On 10/31/2016 3:54 PM, juar...@gmail.com wrote: > > Discontinuing support for 10% of users sounds like shrinking 10% of > customers, lay off 10% of employees, reduce 10% of funds for investments.
I can tell you that the evidence we have does not support the notion that end of life (or the approach we are proposing) will actually result in the attrition of those users. We examined the impact of Chrome's end of life on Windows XP users. The majority of users planned to stick with Chrome even without security updates. We also saw almost zero evidence of Chrome's end of life causing an uptick in Firefox usage or downloads among XP users. - Someone has the statistics details of this "10% user base" for supporting > this decision? What service pack? Where they are? What are the demographics > numbers? How often the browse web? We did look through the data. Yes, there is a geographic skew in XP usage towards countries like Russia and China. In addition, on average, XP users search less and have lower engagement rates than non-XP users, but the difference isn't massive. Peter On Tue, Nov 1, 2016 at 6:48 AM, Aaron Klotz <akl...@mozilla.com> wrote: > Disclaimer: I am not a decision maker on this, these are my personal > opinions, etc, etc > > On 10/31/2016 3:54 PM, juar...@gmail.com wrote: > >> >> Discontinuing support for 10% of users sounds like shrinking 10% of >> customers, lay off 10% of employees, reduce 10% of funds for investments. >> >> - Is really necessary to abandon all XP users? >> > Shifting XP users to ESR is different from "abandonment" FWIW, but IMHO > this move is necessary. As I pointed out earlier in this discussion, the > problems have become more complicated than simply disabling certain pieces > of code when XP does not support them. > > - Is possible to discontinue the most hard to support components? >> Somenthing like "Video conferencing will not work in XP" or like is planned >> for flash. >> > Again, that is not the problem. The problem is more like, "Sorry 90% of > users, we can't give you a better sandbox because of the 10% of users > running an obsolete and unsupported OS," or "This feature is going to be > delayed a release because it mysteriously fails on Windows XP." Meanwhile, > our competitors *do* deliver that better sandbox or *do* release that new > feature before we do. Now we're preserving that 10% of users at the expense > of the other 90%, and it's in the latter category where the growth will be. > That sounds like a pretty lousy growth strategy to me. > > - Is possible restrict the user base affected? Like only XP SP2 and >> older... >> > Existing Firefox system requirements are for XP SP2, so we already > restrict older revisions, but that isn't really the issue here. The issue > is the gulf between all XP releases and newer versions. > > As somebody who has first-hand experience with this, let me assure you: > Debugging XP-specific problems has become very unpleasant. Most developers > don't just have an XP machine sitting around to work with. We can request a > loaner from Release Engineering and debug it through there, but that is > very tedious and time consuming. Turnaround on try builds for Windows XP is > sometimes terribly slow. As XP continues to die off, this will only get > worse, not better. > > I see how Mozilla is important for open web and how firefox user base is >> shrinking. This worries me. >> > Do not confuse shrinking market share with shrinking user base. That is > only the case when the total number of users on the web remains constant, > which is not the case. Having said that, I don't want to see shrinking > market share either. > > I do not believe that we can offer the highest quality experience to the > vast majority of our users by continuing to expend resources on the past. > One of our top-line goals for 2016 has been to build our core strength. I > don't know how we're supposed to do that by intentionally tying one hand > behind our back to support Windows XP. > > Supporting XP might curb short-term market share losses but it will hinder > our ability to deliver long-term market share gains. > > Maybe hiring one or two developers for supporting this user base is >> cheaper than loosing these users. >> > > Hopefully my other remarks in this post have made it clear that XP support > is not an issue of headcount. > > _______________________________________________ > dev-platform mailing list > dev-platform@lists.mozilla.org > https://lists.mozilla.org/listinfo/dev-platform > _______________________________________________ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform
