Tuesday August 17, 2004
Tuesday August 17, 2004
Re: This is about Atom
Beautiful. I hope the more vocal members of the Atom mailing list read it. I think Matt makes some important points that these members need to keep in mind, not that some of them aren't already.
But what Atom says is: If you want to play the weblog game, this is the metadata you need. That, and some other stuff (it's all in the Atom spec).
. . .
What Atom says is: If you want to afford these behaviours, then you have to be weblog-like in this respect and editor-like in this respect, and it doesn't matter what else you do, but those are the shapes we require.
(
Aug 17 2004, 08:32:33 AM
)
Atom
Permalink
Trackback: http://www.brainopolis.com/roller/trackback/lance/Weblog/re_this_is_about_atom
Tuesday June 15, 2004
Re: Worst RSS and Atom article ever
As you may have guessed, I support Atom. But I am distressed when Atom (and RSS) are represented so poorly as in the article that Phil rips apart. What got a chuckle out of me was this line of Phil's:
If you are creating your own RSS feed by continually scraping your own HTML, you really need to fire your entire IT staff.
If this is representative of their current quality I'm glad I let my subscription to New Scientist lapse.
Of course, Dave Winer has to follow that up with his own potshot at Atom:
... it's totally clear now that Atom is no more "open" than RSS is.
The only thing I saw in there that suggested any such thing was a quote from Winer himself.
(
Jun 15 2004, 08:06:57 AM
)
Atom
Permalink
Trackback: http://www.brainopolis.com/roller/trackback/lance/Weblog/re_worst_rss_and_atom
Wednesday March 10, 2004
Musings on Atom/RSS
I've been reading various posts about Dave Winer's proposal to merge RSS into Atom. Most of them have been, shall we say, negative.
On the wiki and the mailing list this sort of thing has been proposed before, in one form or another (RSS1 vs RSS2). I'm tempted to echo Erik (echoing Matt) that I just don't care. And I guess I never have. For a long time I was a follower in matters Echo/Atom, all I cared about was making sure Roller generated proper Atom, mostly a "me too" knee-jerk reaction.
But then last fall I decided to implement the Atom API in Java and created Atom4J. This I care about, but I'm hardly religious. So far I've had little trouble following all the ups-and-downs, the changes to format and protocol (though I'm leary of SOAP and haven't implemented that yet). I like the RESTiness of the current specification, finally REST makes some sense to me. But then Russell pointed out this won't work on most (all?) deployments of J2ME. I've even held back implementing the Authentication layer proscribed last fall - and so far as I can see it has fallen off the table.
Finally, back to Erik, I don't see what's so hard about the Atom Syndication Format or its API. Neither is any harder to hand-carve than any other XML I've encountered (considerably easier than Ant for certain). And although various very nice (and deluded) people have claimed that I am exceptionally smart (well, my wife pretty much has to say that, doesn't she?), I don't think it's true; and if *I* can figure out Atom I'd say just about anyone can.
End Rambling
(
Mar 10 2004, 01:32:00 PM
)
Atom
Permalink
Comments [1]
Trackback: http://www.brainopolis.com/roller/trackback/lance/Weblog/musings_on_atom_rss
Monday March 08, 2004
"that other format"
Not to suggest you read anything into it, I found it funny how this reference was phrased:
2RSS.Com has a public and free service that converts non-RSS feeds, such as those generated by Blogger, to RSS.[Scripting News]
The only non-RSS conversion I found there was Atom, so why didn't Dave just say "... converts Atom feeds, such as those generated by ..." instead? It isn't like Dave is shy about using the word Atom, and he has said several times that he would support Atom - when it is ready. I'm sure there is nothing to it, I just found "non-RSS" amusing.
(
Mar 08 2004, 09:14:37 AM
)
Atom
Permalink
Trackback: http://www.brainopolis.com/roller/trackback/lance/Weblog/that_other_format
Monday March 01, 2004
Atom4J
I more-or-less finished all my work on Atom4J early last week, but I've been having trouble with my CVS connection to OSJava.org (as in, nonexistent) so I haven't been able to checkin my new code. On the off chance you're interested, you can check it out grab the source in a zip file. I won't guarantee how long I'll leave it there, but I'll try and put notice here once I get Atom4J checked into the repository (which is when I'll likely remove the zip file).
