Re: Sticks, carrots, real XML, ultra-liberal parsing

[Henri Sivonen <hsivonen@xxxxxx>]

On Jan 10, 2004, at 20:07, Isofarro wrote:

> What can be done to prevent Atom from being equally difficult to 
> implement in seven or eight years?

The PNG format is still as implementable as it was back when it was 
defined. I think there are three reasons for this:
1) The spec is clear
2) A comprehensive library was available from day one
3) People can't see what binary format looks like in a text editor so 
sloppiness was expected to lead to trouble. (With text people somehow 
seem to expect the software to do similar error correction people would 
do when reading text.)

So I think with Atom we need:
1) A clear spec
2) Libraries that are so good most people don't want to roll their own. 
These libraries need be based on real XML tools and be especially 
available for languages that have a culture of regexp hacks and whose 
XML tools are hard to use. (That is, providing an Atom library for the 
default configuration of PHP4 is more important than providing one for 
Java.)
3) An atmosphere that emphasizes that publishing non-conforming data or 
using an ultra-liberal parser (which rewards those who publish 
non-conforming data) is uncool.

Once the majority is using the good libraries, those roll their own 
have to play by the rules.

> > On the other hand, people *will* publish well-formed feeds (if they 
> > bother trying to publish an Atom feed at all) if the popular 
> > aggregators use real XML processors.
>
> Will _enough_ people publish well-formed feeds?

The situation with Trackback indicates that Movable Type has large 
enough an installed base that a protocol which depends being 
implemented by both sides finds enough lines of communication with MT 
on both sides for the protocol to be perceived as useful. I think there 
will be enough well-formed feeds out there if MT produces well-formed 
feeds.

However, one important question is, why anyone should want to pull the 
Atom feed from MT instead of the RSS 2.0 feed. That is, can Atom 
provide something new and desirable enough that RSS 2.0 doesn't? The 
problem with XHTML is that as a mere reformulation of HTML it doesn't 
allow you to express anything you really, really would want to express 
but couldn't express in HTML.

> What's the argument for doing Atom in XML?

The technical argument is that Atom would benefit from from the 
applicability of XML tools. However, this argument is moot if Atom 
isn't real XML.

The marketing argument is that XML makes it look cool. After all, those 
orange badges say "XML"--not "RSS", because XML has positive marketing 
value.

--
Henri Sivonen
hsivonen@xxxxxx
http://iki.fi/hsivonen/