Charl van Niekerk » Blog: June 2004

Main

Latest

XML MIME Types

Ok, this post derived from a discussion on the comments of previous posts, but I'm just going to summarize everything here for future reference.

You see many MIME types like these:

application/xhtml+xml
application/atom+xml
application/rss+xml
application/rdf+xml

Yes, many different XML Schemas have their own MIME type. The rest will just have to do with application/xml. However, why the +xml bit at the end?

Well, not all user agents support all of these individual XML Schemas. However, that doesn't necessarily mean that they can't do anything with them. Therefore, we always need to let them know that a particular file is XML regardless. The +xml bit at the end signifies this.

Neither Mozilla nor Internet Explorer seem to make use of this (yet). However, Opera does. (Thanks, Anne).

Anything to add? Comment!

Microsoft quickly fixed up Channel 9

This morning everything looks fine using Internet Explorer 6.

Microsoft crashes their own browser

I just read the following post by Anne van Kesteren: Channel 9 does a markover and restyles a bit.

Apparently Channel 9 at Msdn.com just got a redesign, where they claim to have worked on cross-browser compatibility.

I decided to test the site out in three browsers, namely Opera 7.51, and Firefox 0.9, and Internet Explorer 6, on my screen at 800x600 resolution. Here are the results:

Opera 7.51 - Fine
Firefox 0.9 - Fine
Internet Explorer 6 - Broken

Yes, broken. And I have screenshots to prove it. I don't know if it is just my version (6.0.2800.1106 to be precise), but the site is in pieces with the boxes not being aligned correctly.

This must be the joke of the decade or something. I still cannot believe it myself. Microsoft actually goes and breaks their pages ~~for~~ in their own browser, while having other browsers display the pages more-or-less correctly.

I think the problem is that the developers have been using a higher screen resolution. However, many people (like myself) are still using 800x600, so you must cater for that by testing. Obviously, this testing has not been done propperly.

Atom feed

One of the things I like about Blogger is the Atom feed. I also have one available at http://charlvn.blogspot.com/atom.xml.

However, there are a few things wrong with it. Firstly, it is sent as text/xml. It seems like most servers are configured to send .xml as text/xml by default. Why not rather application/xml?

Anyway, Atom feeds have their own content type: application/atom+xml. Some examples of correct feeds:

The first one on the above list I set up correctly to send under the correct content type using ASP.NET It is actually very simple, using the following code:

<% Response.ContentType = "application/atom+xml"; %>

Note that the above is in C#, which is my language of choice (I tried VB and it sucks ~~for me~~ in my opinion). If you want to use that VB though, you must just remove the last semicolon, but you probably know that already. :-)

Another thing that irritates me about my Atom feed, is that it doesn't use propper indentation. However, this doesn't make it incorrect, only irritating.

At least one thing: it validates!

Another thing I don't understand: Why doesn't it make use of of the atom:summary element? It could take the first paragraph out of the post and use it there, or it could even have (my preference) a separate box in the admin section where the author can type it in. Summaries are very important I think when they are written well.

Somehow, I thought earlier on that the entries in the feed didn't have any HTML in them. However, when checking again I see they do. A very nice feature I love! Especially since they use namespaces and not that ugly CDATA method that was used with RSS feeds.

But I have to say, in general the Atom feeds here at Blogger are not too bad. Many thanks go to the developers, because Atom rocks!!!

Modern Search Engine Optimization with Standards

One thing I realize is that there is huge potential for standards-based work in this area. Getting pages to validate is important, because you never know how quirky the search engine's parsers are. It is also very important to use semantic markup. If you look around, you will see that most (if not all) of the top search engine optimizers know about semantic markup, even though many of them are still not really standards-aware.

One of the thing we search engine optimizers always look at is the metadata. meta elements are not so important to Google anymore; however, they are still important to more primitive search engines. Therefore they must still be used. For Google, we (or at least I) make use of RDF.

Not many people really know to what extent Google supports RDF. However, we do know that the Googlebot is looking for index.rdf and atom.xml files.

I think the question is: Is Google looking for index.rdf to extract a feed out of it, or to extract metadata?

RDF is the future of metadata. However, since Google is ignoring meta elements inside a page, why would it be looking at external metadata?

But yet we don't really know Google is cooking behind the scenes. I hope there will be more light shed on it into the future.

However, if Google (or for that matter of fact, any search engine) would ever support RDF metadata, the first thing to be supported would surely be Dublin Core. I have set up a file using only a few Dublin Core elements for a client recently: The Model Shipyard RDF metadata.

Note that in the above example for The Model Shipyard, the file is sent as text/xml since I wasn't really supposed to use server-side scripting and I don't think the server knows about .rdf file extensions yet. Normally, you would use application/rdf+xml for that.

Another example I can show you is for the front page of my mother's website: Joke van Niekerk's Homepage Metadata in RDF. That is sent under the correct content type thanks to ASP.NET, and is also more extensive than the one for Shipyard.

vCard in RDF

I have recently discovered a way to represent a vCard in RDF. I don't really know if one can call this a specification or not, since it is only a W3C Note. However, for the purposes of this post I will refer to it as a specification.

This is probably one of the poorest specifications ever written. I ended up working from the vCard RDF Schema instead. I personally would have liked to see much more information inside that specification from the original vCard RFC. And another thing, that RFC contains much more than what has been included in vCard's RDF Schema mentioned above.

I have set up one example so far. I just hope it is correct, since the specification is so bad I don't really know. The example I created was for my mother: Joke van Niekerk's vCard in RDF.

Link Log - Post 1

My first link-log post. These are just weblogs I read regularly:

There are some other ones as well, but I'll make another post again later.

First post

Hello! This is officially the first post on my new weblog. There is still a lot of work that needs to be done here, but I'll handle it as it comes.

Now at least I can make my posts as long as I want, and I also have something "usefull" to point my URI to from my comments on other weblogs. It's better than having it point to something else which is still currently empty, and is in a language most people won't even understand. I still have some big plans for that site though.

I was very suprised how easy it was to sign up here on Blogger. They also have some nice templates, but I'll have to edit them quite a bit when I have time.

It's a pity that I can't get this site to validate though. Google is inserting some code at the top for the ads, which is fine, but it's breaking all possibilities of validation. I might send them an e-mail to complain, but it's a free service after all, so...

I'll probably get myself some (paid) hosting account one day. I'll see.

Anyway, for those that are wondering, I'll probably be posting all kinds of web design/development and programming-related stuff here when I have time. We'll have to see.

Thanks for reading my first post.

Main

Latest

Archives