On Thu, Apr 12, 2018 at 11:30 AM, Adam Williamson <adamwill@xxxxxxxxxxxxxxxxx> wrote: > On Tue, 2018-04-10 at 08:27 -0400, Gerald Henriksen wrote: >> Shouldn't the priority be to make Fedora welcoming to all, and not to >> limiting the audience based on an artificial size limitation? > > This is actually a tricky equation, because the reason to care about > size is actually *the same*: making Fedora 'welcoming'. Not everyone > has a big fat broadband pipe and a huge transfer cap (or unlimited > transfers). So if we include lots of i18n support we're making Fedora > 'more welcoming' in one way, but *less* welcoming in another way, to a > different (but probably overlapping) set of folks. It's a tricky > conundrum. My experience with crap bandwidth: I'm way more sensitive to the persistent metadata downloads, than that of rare and planned ISO download size. 1Mbps download means a 1.6G ISO is 3h50m and a 1.9G ISO is 4h30m. That difference doesn't even compute, I'm gonna use curl and walk away no matter what. And when I had the pure amazing joy of 350kbps recently, this is now 11 vs 13 hours, again it's the same: craptastic. People who are bandwidth sensitive, especially if they have multiple computers, need to setup a local mirror. *shrug* https://pagure.io/fedora-workstation/issue/46 says it might be too late to use OTC fonts but doesn't definitively state that it is. With five days before freeze: OTC vs subpackage vs grow the ISO by 260MB vs use Fedora 27 fonts. I would push for going in that order through whatever path is the least resistant assuming no already known liabilities and all of those have been tested. Going with Fedora 27 fonts I'd consider the last restort, it's a conservative fallback, it will work but with no improvements for Fedora 28. -- Chris Murphy _______________________________________________ devel mailing list -- devel@xxxxxxxxxxxxxxxxxxxxxxx To unsubscribe send an email to devel-leave@xxxxxxxxxxxxxxxxxxxxxxx
