Wednesday, January 14, 2004|
and another What he said. +1 3:20:25 PM  comments () trackback []
|
|
rss vs. atom The big difference between the being liberal in parsing RSS and being liberal in parsing Atom is that when all of this momentum got started, when RSS really took off, was at the point where there were many people providing feeds that they claimed were RSS. And as Mark points out here, many major and desirable news feeds did not (not sure about now) validate as RSS. Atom does not have NEAR the user base that RSS had when aggregators took off. If there ever was a time to set the bar for producing valid Atom feeds, now is the time. Now is the time to produce Atom toolkits that produce valid feeds and hand them out. Now is the time to make sure that people use existing tools to make sure that their Atom feeds really are valid. If Atom fails because it is too hard to produce a valid feed, then that says something else about the spec. If tools are needed to produce valid Atom feeds, then make sure the tools are out there. Hand out black boxes that say "Content goes in here" and "Atom comes out here," and can be dropped in to other projects. Provide them in Java, C, Python, Perl, Ruby, PHP, C#, and any other language that blogging and aggregation tools are using.
Right now -- before the explosion of Atom feeds -- right now is the time
to enforce validity. |
|
what he said +1 for this. 11:04:15 AM comments () trackback []
|
|
the new blood sport on the web We love to argue about standards, and about what is right. One current topic is Postel's Law, which is in reference to data formats, and it can be summed up as "Be liberal in what you accept, be conservative in what you produce." Reading the current batch of posts, one side is saying that you have to be as liberal as possible, and the other side says that there's a point at which you have to draw the line and move on. The thing that seems ridiculous about all of this is that of course there's a line drawn. Even using the ultra-liberal feed parser that Mark has produced, there will be feeds that fail. If I claim to have a feed, and provide gibberish, should I insist that you have to successfully parse my stuff? If I claim to have an Atom feed, and then point you at an Excel file containing my posts, should I be able to insist that you need to parse my stuff? Of course not. The examples I gave are totally bogus. We can agree on that bit. So there is a line somewhere. Where is the line?
One side says that the line is drawn at the point of well-formed XML.
The other side says that that is not enough. You should be able to deal
with badly formed XML. Well, how badly formed can it be? Can I leave
off closing tags? Can I forget to encode entities? You know, instead
of calling everybody names, if you want to accept non-well formed XML,
you should write up something defining what you will accept and what you
won't. There is a line. There is a point where someone can claim they
have an XML feed, but it won't be XML. You cannot assume that there is
a quiet agreement among everyone about what is and what isn't parseable.
We have already proven that that is not the case. |
Last update: 6/29/2005; 4:20:33 PM .