Update 03/03/04: I was able to check-in my changes and upload files last night. Note to self: update the osjava.org homepage.
(
Mar 01 2004, 12:56:40 PM
)
Atom
Permalink
Trackback: http://www.brainopolis.com/roller/trackback/lance/Weblog/atom4j1
| For further reading on today's posts: |
|
Tuesday February 17, 2004
Writing Atom XML
As I've been working on bringing Atom4J up to spec I've been looking at how I'm generating Atom XML. Currently the objects know how to write themselves (can you say "to String") - which is how Sandler does it as well. At this maintenance stage (where the spec is a moving target) working with a StringBuffer to output this stuff is a pain, and no fun. So I've been considering my options:
- toString() with StringBuffer: works, but a pain to modify, and difficult to write neat (pretty print) XML.
- Templates: I'm thinking Velocity here (since I know it) but others are possible. This is very attractive in some ways, not to mention the *cool* factor, but introduces yet another dependancy and puts that burden on anyone who uses the Atom4J library.
- XMLWriter: Henri wrote a nice XML generating library for quickly generating XML from Java objects. Same issues with another dependancy (though it is a *much* smaller one).
I'd be interested in your thoughts, pro/con for any of these options, or suggest something else altogether.
(
Feb 17 2004, 10:08:19 AM
)
Atom
Permalink
Comments [5]
Trackback: http://www.brainopolis.com/roller/trackback/lance/Weblog/writing_atom_xml
Monday February 16, 2004
Atom4J continues
I've been reticent to talk about this, since the opinion was overwhelmingly that I should join the Sandler effort. I took a longer look at the Sandler source, and I'm not comfortable with the design; two issues in particular.
One, that an Interface is used to hold static variables - which are then inherited through implementing that interface. This appears to be one of those 'religion' issues where you either believe this is a good way to hold values or you don't. I've never formed an opinion on this before, but apparently I don't believe in it.
Two, that the data objects themselves read their values from the XML source. Again, I'm not saying this is wrong, it just doesn't feel right to me. The only real objection I can come up with (and it's weak) is that this doesn't lend itself to 'persistence'. I'm not saying that right, but for now that will have to do. My other concern with this is that it *may* preclude Extending the Atom API, but this is a gut feeling, no evidence to substantiate this claim.
Three, no test cases. From looking at the code it wasn't evident to me how to test Sandler against the example files (Atom3_Min.xml, Atom3_Max.xml, Atom3_Multipart.xml). Without installing it on a server and hitting it with the python client I couldn't verify Sandler worked. I forgot to include Test Cases in the Atom4J source distribution, I'll fix that next go 'round (but they are in CVS). I know, I said there were only two issues - but this isn't a design issue.
The approach I've taken in Atom4J (please shoot me, I wanted a better name) is to separate the data objects from the mechanism by which they are populated. It just seems to me that the object model should not be tied to the persistence (XML in this case). Currently I'm using Digester, which is open to critique - namely that it is slow. The way I've architected Atom4J (and I use the term 'architected' loosely) will allow me to implement other mechanisms later.
(
Feb 16 2004, 09:55:55 PM
)
Atom
Permalink
Comments [8]
Trackback: http://www.brainopolis.com/roller/trackback/lance/Weblog/atom4j_continues
True Atom Clients?
Are there any Atom API clients other than wxAtomClient? There are a number of "feed readers" that support the Atom Syndication Format - but do any of them support the API? I don't think I want to do that research, so I'll lazyweb it and leave it to my dear readers to answer. Thanks!
(
Feb 16 2004, 01:19:09 PM
)
Atom
Permalink
[Link]
Trackback: http://www.brainopolis.com/roller/trackback/lance/Weblog/true_atom_clients
