- This topic has 24 replies, 11 voices, and was last updated 17 years, 1 month ago by
.
-
Posts
-
I know that MyEclipse has given its backing to Xfire. But I want to know if Xfire is really the best option to develop web services. I know Axis. What are the advantages from Xfire over Axis?
wilson,
I know that the team that decided on XFire took the evaluation seriously and looked at everything from bug reports to outstanding issues to community activity and support levels. I will ask them to weigh in with their decision if/when they get a chance. We are in the midst of QA right now and everyone is swamped so replies might not come immediately.Sorry guys, the WS team is very busy right now and haven’t had a chance to weigh in on this thread. I’ll pester them when things calm down.
I’ll add my two cents if I may.
First Axis is INCREDIBLY fussy when it comes to interfacing with .NET-based web services. I had a VERY hard time making code that would cleanly and consistenly call a web service that was written in C#. By the same token, the WebService consumers (that are based on Axis) in WebLogic 8.1 worked very well doing the same thing and were *much* easier to use. So it would appear that WebLogic makes substantial changes to the Axis code base to give it the increased ease of use for developers.
This said XFire has more library bloat compared to Axis and that *does* drive me a bit nuts. But XFire has also (in my initial tests) shown itself to be more robust in the same situations where Axis simply wouldn’t do the job (without modifying generated code).
So there you have my thoughts on the subject… Hope you find them helpful.
Anybody have any idea what Glassfish/Sun App Server is using to implement the new JEE 5 web services that Sun pushed really hard as having fantastic MS-integration?
FYI: here’s a stack comparison courtesy of the writers of XFire.. http://xfire.codehaus.org/Stack+Comparison
It seems that XFire is merging with Celtix and then becoming the part of Apache foundation. Maybe it is not the best thing for MyEclipse to be working on a tool that will become useless in about year or so. Read this post for more details: http://swik.net/Axis-vs-XFire
alukic,
I don’t follow the “useless” assertion… we have very strong ties to the XFire team, where ever the project is going with it’s stack we will likely tread as well.Alright, fair enough. I have been MyEclipse user for a while now, and it is my preferred platform. For that reason I would hate it to go down some path that may not be useful for majority of dev community. When I said useless, I meant that they will be changing their complete codebase, and it may take them a while to get the XFire stable enough for people to use it. Axis has already done this huge step so it is fairly stable. Hence, developing tools for XFire may become a challenging task. I appreciate the fact that you guys are working tightly together with XFire team, but when you say “likely” does that mean “definitely”, “possibly” or “some unknown degree of certainty”?
Don’t get me wrong, I am a big fan of MyEclipse and relly enjoy working with it. That is why I am bit concerned, that’s all. Best of luck.We certainly understand and will continue to deliver the best user-experience we can with the technology stacks we choose to build on. We won’t ship anything shoddy.
@alukic wrote:
It seems that XFire is merging with Celtix and then becoming the part of Apache foundation. Maybe it is not the best thing for MyEclipse to be working on a tool that will become useless in about year or so. Read this post for more details: http://swik.net/Axis-vs-XFire
First of all XFire won’t become useless so easy 🙂 XFire will be developed as long as needed ( as long users have some issue or improvement request ).
Yes, XFire is merging with Celtix, but you can think of it as of XFire 2.0, so any migration from XFire 1.x to XFire 2.x will be as easy as possible, so you can move to XFire 2.0 without any (or minimal ) modification to your code.I’m confused why everyone hasn’t embraced the JAX-WS for their web services. It is part of the Java EE 5 spec and therefore will be supported on every Java EE 5 Application Server. This seems like the starting place for web services. Adding other toolkits would be a bonus, but not a starting point. I have used the JAX-WS and it is extremely easy to use, well documented, and performed very well.
I’m sorry about the ties to the XFire codebase and team (I really liked that product before I started using the Java EE 5 Web Services), but those ties are leading you in the wrong direction. I love Eclipse and MyEclipse, and have converted over almost every company I’ve been at to use them. However, I currently cannot recommend them over Netbeans if the shop is working on Java EE 5 application servers.
I know the Eclipse WTP is looking at Axis2 and you guys are talking about XFire, and to me none of this makes sense. I am simply beside myself waiting for MyEclipse to catch up to netbeans when it comes to Java EE 5. Some one needs to become an expert at it to understand all of the benefits it provides so that they can be ported to Eclipse. I love Eclipse and MyEclipse, but Java EE 5 is the most important technology right now. BTW, I have just downloaded the 5.5M2 for 3.2.2 so I will be checking out the Java EE 5 support of MyEclipse today. I really hope it is better than what I’ve seen in the WTP, because even though it is claimed to be Java EE 5, it doesn’t even come close.
I would love to sell more MyEclipse licenses for you, and I look forward to working with your Java EE 5 updates!
Cheers,
GeoffYou forgot about one small,tiny problem, lot of ppl still use java 1.4 ( e.g because of commercial servers installed on production env ) and can’t move to java 5 so JAX-WS is useless for them ( e.g. for me ).
