Comments on: How buggy is your SOAP stack http://www.dankulp.com/blog/?p=137 Dan's random thoughts on the world Fri, 27 Aug 2010 10:34:45 +0000 hourly 1 http://wordpress.org/?v=3.0.1 By: easternwahoo http://www.dankulp.com/blog/?p=137&cpage=1#comment-16335 easternwahoo Wed, 20 Jan 2010 15:38:41 +0000 http://www.dankulp.com/blog/?p=137#comment-16335 Numbers aside, I have to agree that "CXF just “works”. It does what it says it will do relatively easily". I recently evaluated CXF against Axis2 and the Sun Reference implementation, and chose CXF because, well, it does just work! Numbers aside, I have to agree that “CXF just “works”. It does what it says it will do relatively easily”. I recently evaluated CXF against Axis2 and the Sun Reference implementation, and chose CXF because, well, it does just work!

]]> By: Claus Ibsen http://www.dankulp.com/blog/?p=137&cpage=1#comment-12457 Claus Ibsen Fri, 06 Nov 2009 08:24:11 +0000 http://www.dankulp.com/blog/?p=137#comment-12457 I got inspired by Dans blog and wrote a blog entry about the state of Apache Camel. Unlike Dan I only look at Apache Camel and do not compare against other similar projects. I would love others to do that investigation as I am affiliated with Apache Camel http://davsclaus.blogspot.com/2009/11/how-buggy-is-your-integration-stack.html I got inspired by Dans blog and wrote a blog entry about the state of Apache Camel.

Unlike Dan I only look at Apache Camel and do not compare against other similar projects. I would love others to do that investigation as I am affiliated with Apache Camel

http://davsclaus.blogspot.com/2009/11/how-buggy-is-your-integration-stack.html

]]> By: SteveL http://www.dankulp.com/blog/?p=137&cpage=1#comment-12455 SteveL Fri, 06 Nov 2009 08:11:28 +0000 http://www.dankulp.com/blog/?p=137#comment-12455 This is actually quite interesting. I'd discount the criticality setting, as that may just be a relic of the bug trackers GUI. that leaves numbers filed, which as others point out, could be a function of the number of users. 1. Perhaps its the ratio of bug reports filed/fixed that should be looked at, just as that is how you assess stability of a project. Is the project getting buggier or better? 2. Same for bug longevity 3. I'd also consider grouping by SOAP feature; anything in Java to WSDL could be discounted on the grounds of Moral Wrongness. This is actually quite interesting. I’d discount the criticality setting, as that may just be a relic of the bug trackers GUI. that leaves numbers filed, which as others point out, could be a function of the number of users.

1. Perhaps its the ratio of bug reports filed/fixed that should be looked at, just as that is how you assess stability of a project. Is the project getting buggier or better?

2. Same for bug longevity

3. I’d also consider grouping by SOAP feature; anything in Java to WSDL could be discounted on the grounds of Moral Wrongness.

]]> By: Claus Ibsen http://www.dankulp.com/blog/?p=137&cpage=1#comment-12451 Claus Ibsen Fri, 06 Nov 2009 07:29:59 +0000 http://www.dankulp.com/blog/?p=137#comment-12451 Of course bug counts matter, but its just one metric out of many. Its also a testimony of the committers if they oversee the issue tracker and react to those issues reported. If the bugs are not reported with enough details they should be asked to provide those details or the bug can be closed with incomplete etc. Remember this is bugs, not requests for new features, improvements etc. Its bugs! So having bugs open for a long time is because the maintainers do not oversee their project well enough. Dan great post, got me inspired to try look at the state of Apache Camel . Of course bug counts matter, but its just one metric out of many.

Its also a testimony of the committers if they oversee the issue tracker and react to those issues reported. If the bugs are not reported with enough details they should be asked to provide those details or the bug can be closed with incomplete etc.

Remember this is bugs, not requests for new features, improvements etc. Its bugs!

So having bugs open for a long time is because the maintainers do not oversee their project well enough.

Dan great post, got me inspired to try look at the state of Apache Camel .

]]> By: Yaytay http://www.dankulp.com/blog/?p=137&cpage=1#comment-12446 Yaytay Fri, 06 Nov 2009 06:37:49 +0000 http://www.dankulp.com/blog/?p=137#comment-12446 Dan, The text of your article is all good, but the numbers are meaningless. When comparing bug counts you need to also consider the number of users and the ease with which bugs are filed. Without that information bare bug counts have no context. Dan,

The text of your article is all good, but the numbers are meaningless.
When comparing bug counts you need to also consider the number of users and the ease with which bugs are filed.
Without that information bare bug counts have no context.

]]> By: Eran Chinthaka http://www.dankulp.com/blog/?p=137&cpage=1#comment-12422 Eran Chinthaka Fri, 06 Nov 2009 02:25:55 +0000 http://www.dankulp.com/blog/?p=137#comment-12422 Dan, Thanks for the very interesting comparison of the projects. Also I respect and value your effort on giving constructive criticism on Axis2 and we always welcome feedback. But in general I'm not a believer in using bug counts as a way of comparing different projects or to measure the quality of software. There are few reasons behind my belief. A software might have 0 bugs for 10 years because there aren't enough users using this software or the users care enough to post a bug. On the other hand a project having lots of bugs doesn't mean it doesn't work. May be there are thousands of users who really use it and care about improving the project. There are more reasons that can be found here : http://www.sqablogs.com/jstrazzere/267/Misuse+and+Abuse+of+Bug+Counts.html. We all know very well that CXF is good is some cases and Axis2 is good in some other cases. But we can never say this is bad and the other is good. Once again thanks for this post and please continue helping us with these kind of feedback. We really appreciate it. Dan,

Thanks for the very interesting comparison of the projects. Also I respect and value your effort on giving constructive criticism on Axis2 and we always welcome feedback. But in general I’m not a believer in using bug counts as a way of comparing different projects or to measure the quality of software.

There are few reasons behind my belief.

A software might have 0 bugs for 10 years because there aren’t enough users using this software or the users care enough to post a bug. On the other hand a project having lots of bugs doesn’t mean it doesn’t work. May be there are thousands of users who really use it and care about improving the project.

There are more reasons that can be found here : http://www.sqablogs.com/jstrazzere/267/Misuse+and+Abuse+of+Bug+Counts.html.

We all know very well that CXF is good is some cases and Axis2 is good in some other cases. But we can never say this is bad and the other is good.

Once again thanks for this post and please continue helping us with these kind of feedback. We really appreciate it.

]]>