Windows XP and Vista Long Term Support Plan

[Peter Dolanjski]

>
> On 13/10/16 18:52, Justin Dolske wrote:
> > I'd suggest placing bounds on it, so we can all be on the same page and
> > comfortable with what's being committed to. E.G. "We project that at
> > current rates, XP usage is expected to drop to X% by the normal end of
> > ESR52. If, at the normal end of ESR52, usage remains above X+e%, we will
> > further extend the life of ESR52 for an additional Z months. After that
> > Z months, ESR52 will be discontinued irrespective of usage."


We can do that if it is useful, beyond a commitment for a final end of life
date.

On 10/13/2016 1:37 PM, L. David Baron wrote:
>
>> So if we're actually talking about possibly maintaining ESR52 for
>> three years, I think we'd be better off not doing this, and
>> continuing to support XP and Vista in our non-ESR releases.
>>
>> If we're willing to make a firmer commitment to ending support, I
>> think the plan becomes more reasonable.
>>
>> Then we need to make that commitment.
>

Based on an upper and lower bound projection, the end of ESR52 (likely
April-ish 2018) shows that WinXP should be between 2.5-4.2% our Firefox
Average Monthly ADIs.  By January 1st, 2019, the range is 0.8-2.6% of
monthly ADIs (though that's trending pretty far out).
I think we can say with some certainty that the worst case scenario (in
terms of length of support) is the end of 2018.

Peter

On Thu, Oct 13, 2016 at 4:12 PM, Aaron Klotz <aklotz at mozilla.com> wrote:

>
> On 10/13/2016 1:37 PM, L. David Baron wrote:
>
>> So if we're actually talking about possibly maintaining ESR52 for
>> three years, I think we'd be better off not doing this, and
>> continuing to support XP and Vista in our non-ESR releases.
>>
>> If we're willing to make a firmer commitment to ending support, I
>> think the plan becomes more reasonable.
>>
>> Then we need to make that commitment. Windows XP is no longer a minor
> inconvenience such that we can just hack around its missing APIs via
> dynamic linking. It is different enough from modern Windows that XP has now
> become an impediment to shipping new features such as a11y on e10s and the
> newest sandboxing mitigations. Continuing support for XP on non-ESR is an
> option that completely opposes the objective of building core strength.
>
> _______________________________________________
> firefox-dev mailing list
> firefox-dev at mozilla.org
> https://mail.mozilla.org/listinfo/firefox-dev
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.mozilla.org/pipermail/firefox-dev/attachments/20161014/617fb916/attachment.html>