Apache Tomcat - The Shine is Gone
After using Tomcat as the preferred Java-Servlet-Container for the last couple of years, I have been using Version 5.0.x built-in clustering support for only a few months now.
Single server servlet programming and deploying was almost always Tomcat independent; one hardly needed to look at Tomcat's source-code and while Tomcat's documentation always sucked, any Servlet.jar file's java-doc was good enough to write programs that would work well inside the container.
If I needed to find out about certain tags in one of the descriptors like server.xml, web.xml, context.xml, etc. there was always BEA's Web site to look for a helping hand.
With clustering all that changed
Documentation about Tomcat-Clustering is almost none-existent on the Apache site. This "how-to" is as detailed as it gets there. But no problem, we are talking about open-source, right? All we have to do is to look at the code and we know what's going on, right?
Well, not exactly. The code in the clustering package is a mess. No documentation in the source files, not a single javadoc tag to be found.
Now, there are open-source projects out there, containing well-written code and documentation; Jason Hunter's JDOM or Marc De Scheemaecker Nano-XML come to mind. I had expected so much more when it comes to an established project like Tomcat, hosted on the prestigious Apache site.
Single server servlet programming and deploying was almost always Tomcat independent; one hardly needed to look at Tomcat's source-code and while Tomcat's documentation always sucked, any Servlet.jar file's java-doc was good enough to write programs that would work well inside the container.
If I needed to find out about certain tags in one of the descriptors like server.xml, web.xml, context.xml, etc. there was always BEA's Web site to look for a helping hand.
With clustering all that changed
Documentation about Tomcat-Clustering is almost none-existent on the Apache site. This "how-to" is as detailed as it gets there. But no problem, we are talking about open-source, right? All we have to do is to look at the code and we know what's going on, right?
Well, not exactly. The code in the clustering package is a mess. No documentation in the source files, not a single javadoc tag to be found.
Now, there are open-source projects out there, containing well-written code and documentation; Jason Hunter's JDOM or Marc De Scheemaecker Nano-XML come to mind. I had expected so much more when it comes to an established project like Tomcat, hosted on the prestigious Apache site.
Up to a couple of days ago - before I had read in the Allan Holub's article in the January 2005 issue of the Software Developer Times - I thought it's just me, not being sharp enough to understand the Tomcat code, which for others must be so obvious that it doesn't need documentation - and the code that I considered a mess, in fact was brilliantly written and structured.
Allan Holub, whose opinion I value much for quite some time now, however concludes in his article:
Allan Holub, whose opinion I value much for quite some time now, however concludes in his article:
- Tomcat's documentation is virtually unusable. It's a hodgepodge of inadequate .html files.
- The code is a mess. It's poorly structured, poorly documented and poorly written.
- You have to do a lot of work to use an open-source product like Tomcat.
- You have to test new versions rigorously.
Carlsbad, Batiquitos Lagune
Panasonic DMC FZ-15 • 1/640 @ f/4.6 ISO 64 • Jan.16.2005 14:27
Panasonic DMC FZ-15 • 1/640 @ f/4.6 ISO 64 • Jan.16.2005 14:27