IRC conversation:

<cosimoc> drago01, do you have any sort of insight that more websites use base 2 vs base 10?
<cosimoc> drago01, I don't think we should change it back to be honest
<cosimoc> it's going to be inconsistent with something no matter what we choose
<drago01> cosimoc: not sure have to check dropbox, skydrive etc
<cosimoc> right
* mccann (~mccann@vtelinet-216-66-97-217.vermontel.net) hat #gnome-os betreten
<drago01> cosimoc: initally I did not care much but after the discussion I had to today (aka "real world case") it looks like a mistake to me
<cosimoc> using base 10 is at least consistent with what manufacturers use
* gcampax hat die Verbindung getrennt (Ping timeout: 600 seconds)
<cosimoc> which is the reason why we did it in the first place
<drago01> yeah I understand that (I wrote that in the bug)
<drago01> but the inconsistency with there has been around "for ever" so people are more or less used to it
<drago01> but having someone reupload a file multiple times due to that is bad
<cosimoc> I agree that is bad, but I also don't think people checking sizes from two very different places and expecting them to match is such a common thing
<drago01> dunno about that ... but seeing a different value for the same thing can be very much a source of confusion
* mclasen_ (~mclasen@66.187.233.207) hat #gnome-os betreten
* fabiand hat die Verbindung getrennt (Verlassend)
<drago01> cosimoc: https://bugzilla.gnome.org/show_bug.cgi?id=695146
<Services> Bug 695146: normal, Normal, ---, nautilus-maint, UNCONFIRMED, Go back to MiB
<drago01> cosimoc: seems not to be that uncommon
<drago01> cosimoc: (was googling for "dropbox Mib")
<drago01> cosimoc: again different value for the same thing
<cosimoc> drago01, did you want to send a different link?
<drago01> cosimoc: err yes sorry
<drago01> cosimoc: http://forums.macrumors.com/showthread.php?t=1358911
<cosimoc> drago01, so the day google drive or dropbox switch to base 10 we're going to change again? :)
* ssam2 hat die Verbindung getrennt (Leaving)
<drago01> cosimoc: if other popular sites / services do it to and people start to get confused I'd say yes
<cosimoc> drago01, I don't know...it's also confusing for people that buy a 100gb drive that it doesn't show up as 100gb on gnome...
<drago01> cosimoc: also random sites that tell you "download foo x MB" will end up using MiB most of the time
<cosimoc> (citation needed)
<drago01> let me find one
<drago01> or tons of them
<drago01> that should be easy
<drago01> cosimoc: https://fedoraproject.org/
<drago01> cosimoc: 916MB here means 916Mib
<cosimoc> I don't disagree with you that this whole situation is unfortunate
<cosimoc> but a single standard doesn't exist
<drago01> cosimoc: I know about the harddirve thing ... but have you seen that many people get confused because of that? I mean it isn't ideal but that is the lesser evil as this inconsistency has been around forever
<cosimoc> expecting people to think in base 2 is just not correct IMO
<drago01> I don't expect them to think in any bases ... they I expect them to see the same values for the same thing
<cosimoc> drago01, anyway I don't think I'm going to change it for this release
<cosimoc> drago01, as I said there's always going to be inconsistencies
<cosimoc> one way or another
<cosimoc> you have no guarantees which base a random website will use
<cosimoc> so the choice is between being consistent with what some popular services do
<drago01> cosimoc: is there any website you know of that shows MB (as in base 10) ?
<cosimoc> or be consistent with HW manufacturer
<cosimoc> drago01, I never checked tbh
<cosimoc> windows 8 apparently still uses base 2
<cosimoc> drago01, in any case, let's put it up for discussion for next cycle
<cosimoc> I think we're too tight on time right now to make any decision for 3.8
* tacg hat die Verbindung getrennt (Ping timeout: 600 seconds)
<drago01> cosimoc: well I just have had an interaction where it caused confusion and after thinking about it I am convinced that this is wrong
* gpoo (~gpoo@sai.cs.UVic.CA) hat #gnome-os betreten
<drago01> cosimoc: it does not match what most webservices or even other applications do (firefox reports MiB for example) 
<drago01> cosimoc: sure it is not that urgent to fix it ASAP
<drago01> cosimoc: but it *is* a problem
<cosimoc> drago01, I thought Firefox used the glib format too
<drago01> cosimoc: it does?
<drago01> cosimoc: no
<drago01> cosimoc: 6.2MB file in nautlius downloads as 5.9MB
<alex> I wonder what firefox does on osx then
<alex> osx uses base10
<drago01> alex: good point let me try
* fmuellner (~fmuellner@194.Red-88-22-202.staticIP.rima-tde.net) hat #gnome-os betreten
<drago01> cosimoc, alex : base 10
<drago01> cosimoc, alex : err base 2 actually
<drago01> shows 5.9 for a the 6 Mb file
<drago01> so it is a mess there too
<drago01> but just because they do it don't mean we have to do the same mistake
* tacg (~tiago@bl19-187-178.dsl.telepac.pt) hat #gnome-os betreten
<drago01> cosimoc: mind if I copy that conversation in into the bug for reference?
<drago01> cosimoc: seems we will not agree now anyway
<cosimoc> drago01, sure

There's no way to win in this game. Best not to play. At least we are internally consistent now. Nautilus, filechooser, gnome-disks, gnome-system-monitor, control-center, all use GB for file sizes (and GiB for memory sizes, where it applies).

(In reply to comment #2)
> There's no way to win in this game. Best not to play. At least we are
> internally consistent now. Nautilus, filechooser, gnome-disks,
> gnome-system-monitor, control-center, all use GB for file sizes (and GiB for
> memory sizes, where it applies).

Being internally consistent is easy (as we control it) but where are inconsistent with the "rest of the world" (tm) (modulo OS X) we cannot fix the others to be consistent with us so we should do what the majority does to minimize inconsistencies.

I'd prefer GNOME-wide consistency so fixing this in Nautilus only (this bug report) should be a no-go, so I propose a WONTFIX.

(In reply to comment #4)
> I'd prefer GNOME-wide consistency so fixing this in Nautilus only (this bug
> report) should be a no-go, so I propose a WONTFIX.

Who or what is stopping us from fixing this everywhere else? 

Pretending that problems do not exist and close WONTFIX is just stupid.

Feel free to present a solution.

(In reply to comment #6)
> Feel free to present a solution.

I did go read my comments again. Writing patches to do this is trivial I can do that once we agree on the what we want to do.

Honestly I have no idea what your problem actually is ...

(In reply to comment #7)
> (In reply to comment #6)
> > Feel free to present a solution.
> 
> I did go read my comments again.

You don't even have to do that, just reading the subject is sufficient ...

Ah. WONTFIX then, but I also wrote that before already.

(In reply to comment #9)
> Ah. WONTFIX then, but I also wrote that before already.

OK, you are free to put your head in the sand and ignore real world issues .. so how about staying out of the discussion if you don't have anything useful to contribute?

*** Bug 735797 has been marked as a duplicate of this bug. ***

Maintainers have expressed that they don't intend to revert this, and nothing new has been brought to the table in years. Closing this as WONTFIX which is effectively the same (but better) as saying no and letting the open report adding noise to the list of bugs.