Revision as of 09:53, 19 July 2008

This was just an attempt and became invalid when the community rejected it. Read my arguments if you’re interested.

This page is was meant as an experimental structure-enhanced talk page for resolving the issue about IEC prefixes (Ki, Mi, Gi) with amounts of data. If you need more information, please see binary prefixes. The dispute is about whether or not these should be used and whether the SI prefixes (K, M, G) should be used in the binary sense, in the decimal sense, or both, depending on context.

This page works as follows: Major points have their own level-2 heading either under the #Pro section or under the #Contra section (that is pro or contra IEC prefixes). As the page is hierarchial, sub-points get a higher-level heading, like #Does it matter?. Please take care of this hierarchial structure.

Each point starts with a list of users who endorse. Please use #~~~ to endorse. After that list, each point has an “Arguments” section. Arguments are numbered. To refer to a previous argument, please use the number. Sign your arguments with --~~~~. Also, please only add new arguments, and make it as short as possible. Do not fly off on a tangent, and always argue to the point. Feel free to add new points if they really are new. Also, check if it fits under an existing point as a sub-point.

Pro and Contra have their own endorse-list and argument section. This is for an overall view. Please do not add arguments here that would fit somewhere else, but only those that concern the weighing and evaluation of the points in Pro and Contra.

The #Solutions section at the end provides possible agreements. You can endorse one or several solutions. The “Details” sections work like the “Arguments” sections. There you can supply details for possible implementations.

Pro

1. Quilbert (talk)

Arguments

1. If current literature is being inconsistent, that point is worthless. Encyclopedias and user-oriented IT magazines are written by professionals, but not by scientists, so their editors adhere to the old, inconsistent but familiar notation. We do not have to take over everything from professional encyclopedias, especially not their errors. Using IEC prefixes does not violate WP:NOR or the like because most major norming facilities prescribe them. --Quilbert (talk) 11:16, 11 June 2008 (UTC)

The IEC prefixes provide the necessary consistency

1. Quilbert (talk)

Arguments

Consistency is needed

1. Quilbert (talk)

Arguments

1. I searched the net recently, trying to find out whether Iomega Zip drives give their capacity in binary or decimal units. I haven’t found the answer yet. It were a great advantage if the information which unit is used were accessible directly from the article, e.g. when MiB is linked to binary prefixes. --Quilbert (talk) 11:16, 11 June 2008 (UTC)
2. Units are a mathematical measure. Data amounts can be part of mathematical formulas, such as the fraction ${\displaystyle {\frac {F}{V}}}$ filesize over volume size of a volume that a file uses up. Its unit would be e.g. ${\displaystyle {\frac {GB}{GB}}={\frac {2^{30}}{10^{9}}}\approx 1.074}$. This is mathematically more the unsatisfactory. Similarly, we would have ${\displaystyle KB/s=K{\frac {B}{s}}\neq {\frac {KB}{s}}}$, making “K” a nonlinear, context-dependant operator rather than a unit. --Quilbert (talk) 11:16, 11 June 2008 (UTC)

There is current literature that uses them

Arguments

The SI units have been defined in a decimal way much earlier

1. Quilbert (talk)

Arguments

1. They have been used that way since many centuries, and the norm dates back to 1960. --Quilbert (talk) 11:16, 11 June 2008 (UTC)

Many standardization facilities have taken over the IEC norm

1. Quilbert (talk)

Arguments

1. There is no operative norm concerning the matter other than this. --Quilbert (talk) 11:16, 11 June 2008 (UTC)

The norm has been taken over by storage medium manufacturers

1. Quilbert (talk)

Arguments

1. Hard drive manufacturers use decimal units. That is because of the norm. Of course they only adopted it because they were out for their own benefit, but that is irrelevant here.

The benefit of binary units in an IT context is fading

1. Quilbert (talk)

Arguments

Contra

Arguments

Current literature doesn’t use them

1. Quilbert (talk)

Arguments

The IEC prefixes sound silly

1. Quilbert (talk)

Does it matter?

Yes

1. Quilbert (talk)

Arguments

1. It is not desirable feeling like making fun of oneself whenever one needs to use the units. --Quilbert (talk) 11:16, 11 June 2008 (UTC)

No

Arguments

1. “Sounding silly” is not an objective argument, it doesn’t say anything about practical advantages or disadvantages. --Quilbert (talk) 11:16, 11 June 2008 (UTC)

SI prefixes have been used in a binary sense with bytes for a long time

Arguments

And that is very suitable in the computer universe

Arguments

Excessive exactness is not needed

Arguments

Solutions

1. Use the prefixes K, M, G, T in binary sense, as it was before manufacturers started adopting the norm.
2. Adopt the norm partially, i.e. use K, M, G, T in the binary or decimal sense, whatever is common practice.
3. Fully adopt the norm, but avoid the binary prefixes.
4. Fully adopt the norm and use primarily Ki, Mi, Gi, Ti.

Solution 1

Endorse

Details

Solution 2

Endorse

Details

Solution 3

Endorse

1. Quilbert (talk)

Details

1. This solution would keep the units the readers are used to, avoid the IEC prefixes, which to some sound ridiculous, but nevertheless conform to the standard, as it uses the SI prefixes in the new (“right”) way. In my opinion, it would be best to use templates with tooltips for clarification: Typing {{MB10|11}} would render as 11 MB (note the tooltip), and {{MB2|11}} as 11.53 MB. The automatic calculation is no problem. Additionally, footnotes with clarifying text could be used. --Quilbert (talk) 14:04, 11 June 2008 (UTC)

Solution 4

Endorse

1. Quilbert (talk)

Details

Articles affected by the new MOSNUM guideline introduced on 7 June 2008

On 7 June 2008 the WP:MOSNUM guideline was revised. The new guideline has been interpreted by some editors as justification for replacing unambiguous IEC units (MiB, GiB etc) with ambiguous ones (MB, GB, etc). In some cases an alternative method of disambiguation has been introduced in their place and in others not. Below is a list of affected articles

The case for deprecation ...

... is a weak one.

These are arguments for and against deprecation.

for

1. IEC prefix is rare and unfamiliar to many readers
2. etc

against

1. IEC prefix is unambiguous, simple to use and simple to understand
2. IEC prefix is supported by national and international standards bodies (IEC, BIPM, IEEE, NIST)
3. use in scientific publications is increasing: 1999-2001 (17 hits); 2002-2004 (34 hits); 2005-2007 (53 hits)
4. the alternative (binary use of SI-like prefixes) is deprecated by the same standards bodies
5. deprecation (of IEC) increases the difficulty threshold for disambiguation, reducing the rate at which articles can be disambiguated by expert editors
6. in turn this reduces the total number of articles that can be further improved by less expert editors with footnotes etc (if that is the consensus)
7. deprecation is interpreted by some editors as a justification for changing unambiguous units into ambiguous ones (per above list)
8. removing IEC prefixes from articles, even when disambiguated with footnotes, destroys a part of the information that was there before, because it requires an expert to work out which footnote corresponds to which use in the article
9. etc