Bill de hÓra: JORAM in particular is a quality JMS implementation.
Yup. Wellstorm uses JORAM internally and it has proven solid. (The ObjectWeb logger did hijack our logging, an issue we still need to work out). I really haven't caught any JORAM errors, so it meets the ultimate test of infrastructure: It's just there.
If you can't say anything else for Java, you can say it has a great choice among infrastructure packages. Last weekend I added a scripting capability using Rhino. Got the idea at lunch Friday. Took an hour to get something working, the rest of the time framing a decent web UI. Demoed it today.
Hugh Winkler holding forth on computing and the Web
Monday, January 23, 2006
Encoding the XML infoset in HTML forms
Every month or two I find myself remarking on the awkwardness of constructing XML on the client side. Are HTML clients forever condemned to using XMLHttpRequest? Must we await widespread support for Xforms?
In the spirit of The Simplest Thing That Could Possibly Work: Why don't we just standardize encoding the XML infoset in HTML forms?
Here's how you could POST an Atom entry (compare to example in draft 07):
That form encodes as
The pattern is: Name your HTML form elements using XPath syntax.
(What if all Atom powered blogs supported POSTing of form data in addition to the Atom syndication format? I guess I could construct a website that exposes a form like this, and permits you to direct your entry to any blog service. Not sure how useful that would be, but it doesn't seem any less useful than enabling a rich client do it ;) ).
I'm just reiterating the point I made here some time ago: form data is a perfectly good hypermedia representation, and it has a ton of support already in place.
In the spirit of The Simplest Thing That Could Possibly Work: Why don't we just standardize encoding the XML infoset in HTML forms?
Here's how you could POST an Atom entry (compare to example in draft 07):
<form action="http://localhost/foo" method="POST">
<input type="hidden" name="/entry/@xmlns" value="http://www.w3.org/2005/Atom"/>
Title: <input type="text" name="/entry/title">
Link: <input type="text" name="/entry/link/@href"/>
Id: <input type="text" name="/entry/id"/>
Updated: <input type="text" name="/entry/updated">
Summary: <input type="text" name="/entry/summary">
<input type="submit" name="submit" value="Submit">
</form>
That form encodes as
%2Fentry%2F%40xmlns=http%3A%2F%2Fwww.w3.org%2F2005%2FAtom&%2Fentry%2Ftitle=Atom-Powered+Robots+Run+Amok&%2Fentry%2Flink%2F%40href=http%3A%2F%2Fexample.org%2F2003%2F12%2F13%2Fatom03&%2Fentry%2Fid=urn%3Auuid%3A1225c695-cfb8-4ebb-aaaa-80da344efa6a&%2Fentry%2Fupdated=2003-12-13T18%3A30%3A02Z&%2Fentry%2Fsummary=Some+text.&submit=Submit
The pattern is: Name your HTML form elements using XPath syntax.
(What if all Atom powered blogs supported POSTing of form data in addition to the Atom syndication format? I guess I could construct a website that exposes a form like this, and permits you to direct your entry to any blog service. Not sure how useful that would be, but it doesn't seem any less useful than enabling a rich client do it ;) ).
I'm just reiterating the point I made here some time ago: form data is a perfectly good hypermedia representation, and it has a ton of support already in place.
Subscribe to:
Posts (Atom)