DWR version 1.0 released
java.net on DWR
java.net have just published a step-by-step tutorial on writing a web chat app using DWR.
The final result is a web-chat app, in about 100 lines of code - client and server. The point being how easy it is to get started with DWR.
We could have gone into detail about using Scriptaculous for fancy effects, or created more server side features like one-to-one messages and so on. The point however is to demonstrate how easily DWR can take data in Java on the server and transparently turn it into Javascript arrays and lists and then put that data in the browser.
The strange things people look for on web-sites
I've just updated the DWR website to use Drupal rather than SnipSnap. And in doing so quite a few of the URLs have changed, so I waved my Apache/mod_alias wand around to point people where they should be going. I've had the naive aim that I should be able to nearly get rid of 404s.
I've been really surprised by the bizarre things people look for on my web site. For example:
- robots1.txt: Is that just a typo? Or is there something I'm missing. All the hits are from AOL users. But that could be a coincidence.
- DWR实时系统.files/drupal.css: I guess that's a typo, but the drupal style sheet isn't related to DWR in any way. So I don't know how you would come to make such a typo.
- siteinfo.xml: I guess someone is hoping there is a hidden Eclipse plug-in for DWR and are trying to find it.
- DWR%20-%20Overview%20Getahead_archivos/style.css: There is a referrer, but not with links to that file.
- There are a lot of hits for files that exist, for example intro1.jpg (one of the pics on the front page), but with paths that are just wrong. Why would (quite a few) people be looking for files that they know to exist somewhere in other random places?
Has anyone else wondered where all this wierd stuff is coming from or found bizarre URLs?
Update: Some of these are tools, and some just have to be human - the Eclipse one for example, but I've just found a new one that is a bot. What seems like a broken Safari. Get this:
- /dwr/atom/feed followed by /dwr/atom/atom/feed followed by /dwr/atom/atom/atom/feed followed by ... he gave up at 16 atoms!: The user-agent is 'AppleSyndication/38' which it appears is how Safari's RSS support manifests itself.
Further Update: I don't think this Yahoo FAQ explains any of the URLs above, but it does explain a valid reason why some bots from search engines go looking for semi-random URLs.
DWR version 1.0RC3 Released
DWR version 1.0 Release Candidate 3 has been released. DWR (Direct Web Remoting) is easy Ajax for Java.
The download contains a number of small bug fixes. The biggest change since the last announced release is permissions to the BeanConverter to restrict the properties that are converted.
Please help and test it out in case we need to make any last changes before we release 1.0 final.
Canvas element for IE
<canvas .../> is an HTML element that first popped up in Safari, where it helped Apple create Dashboard.
It is being standardized by What-WG
and added into Mozilla 1.5. I guess that Konqueror may be able to get support easily from Safari, it is on the list at least.
I've thought a number of times that this is all irrelevant for the next 5 years at least. Currently there is no <canvas .../> in IE, and I can't see Microsoft adding one into IE7, so we'd need to wait for IE8, or for IE to fall to less than 10% market-share. Even at the current 1% a month we've got another 5/6 years to go and I'm sure IE7 will slow down the decline.
So while <canvas .../> is neat I didn't plan on adding it into my list of things to investigate, until now.
I stumbled across the magic incantation: 369303C2-D7AC-11D0-89D5-00A0C90833E6 which happens to be the GUID for DirectAnimation. DirectAnimation is part of ActiveX, but so is XML.HTTP which is the foundation of Ajax in I.E.
You can see a demo of DirectAnimation - I.E. only, obviously. And it is described by WebDevelopersJournal if you want to know more.
So here is the challenge: Write some Javascript that detects lack of support for canvas, and support for DirectAnimation. It then replaces the <canvas .../> element with an <object .../> element which invokes DirectAnimation. The Javascript will then have some wrappers that convert the <canvas .../> Javascript function into DirectAnimation calls.
I have no idea if it is possible, but it sounds like it could be very interesting if it works.
Update: Emil A Eklund has implemented the Canvas Tag for IE using VML