Use IEC standard unit notation (KiB, MiB, GiB, TiB) in AddOns

[kriukas]

Maybe overkill after posting in mainline development, so feel free to delete as duplicate. But perhaps in this post it is possible to insert the result of the informal initial audit/investigation which Add-Ons are impacted.

For all dear Add-Ons developers consideration for implementing unambiguous alignment in implementations where possible+reasonable:

Use IEC standard unit notation (KiB, MiB, GiB, TiB)

 

[dave14305]

I’m not sure how you can be so bad at posting links. The links in both forum posts are broken, and the link in the RTRMON Github issue is also invalid.

 

[kriukas]

I’m not sure how you can be so bad at posting links. The links in both forum posts are broken, and the link in the RTRMON Github issue is also invalid.

Pardon. Fixed. OTOH - When forum link-insertion dialog is like 100px wide - one doesn't WYSIWYG...

 

[Squall Leonhart]

nah

 

[Ripshod]

Why try to change what's normal?

 

[Ripshod]

Pardon. Fixed. OTOH - When forum link-insertion dialog is like 100px wide - one doesn't WYSIWYG...

Double post.

Use IEC standard unit notation (KiB, MiB, GiB, TiB)

For dear mainlines developers consideration for implementing unambiguous alignment in implementations where possible+reasonable: Use IEC standard unit notation (KiB, MiB, GiB, TiB) EDIT: URL fixed.

www.snbforums.com

 

[Tech9]

I personally don't like Kibibytes and Mebibytes. Voting against changes.

 

[kriukas]

Why try to change what's normal?

It depends on perspective. Cause sometimes "normal" is ambiguous. That's it. No more, no less. And that's why ISO-level standard was agreed upon to unclutter. ISO/etc. standards are not needed at all. Plugin wire / memory and here you go. But somehow we all want some level of quality. That was the point.

 

[kriukas]

Why try to change what's normal?

Even Asus/Merlin does "new normal" now:

scsi 0:0:0:0: Direct-Access Patriot Rage Pro 1100 PQ: 0 ANSI: 6
sd 0:0:0:0: Attached scsi generic sg0 type 0
sd 0:0:0:0: [sda] 241637376 512-byte logical blocks: (124 GB/115 GiB)

from dmesg 3 mins ago

OK, humour aside - MiB / GiB is simply convenient when working with permanent storage. If you attach a larger drives - you do not have to do mental math, as difference between GB & GiB grows.

 

[Ripshod]

I just find the SI units simpler. For those not mathematically inclined 1kB (1000 bytes) is a lot easier to work with than KiB (1024 bytes), especially when working with terrabytes.
We're working with the general public here, not engineers.

Even Asus/Merlin does "new normal" now

Whose choice was that, Asus, @RMerlin or both.

In the 70s the uk was forced from imperial to metric. Now the self-appointed bureaucrats in Geneva want us to use a derivative of binary. Not gonna happen fully as the general public don't work that way.
Fine for engineers but I don't think so for a worldwide public forum.

 

[Tech9]

For those not mathematically inclined 1kB (1000 bytes) is a lot easier to work with than KiB (1024 bytes), especially when working with terrabytes.

Works better for the marketing as well. The same size storage device is more KB than KiB, etc. I just checked a few large storage manufacturers and none use TiB for their products storage capacity.

 

[ColinTaylor]

Whose choice was that, Asus, @RMerlin or both.

Neither. The decision to display that information comes from the Linux developers. It's been displaying it like that for at least ten years (probably much longer). IIRC there appeared to be a push about twenty years ago to clean up some of Linux's output to use the correct units.

There are well established norms for certain things, e.g. memory in 1024's, bitrates and disk capacity in 1000's. Problems arose IMHO when software programmers who were used to dealing with memory addressing started to incorrectly apply "their units (1024)" to things they shouldn't, like bitrates. Microsoft was/is one of the biggest offenders of this - compare how MacOS and Windows would report the size of the same disk. Then there are ambiguous areas like download speeds, is the test reporting the true bitrate (including all overheads) or just the number of bytes transferred. If it's the latter should the "byte-rate" be in 1000's or 1024's? (And don't get me started on flash memory cards.)

 

[dave14305]

ASUS QoS page converts the input bandwidths from MB/s to KiB/s when saving in nvram, but the underlying tc commands assume non-iec input unless told otherwise via the -iec option.

 

[kriukas]

I just find the SI units simpler. For those not mathematically inclined 1kB (1000 bytes) is a lot easier to work with than KiB (1024 bytes), especially when working with terrabytes.
We're working with the general public here, not engineers.

Case closed - every party have its own opinion. Discussion was open - and that's great

Whose choice was that, Asus, @RMerlin or both.

I don't have a clue - just selling what I bought. Was simply debugging failed partition/fs on USB drive and saw from dmesg output. And bit**ed a little. From my POV, if its in the dmesg and standard drives, it comes from upstream linux.

Great for me, as many storage management commands lets you freely set almost any storage unit - except bits if I remeber correctly.

In the 70s the uk was forced from imperial to metric. Now the self-appointed bureaucrats in Geneva want us to use a derivative of binary. Not gonna happen fully as the general public don't work that way.
Fine for engineers but I don't think so for a worldwide public forum.

Why self-appointed? I would understand ranting about the EU crappy legislation going of the reasonable rail, but as per Wikipedia, organization is old and multi-party, made many standardization work, noble people work there. Of course, I do not have insights regarding last decades. Maybe there are the same problems in Geneva as in EU.

 

[killweav]

However you want to calculate it (base 10 or base 2), just know that in some parts here in the US, people call it jigga-byte. As in "I want the 32 jigga-byte USB drive."