In organizations of all types around the world, executive decisions are met with objections from subordinates. For open source software projects, those objections often take on the tone of philosophical treatises. Such has been the case the past two weeks with the Mozilla organization’s decision (which had been in the works for several months) to reduce the amount of data that its Firefox browser presents the user about its version number to near-zero.
The decision had been made as part of the group’s transition last year to an expedited release system, with the goal of finalizing new major or feature releases of Firefox every six weeks, following a staggered 24-week development cycle. At that rate, we could be at Firefox 10 by the beginning of the year. Does that fact really matter? Perhaps not, say Mozilla contributors such as community manager and Firefox product manager Asa Dotzler, who supports discontinuing the hype over version numbers as any big deal that real-world users care about.
But among those who do care about version numbers are the large, and hopefully still growing, number of developers of Firefox add-ons, the free or low-price functionality enhancers that, for the last five years, have truly differentiated this browser from its competition. For add-ons to work properly, they have to be validated against the browser versions that users are most often downloading – typically the current ones. If those current versions literally update themselves every six weeks, then users may come to discover the add-ons they installed yesterday failing to work today.
And if they can’t see the change in the Firefox version number, they might not know why.
“We are moving to a release model that makes version numbers meaningless and unactionable for consumers,” Dotzler wrote members of the Firefox development community in a newsgroup message August 15. “In that model, version numbers go away.”
Since that rather decisive and unambiguous post, Dotzler has been in full damage control mode, particularly with contributors who feel the number-erasing decision was made before their opinions could be fully heard and reasoned with. Within a few hours, contributor Tyler Downer wrote, “Frankly Asa, I’ve been with Mozilla for 3+ years, I’ve touched tens of thousands of bugs, [and] I have to say your attitude, is driving me out the door… By not having the common courtesy to explain your decisions or ideas to me and the other members of the community, Asa and Mozilla in general [are] showing how it truly feels about the community that is behind it. Mozilla is better than the community, and does not have to explain anything to them… Thank you for showing to me that Mozilla is not the community that I need to be a part of.”
This was not (by far) Downer’s final say on the subject. But while other contributors raised objections that were less personal and more directed toward the product, Dotzler’s responses for the next several days tended to be directed toward Downer and his allegations about Mozilla’s attitude.
One such product-directed objection came from a contributor who had noted Mozilla’s success in the past with releasing security updates on an independent cycle, and who believed merging the security update track with the feature release track may compromise the browser’s integrity. “There has been a distinction between ‘security’ updates and other forms of updates,” contributor Eliot wrote. “Having that distinction provided a confidence to users that they lacked with Microsoft when, for instance, they rolled out Microsoft Genuine Advantage, and other functional changes through an update that then caused users problems.”
Contributor Matt Brubeck cited his experience as a developer for Amazon.com, saying that in the context of Amazon’s Web site, it might not make any difference at all to Amazon’s customers if it had published a Web site version number on the site itself. Hypothetically, Brubeck contended, users don’t really care about version numbers, but added, “We’re not in my hypothetical world. Here in the real world, many users do expect client-side software to have version numbers, and to find the number in a particular place. By proposing a change like this, we may have the opposite of the intended effect: We are focusing more attention on the version number, and making the transition to rapid release less seamless.”
“I’m rather unconcerned about the actual change,” wrote contributor Robert Kaiser, “but I’m very concerned about the public message and the loss or gain of peers, as both are what made us a large contender on a now-fast-moving market and if those turn against us, we have lost and the web has lost the only really open voice. We cannot let that happen.”
More objections were made by Linux supporters who noted that general users in a Linux environment do not have permission to update their applications, or to have their apps update themselves, for reasons of administrator permissions and standard user security. As a result, Linux-based Firefox users may be assured, they argued, of almost never having the current version – and that could cause new security headaches if vulnerability patches and updates continue to be rolled into the single rapid release cycle.
There was also this: “I am a Firefox engineer, and when people tell me about a problem, I want to know which version they are using. That’s always my first question. This won’t change.” This from developer Kai Engert, who has crafted his own temporary solution to the dilemma: a Firefox add-on that simply reports the browser’s version number in the status bar. (Sadly, this add-on too may be subject to behavior trouble on account of Firefox’s rapid release.)
Users who tend to report trouble to engineers like Engert directly work in enterprises, and that led Mozilla’s Dotzler to reiterate a response that got him in hot water the first time around: Firefox is not centered around use in the enterprise. Wrote Dotzler:
Firefox continues to be focused on individual users and not enterprise. If we decide to change that, with changes to how and what we deliver as Firefox, as a result of the discussions in the Enterprise Working Group, then we’ll do so but that is a separate discussion. For the time being, Firefox is designed for individuals and not enterprises and I don’t think the product or UX team should be focusing on developing an enterprise-friendly feature set until our mandate to focus on individual users has been changed.
But much of the rest of Dotzler’s responses have taken this tack: “Tyler, you’re painting a picture that’s just not accurate… The only fair way to measure the success of our rapid release migration story is to look at the migration from 4 to 5 and starting today from 5 to 6… You also lament that we won’t have a better system until Firefox 9 or 10 (which is not completely accurate)… and you seem to miss that the version number removal is a proposal that wouldn’t see the light of day until just such 8 or 9 releases.” So presumably November.
The Troubleshooting page in current versions of Firefox (Help -> Troubleshooting Information) does show the version number, and may continue to do so indefinitely. That’s only slightly removed from the “About” box (Help -> About Firefox), but user experience engineers have argued that the “About” box is where users of any application expect to see its version number. In response to Engert’s contention that users in trouble may not know be able to intuit this relocation, Dotzler responded, “I have been in close contact with the people who do more Firefox user support than any other individuals in the world and they believe that About->Troubleshooting is far superior for diagnosing Firefox problems than About->Support so this isn’t just my opinion. It’s coming from those people who understand this process better than me and you and just about anyone else on this thread so far.”
It was this attitude that tipped the balance for one contributor named Steve:
The summary dismissal of the enterprise community is relegating a once treasured and useful tool of the enterprise developers to the same category as Google (making Firefox second fiddle to Chrome),” he wrote August 16. “It has also left some of the most influential members of the community uninterested in continuing to champion and support the seemingly ‘legacy’ browser that Mozilla is fast becoming. This calls into question the wisdom of leadership that considers Grandma Jane running Windows Vista 32-bit and surfing the web to take care of her cat a more important user than CEO Joan who uses the same plug-ins, same browser, same WWW to make decisions about tens or thousands of lives, their well-being, and their prosperity.