User: Password:
Subscribe / Log in / New account

PostgreSQL and the SQL standards process

PostgreSQL and the SQL standards process

Posted Sep 21, 2011 22:02 UTC (Wed) by n8willis (subscriber, #43041)
In reply to: PostgreSQL and the SQL standards process by Simetrical
Parent article: PostgreSQL and the SQL standards process

That's true; I hope I did not suggest somehow that all standards bodies are the same. However, even W3C standardization processes can be slowed down by antagonistic participants with their own agenda. Open source projects are not impervious to being said troublesome participant, either, naturally. And I have heard from W3C folks that not as many FOSS projects get involved as the group would like; I chalk that up largely to Berkus' Challenge #2: it's hard to get someone free to sit in on phone and in-person meetings.


(Log in to post comments)

PostgreSQL and the SQL standards process

Posted Sep 21, 2011 22:18 UTC (Wed) by Simetrical (guest, #53439) [Link]

Different parts of the W3C do things differently. The key Working Groups (HTML, WebApps, CSS, SVG) tend to follow whatever the major browsers' implementers agree on, so no one other than implementers can really cause much trouble. Of course, sometimes implementers won't agree on something, like media format support for <video>. But there's not much to do about that. Fortunately, right now all the major browser implementers are largely committed to interoperability, so we don't have a lot of obstructionism.

The active participants in the Working Groups I'm involved in tend to almost all be employed by one of the major browser implementers: Microsoft, Mozilla, Google, Apple, and Opera. Some of these (Mozilla, WebKit, Chromium) are FOSS. There are a nontrivial number of people participating in their spare time, too. I was one for a few years before I started getting paid to do it this year.

The groups I'm active in (HTML and WebApps) don't do anything significant by phone or in-person meetings. It's all mailing lists, IRC, and bug trackers. They're really run like open-source projects.

PostgreSQL and the SQL standards process

Posted Sep 22, 2011 18:42 UTC (Thu) by jwakely (guest, #60262) [Link]

Different parts of ISO do things differently too: the ISO C++ committee may not be perfect but by the sound of things it's a lot more open than the SQL one.

PostgreSQL and the SQL standards process

Posted Sep 23, 2011 5:35 UTC (Fri) by ncm (subscriber, #165) [Link]

That's a recent change. Curiously, it hasn't been a result of changes in ISO rules, but just interpretations of them. Committee Drafts are still secret, but Working Papers, effectively identical, are not. Likewise, published Standards. Committee mailing lists and issues lists exist entirely outside ISO's domain. Even voting rights make much less difference than one might think; actually doing the work carries more weight, because when in doubt, voters vote with whoever seems to know more, and most voters are always in doubt.

PostgreSQL and the SQL standards process

Posted Sep 22, 2011 22:06 UTC (Thu) by justincormack (subscriber, #70439) [Link]

A lot of the rest of the w3c doesnt work like this though, and a lot of this appened because of html5. Before tht there was a mess of unused xml standards...

PostgreSQL and the SQL standards process

Posted Sep 22, 2011 22:17 UTC (Thu) by Simetrical (guest, #53439) [Link]

Yep, definitely. But all the W3C standards I can think of that are undergoing active development, and that are targeted at web browsers, work entirely in the open these days. There may be exceptions. But even the worst WG in the W3C publishes its standards for free on the web, and is required to listen to feedback from the general public, so it's still way better than the SQL situation described in this article.

listening to feedback

Posted Sep 23, 2011 13:29 UTC (Fri) by pjm (subscriber, #2080) [Link]

What does "required to listen to feedback" mean? Sure, anyone can post to the mailing list, but my experience with the CSS mailing list (www-style) is that it can take months to get a reply, and the eventual reply can still be "we've decided not to address this" without giving any reason. There's a w3c policy document that says that this "should" not happen; yet happen it does.

While it's true that the CSS 2.1 specification is published on the web, it's also true that the published specification contains contradictions and does not contain enough information to pass the test suite.

Yes, it's like an open-source project: no-one's under any obligation to address the bug reports you file.

listening to feedback

Posted Sep 23, 2011 17:40 UTC (Fri) by Simetrical (guest, #53439) [Link]

Yep, nothing is perfect by any means. It's a far cry better than what's described in the article, though. For what it's worth, some of the more recent specifications are far more detailed, and hopefully do contain enough information to pass their test suites. CSS 2.1 is ancient, although it only just made REC.

But all the CSS specs are less precise than the newer HTML/DOM specs. At some point we're going to have to rewrite CSS from scratch with the same level of precision as HTML and DOM have been rewritten. We've learned a lot over the last decade about how to write standards.

Copyright © 2017, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds