What defines a standard?

Question

What defines a standard like HTML5, C++0x, etc.? Is it just that you hand something in to W3C/ANSI/ISO/..., they produce a couple hundred pages long document and suddenly it's a standard?

Can't I, as an individual, create something and standardize it by myself? I surely could produce a couple hundred pages long document which describes my creation in every detail. So what is the benefit of

publisher: W3C

Stand...

?

What really defines a standard is when people outside the standardization group cares about it. There's a ISO standard for AC power plugs but besides Brazil no one uses it. It may be an ISO standard but is far from being a de facto one, for example. If many people cares about the DaveO Standard #1:2001 then it *is* a de facto standard. — Vitor Py, Mar 08 '11 at 17:35
As one of my co-workers used to say: The good thing with standards is that you have so many of them to choose from. :-) — , May 10 '11 at 12:17
Worth pointing out that ANSI is actually a member organization of ISO. — Parthian Shot, Jun 25 '15 at 01:55

score 13 · Accepted Answer · answered Mar 08 '11 at 12:32

A standard isn't just some mystical group declaring "What Shall Come To Pass". It's defined as a common set of rules and goals to which a collection of individuals or groups that, as a whole, agree to adhere. In the case of the W3C, the browser developers have all tacitly agreed to adhere to the set of rules that is constructed by the committee put in place to handle these questions. I don't know what the make up of the top level portion of the committee is, but I do know that participants to the construction of the standard is inclusive of hundreds of web (html/css current) experts who are all considered leaders in the field, and, as I recall, each of the major browser producers has representatives on the committee.

Since there is nothing binding about the standard, it is just a guideline document to help browser producers make it easier for developers to produce pages that render properly in their software.

As for creating a standard, yes anyone can create a standard, publish it and call it a standard. It won't mean anything unless some collection of individuals buys into it and agrees to adhere to its principles and rules. As an example, our Internet Services team has created a standard for code production. It follows most general guidelines but deviates in specific areas. The junior developers and interns had no real say in it, but the leads got together and decided that these rules were the best rules for producing software for the team. Once we were all in agreement and put it into practice, it became a standard (for us).

Look through the RFCs at http://www.ietf.org/rfc.html and you get a feel for what goes into the standards process. Many RFCs are still on the standards track. This site will let you review comments on RFCs as well as obsoleted RFCs. Other standards organizations follow similar processes. — BillThor, Mar 08 '11 at 15:05
@BillThor - I'd put forward that most RFCs are actually treated as standards. 802.11 is an RFC, yet companies use it as a measure of "compliance". Again, it boils down to what a common group of individuals (companies) agree will behave as the standard. RFCs are a good place to look at common standard criteria though. IEEE also maintains good standards documents. — Joel Etherton, Mar 08 '11 at 15:13
I was considering mentioning something like. _Violating RFCs is done at your own risk, as they are de-facto standards._ That's another part of the standards process, in that compliance often begins well before the standard is accepted. (Is this a chicken and egg kind of situation?) — BillThor, Mar 08 '11 at 15:23
"Since there is nothing binding about the standard, it is just a guideline document" - this is often true in our field of software engineering, if you look at many other ISO (or similar) standards and such there often are legal requirements to follow those. — johannes, Feb 20 '13 at 02:59

score 1 · Answer 2 · edited May 09 '11 at 22:32

When speaking of "file format standards" or "data format standards" (Not standard methodologies), there are another 2 concepts that mix with "Standard":

(1) "frequently used data format" or "frequently used file format"

This case, is a data / file format that is commonly used, sometimes has subbranches, but is not regulated by a organization to be considered a standard. I believe JSON started like this, and eventually became a standard.

(2) "open, not frequently used, data format" or "open, not frequently used file format"

PDF started like this. Some programmers used PDF, but did not know its internals. Later, Adobe "opened" the format and it became available, even if it wasn't commonly used. The same goes for the PNG image format.

So, in theory you may create your own data / file format, expect it to become commonly used, and later become standardized by an organization.

What defines a standard?

2 Answers2