July 19, 2004
Congratulations to the JBoss guys for the J2EE certification. That's a Very Good Thing to see that Open Source projects get certified. with JOnAS and Geronimo certifications coming soon, the J2EE market could be shaken up.
That's great because it'll bring innovation (EJB3) from both the Open Source and the proprietary stack (BEA, IBM, Oracle) under the stewardship of the JCP. It'll also increase choice which is also a Very Very Good Thing.
When I see people on TheServerSide complaining that there is too much choice in Java (Hibernate or JDO, Ant or Maven, JBoss or Weblogic, ...), it reminds me of my use of Emacs and Vi. I'm very happy to be able to use one tool or the other for a given task and not to be bound to an inadequate tool.
July 16, 2004
The new WYSIWY – M – G editor of Blogger is a hell to use when posting html + JavaScript code (like in my previous blog). It also doesn't work well on Firefox and doesn't work at all on Safari.
What's worse is that it change the HTML I type when I edit again the same post. That'd be great if the HTML editor could be set by default.
[update] The new version of the editor pushed by Jason is working fine and keeps the HTML version by default. Thanks Jason!
July 16, 2004
[update: the bookmarklets were broken for a long time but they now work again on Firefox 3 and Safari 3]
In a comment on his blog, I proposed to Cedric a simple solution to spilt a HTML page in two within the browser. It's just a one line JavaScript but it can be handy to use it,
split document in new window -> drag to your bookmark folder
When you want to split a html document, click on the bookmark and you're done (it works on Safari and Firefox)
Cedric then suggested to split the document in the current window. So here come two new bookmarklets:
July 15, 2004
Following a blog of Tim Bray, I downloaded Sage, a Firefox news reader.
To evaulate it, I tried to import my bloglines subscibtions under the OPML format but it failed.
So I did the 101 test: I tried to import an OPML file that was exported by Sage and it also failed...
At least, the good old drag'n'drop method is working.
July 13, 2004
Maven 1.0 has been released.
I haven't tried Maven since the RC1 (or RC2, it was months ago...) but I was really underwhelmed by it. To be fair, I'll give Maven 1.0 a try when one of the projects I follow move to it. But I'm not sold on the usefulness (or the relevance) of Maven compared to Ant 1.6.
July 13, 2004
Question: How do you get procedural programming and #include in Java?
Answer: You just have to create a JSP file of 2000 lines containing only Java functions and include it in every other JSP page.
This is MVC model 1 at its ugliest...
July 12, 2004
Given all the interviews, articles, blogs about JUnit (I won't even try to link to them...), I think they should replace the quote by Martin Fowler on the JUnit Homepage
"Never in the field of software development was so much owed by so many to so few lines of code".
by
"Never in the field of software development was so much talked by so many about so few lines of code".
Yes, I use JUnit heavily but I don't think it's the main key aspect to good software development as so many would like us to believe.
July 8, 2004
When I see the similarities between Gmail and the beta of Google Groups 2, I wonder how long we will have to wait before being able to read and post to groups from Gmail...
July 8, 2004
I'm reading the early draft of EJB 3.0 bits by bits and was quite amused by the "universal context model". Doesn't it sound like the (in)famous JBoss' "unified class loader"?
I wonder if the expert group will follow the same version scheme! ;-)
July 2, 2004
Very interesting blog from Dion, Scared of annotation hell.
The EJB specification made a good job to separate the different roles (bean developer, assembler, deployer and I forgot the others). One of the interesting side effect was that only a subset of the deployment descriptor was relevant for a given role.
However, it didn't work out as expected because of the lack of smart applications to read/write the deployment descriptor.
One classic example is transaction demarcation: only the bean developper can set the transaction demarcation attribute transaction-type and it can't be changed later on (if the bean use bean-managed transaction demarcation, you can't make it container-managed).
I think that the EJB 3.0 Expert Group needs to think about that issue and maybe keep (optional) deployment descriptor outside the bean (for information such as table name) and use annotations for information that can't be modified without affecting subsequently the EJBs (such as transaction demarcation).