I had a need in one of my projects to query for different XML schema
documents from my RESTful service. Looking around the internet it
seemed that the de facto way of doing this would be:
application/vnd.myformat+xml
Or, just register "myformat" with IANA and have:
application/myformat+xml
One huge downside of this though is that my browser does not know how to
render application/myformat+xml. So, my question to this list is, why
aren't MIME properties used to identify XML schema? For example:
I tried this approach with Firefox and Safari and both browser ignore
the attribute and render the document as XML.
Would there be any reason NOT to use and promote this approach other
than it goes against current de facto convention? I blogged about this
a few months ago and have yet to receive a logical reason why this
approach isn't superior:
I had a need in one of my projects to query for different XML schema documents from my RESTful service. Looking around the internet it seemed that the de...
There was a whole discussion and even a very detailed proposal for something like this a while back. Here's what I wrote back then [1]. Interestingly, after a...
Any reasons why is having a zillion types is a bad thing? Media-"type" is a type, and generic types don't convey much to clients. Subbu ... http://subbu.org...
Hmm. Let's say, if we fix the interface to be uniform, and fix all media types to be, say a dozen, where would that take us? By the way, the point I am making...
... No, I meant that having lots of media types is a bad idea for interoperability for the same reason having lots of interfaces is a bad idea. But to Bill's...
... Yes, interop is the key. But the downside is that most people interpret this argument to mean that they should limit themselves to "standard" media types...
... Registration of vnd.redhat.blahblah will not allow Firefox and Safari to render my documents though :) I see your point of why namespaces don't replace...
Not sure what you mean. You mean Peter's suggestion? Of looking at the User-Agent header? Not very feasible to add to user code (and ugly to add to user...
When you find that the client does not support application/ vnd.redhat.blahblah+xml, redirect the client to some blahblahdocument.xml. This may work since when...
Browser rendering issues usually happen when the server doesn't properly handle content-neg; i.e. resources are defined for limited (only one) representation...
... Doesn't really help to transform things to HTML when you're trying to debug an XML problem. Customer Support: Hey Bill, this XML message won't post to the...
Bill: i provide a simple 'viewer' in HTML/Ajax that makes it possible to see all the various (text-based) messages no matter the media-type. very simple to...
... Based on the long discussion on this topic i think it is clear that your only real choice is to learn to love `curl`, `xml_pp` and `less`. ... That...
... Or use CLI curl. I'm losing track of the number of ways to run GNU/FOSS on Windows (used to just be cygwin), I only came across uwin today: ...
Eric J. Bowman
eric@...
Dec 12, 2008 5:06 am
Eric: Yep, lots of ways to go. I've even run WFetch.exe via WINE over Ubuntu (no SSL, but it covers most cases). The point, I think, is that the Web browser is...
... Yup, I got that point, I just like mentioning that every GUI or CLI tool discussed regarding REST development is available on Windows... ...
Eric J. Bowman
eric@...
Dec 12, 2008 8:06 am
... Turns out it was written by Dr. David Korn: " The best advice I can give is to not use Windows unless you have to, or if you have to, use a UNIX operating...
Eric J. Bowman
eric@...
Dec 17, 2008 8:45 pm
... Sigh... ... -- Bill Burke JBoss, a division of Red Hat http://bill.burkecentral.com...
Hi Stefan, Stefan Tilkov wrote: ... I use the HTTP Link: header [1] to do this kind of thing - defining a link between the representation returned and its XML...
Off-the-cuff, I like Bill's proposal better, b/c 1) we already know it's XML, so anything that knows how to render XML is happy 2) we already have an XML...
(1) The proposal was not XML specific. (2) Absolutely wrong :). Every application defines their own namespaces .. it may be possible use a uniform interface to...