppose that's often how things work out - seized by the law of unintended consequences.
Golem.de: What were your prior projects that made you confident that programming an own client would be possible?
Harris: I'd been a freelance programmer before I started work at the university, developing custom database systems for clients like auction houses. Dull stuff, but it gave me a good understanding of how users can break even the most meticulously-crafted program. It also taught me the value of simplicity, and gave me a really good set of tools I could use to build new programs. I still place a huge value on software tools - I think I actually love writing good tools more than I do writing the applications that use them.
Through good times and bad
Golem.de: When did Pegasus Mail grow into a full time business for you and how did you feel about that change?
Harris: That would have been about 1992, I think. I was putting so much time into it that I simply didn't have a life any more, so I had to make a choice between it and my job at the university. The decision was made easier when Novell bought a license for the program - for a while, it was included in the Red Box until they came out with their own solution.
As for how comfortable I was with it... Well, it was definitely a risk. The software was free even back then, and I couldn't see any way I'd be able to continue making a secure living, but fortunately enough people bought manuals that I did reasonably well for a time. Would I take the same risk again now? Hard to say - it would depend on how much I believed in the program, how much benefit I thought it might offer to the broader community.
You see, it's never been about money, although money has always been the biggest problem. But if I hadn't believed that the programs I wrote were solving problems for people who had real needs, I would never have started out in the first place.
Golem.de: What were the high times for the client and when and why did usage decline?
Harris: Things were great until around 1996, then Microsoft introduced this thing called "Outlook Express". Frankly, it was a poor mail program in a great many ways, but once something is in the OS, it becomes nearly impossible to compete with it. Once Outlook was there, my position was always going to become a niche spot. A useful niche spot, to be sure, but a far cry from the early days when I had many millions of users.
Was I bitter about that? A little, yes. It was annoying to have taken the plunge, to have provided something for the public good, only to have a multinational corporate (and at that time a highly reviled multinational corporate) come along and put you into difficulty through what seemed like a very unfair distribution model. But business is business, I guess, and I've never been a businessman in any recognizable sense, so I suppose I had no real right to complain.
Sure did leave a sour taste, though.
Golem.de: How was contact with and from other developers and clients - if there was any?
Harris: Less than you might think, in fact. I occasionally had contact from other authors - for instance, Steve Dorner, of Eudora fame, reached out a few times: I liked him - he was sensible, pleasant and reasonable. Unfortunately, living in New Zealand and not being especially rich meant that I couldn't regularly make it to developer events or trade shows, so I was always a bit remote, a bit cut off from where things were actually happening, a bit outside the loop. I'd have loved to have been more involved with things like the The Internet Engineering Task Force (IETF)(öffnet im neuen Fenster) , too, but the same sorts of problems applied there as well - too far away, too few resources.
Golem.de: When and why did you decide to cease development?
Harris: I need to be very clear about this - I never did! It's something I talked about back in 2007, but I never actually stopped. I think a number of comments I made got a bit taken out of context at the time.
2007 was a horrible year in a great many ways. I was under real pressure, my morale was low, I wasn't enjoying work, and times were financially very hard. It would have been easy to toss it in, but there was always a hard core of supportive people urging me on, and I would always have felt bad about abandoning my user community. I'm not quite sure how, but I pushed my way through it, and since then I've managed to recover most of my interest in the work.
Once I became reconciled to the fact that I was going to have to be very careful financially, and started living more simply so I could keep within my means, things became a bit easier... But it's definitely been a hard transformation.
Golem.de: Did you consider growing the business to "keep up with the big boys"?
Harris: Never once. As I remarked above, I'm not a businessman, and I'm also not a very good team player. I had put so much of my life into the programs that I think I was almost physically incapable of giving up any amount of control over them. Part of me revels in being different, in not conforming - and to grow the business, I would have had to become much more conventional. It was never going to happen.
