| 1 |
On Tue, Oct 10, 2017 at 11:56 PM, Duncan <1i5t5.duncan@×××.net> wrote: |
| 2 |
|
| 3 |
> Alec Warner posted on Tue, 10 Oct 2017 15:28:41 -0400 as excerpted: |
| 4 |
> |
| 5 |
> >> Please consider switching from your current 13.0 profile to the |
| 6 |
> >> corresponding 17.0 profile soon after GCC 6.4.0 has been stabilized on |
| 7 |
> >> your architecture. The 13.0 profiles will be deprecated and removed in |
| 8 |
> >> the near future. |
| 9 |
> >> |
| 10 |
> >> |
| 11 |
> > Can you commit to a deadline on this? |
| 12 |
> > |
| 13 |
> > Its OK to be wrong (e.g. say 1 month but remove in 3); but "near future" |
| 14 |
> > is not actionable by readers. |
| 15 |
> |
| 16 |
> Will the 13.0 profiles be removed all together, or per-arch? |
| 17 |
> |
| 18 |
> If they're removed all at the same time, then the time-limiting factor |
| 19 |
> will certainly be how long it takes the last arch to stabilize gcc-6.4+, |
| 20 |
> something that's likely not entirely predictable but that might take some |
| 21 |
> time, given gentoo's known issues with straggling archs. |
| 22 |
> |
| 23 |
|
| 24 |
> If the existing profiles will be deprecated and removed per-arch, with |
| 25 |
> some fixed time after gcc-6.4+ stabilizes on that arch as a goal, then |
| 26 |
> the time for most popular and best maintained archs may be predicted now, |
| 27 |
> but the time will differ for each one, so the best that could be done |
| 28 |
> would be either a time range or a list of the known ones, with presently |
| 29 |
> unknowns being added to the list in further revisions of the news item. |
| 30 |
> |
| 31 |
|
| 32 |
So my point isn't to be pedantic (that is why I said its OK to be |
| 33 |
incorrect.) |
| 34 |
|
| 35 |
"In the near future" to me could mean: |
| 36 |
|
| 37 |
1) tomorrow |
| 38 |
2) next week |
| 39 |
3) next month |
| 40 |
4) next quarter |
| 41 |
|
| 42 |
If we wrote: |
| 43 |
|
| 44 |
"The 13.0 profiles will be removed in six weeks, upgrade before then." Its |
| 45 |
clear to the reader that |
| 46 |
they should schedule this effort before the six weeks is up. It matters |
| 47 |
less if the six weeks is true; the email |
| 48 |
sets expectations regardless of the truth. |
| 49 |
|
| 50 |
We could rewrite it further to avoid the pedantry and say: |
| 51 |
|
| 52 |
"Please upgrade away from the 13.0 profiles in the next six weeks." |
| 53 |
|
| 54 |
This also sets expectations for readers, but avoids any specific guarantee |
| 55 |
around when Gentoo developers actually delete the 13.0 profiles. |
| 56 |
The reality of when the work is done matters significantly less than the |
| 57 |
expectation setting (as you imply there will likely be unknowable delays in |
| 58 |
deprecation and so forth, but users shouldn't take that as an opportunity |
| 59 |
to delay upgrades.) |
| 60 |
|
| 61 |
|
| 62 |
|
| 63 |
> |
| 64 |
> The other alternative might be to word it something like (1 year can be 6 |
| 65 |
> months or whatever instead, if that works better): |
| 66 |
> |
| 67 |
> "13.0 profiles are set to be removed one year after the last arch |
| 68 |
> stabilizes gcc-6.4+, with the goal for the gcc stabilization being the |
| 69 |
> end of 2017, meaning 13.0 profile removal is planned for the end of 2018 |
| 70 |
> if all archs meet their gcc-6.4+ stabilization goal." |
| 71 |
> |
| 72 |
> -- |
| 73 |
> Duncan - List replies preferred. No HTML msgs. |
| 74 |
> "Every nonfree program has a lord, a master -- |
| 75 |
> and if you use the program, he is your master." Richard Stallman |
| 76 |
> |
| 77 |
> |
| 78 |
> |
Archives