Firefox Support in Onfolio

I was just going through all of the requests we get for enhancements to Onfolio so I could issue Beta invitations for those people who were nice enough to write to us. During this process, I was reminded of what an interesting path we took to get to the point of integrating Onfolio 2.0 into Firefox.

At the beginning of the development process for Onfolio 2.0, Firefox wasn’t really on the radar screen. Support for browsers other than internet explorer was high on the list of customer requests, but the requests were spread out over a large number of browsers, making it hard to address them without a huge amount of work. Our theory was that doing the work to fully integrate into these browsers would dilute the effort to add to the core value of Onfolio in the areas of shared collections, modeless/advanced capture, advanced searching, structured report publishing, integration with EndNote, RSS reading and so on. Then, two things happened.

The first was that we got the results back from our customer survey in June and saw that 25% of people who were using Onfolio on a daily basis were also using browsers other than Internet Explorer. This was surprising because it was a lot higher than the 5%-6% market share that alternate browsers were purported to have at the time, and it was surprising because we did not work very well with those browsers. I think that most of these people used Internet Explorer when they wanted to use Onfolio, or used the Onfolio Deskbar to capture items to Onfolio or access items stored in Onfolio. We knew these people would be a lot better off with the kind of integration Onfolio has with Internet Explorer, but we still had too many other browsers that we would need to work with to make integration cost-effective.

The next thing that happened pretty much solved that problem. I was listening to the radio on July 5^th while driving and heard NPR’s Larry Abramson telling me that I should think about using a different browser. His story was in response to an announcement by US-CERT and some follow-up newspaper articles that came out in late June recommending browsers other than Internet Explorer as a way of avoiding internet attacks based on known holes in IE. When the initial CERT announcement and newspaper articles came out, we thought that we might need to rethink our initial decision to not support Firefox. When the NPR story hit, we were sure we were going to have to support it. That marked a level of mainstream attention that most computer issues never get.

Something that the CERT follow-up stories had in common was that while they recommended “other browsers” in general, they usual tacked on the phrase “such as Mozilla’s Firefox.” This seemed to start an information cascade that changed a lot of people’s preferences from wanting support for some other alternate browser to wanting support for Firefox in particular. It also prompted a large number of people to make this request for the first time. When I look back on our customer request data, I see that in early July, requests for support for alternate browsers jump to over 1/3 of all requests for new features and that Firefox is the preferred browser at a rate of 6 to 1 for all other browsers. Finally we had a dominant alternate browser that would make the work we would have to do pay off. An article by Walt Mossberg in September reinforced the Firefox movement.

JJ put together a prototype over a couple days to show that we could embed the Onfolio collection explorer as a pane in Firefox and that we could get our buttons displayed on their toolbar. With this proof-of concept in hand, the team went at it hard for about four weeks, we moved our schedule out, and we ended up with working Firefox integration. Since Onfolio is so intricately embedded into the browser, there were a number of technical hurdles to overcome, but in the end it looks like we will have complete parity with the level of integration we have with Internet Explorer. I’ve been using an Alpha version for the last couple weeks, and I’m very happy with the way it came out. If you want to try it out yourself, you can sign up for the Beta here.

As a side note, we did make some improvements in the way we work with some other browsers, and we made a lot of improvements to the Onfolio Deskbar to improve the Onfolio experience with other browsers as well. We were not able to make any progress with Opera, however, which was the second-most requested alternate browser after Firefox.

Comments

[Joe Cheng]

Hi Sebastian, I couldn't resist weighing in with my perspective on this. I've been a Firefox user since... well, I don't know the version, but it was called Phoenix at the time. It was clear to me and many thousands of others even then that this browser was going to be big. The moment I realized that all of my IE keyboard shortcuts worked in Phoenix was the moment I realized these guys were really serious about taking on IE.

Some of us believed right from the beginning (of Onfolio 1.0 development) that eventually we were going to have to do a full Firefox integration. The problem was always the impedence mismatch between the Mozilla platform and our development platform (.NET/COM). I personally was convinced that we would have to reimplement the product, or at least all the user interface code, for the Mozilla platform--so great is the chasm between XUL and .NET.

But JJ always suspected there was a way to slam our code in. Seriously, if pigs had a C interop layer, JJ could make them fly. I know that now, but before he showed me the first Onfolio/Firefox prototype, I would have bet a lot of money against his proposed interop mechanism. And even once he proved that we could force our UI into the browser, there was a lot of grungy work to be done to make the integration seamless and support all the features we have in IE, and a high probability that we'd eventually run into at least one showstopper problem with the Firefox integration that would doom the whole feature.

Anyway, I don't want your readers to get the impression that it took the mainstream press to clue us in to the goodness of Firefox. The desire to integrate was always there, but the development cost was just too hideously huge to justify when there were so many areas of our 1.0 release that had the potential to be so much more. Luckily, when CERT and the media raised the stakes, it forced us to take a closer look at the technical challenges, and it turns out they are merely very difficult rather than completely insurmountable.