Of course, lack of funds means that things like new hardware and software, promotion and marketing - they all become difficult. I also have no way of employing anyone, so whatever has to be done has to be done by me on my own, with as much help as my very dedicated and loyal volunteer beta test teams can offer... But those were the choices I made, and for both better and worse, I will live with them.
A few people offer regular ongoing financial support - without their help, things would probably have become impossible by now, so they have my unending gratitude.
Golem.de: How did things go from there? There has been a steady flow of releases since 2007 and version 5 ist on the horizon for a while now - when will it come and what will it be like?
Harris: Back in 2011, I was faced with a really hard decision: I was dealing with a codebase that had been largely written in the 1990s and was, frankly, totally out-of-date. Maintaining it was becoming a real problem, and trying to keep up with the buzz-words and trendy interfaces was nearly impossible. I had to choose between abandoning all the code and starting again, or completely overhauling the existing code to make it more "modern" and easier to maintain.
Unfortunately, I chose to overhaul - it was probably the wrong decision. I've largely ended up rewriting the bulk of both programs in any event, but without the advantages I would have had if I had started afresh. It never even crossed my mind that I might end up having to rewrite nearly a million lines of core and support code, nor that I might have to learn so many new things - something I'm really not very good at doing: I work very well with my own code, but much less well with code written by other people. And it's never-ending: at the moment, I'm having to deal with absurdities like OAUTH2, all the while still trying to wrench two major commercial-sized programs into a form that has long-term maintainability.
In some ways, version 5 has become a bit embarrassing - a bit like Ventura Publisher in the bad old days: I'm sure a lot of people think it's "vaporware", and it's frustrating that the only person who knows how much work and effort has already gone into it is me: I really wish there was something I could show the world to prove how committed I am to this transition and how much I've already done, but the reality is that it's such a big project with so many interconnected parts that it has passed beyond the ability of one person to deliver in a timely fashion. That's not to say that it won't happen, but it's taking far longer than I would like it to.
I'm confident the result will be good and that it will offer valuable service to a wide range of people, although I also have no doubt that it will still be a niche product - a product for people who are focused on handling large amounts of e-mail efficiently and quickly, with a strong emphasis on privacy, encryption and data security.
The future of Pegasus Mail
Golem.de: What is your personal view of the whole project - after all your software changed the lives of millons of people all over the world in giving them a means to communications.
Harris: Again, I find this comment surprising, because development has never ceased, and I have never shut down. The program has been available for download on a completely uninterrupted basis from February 1990, and there has never been a day since that time when development has officially ceased, or when anyone who wished to do so could not have downloaded the software.
In the early days of the program, I used to say that freedom of speech is worthless if nobody can hear you - that's what the programs have always been about. It may be that those priorities are less important now that there are so many more options available, but in the the modern world, the struggle is morphing into one of data security - providing a service where corporations aren't mining every message you send and receive to work out what they can sell you or advertise to you: a service where you can't have your mail secretly inspected by foreign government agencies; one where your data remains *your* data. There will always be people who really value privacy and integrity, and as long as there's a need to provide them with solutions, I'll always attempt to fill it as best I can.
Golem.de: As we´re a publication exclusively targeting a german speaking audience I have to ask: why was the client available in english and german of all languages?
Harris: One of my beta testers is an extremely talented man called Sven Henze: he puts in a huge amount of effort to provide a regularly-updated high-quality German interface to the program, for many of the same reasons that I provide the program in the first place. I have other translators who put in hard work as well, but Sven... Well, I really don't understand how he manages to devote so much time and do it so well.
Golem.de: What are your future plans with Pegasus Mail and how do you want it to be used and remembered?
Harris: My main plan is to get v5 out the door and show people that there really still is a place for a high-quality, independently-written mail program, that there are ways of doing things that aren't the Microsoft Way or the Google Way - that there can still be some variety in the world. And as long as people want to use it, I plan to carry on producing and supporting it. After all, what else could I possibly do?