tag:blogger.com,1999:blog-6406593750327945950.post1718141255510051649..comments2008-10-21T22:55:26.035-07:00Igor Minarhttp://www.blogger.com/profile/03520548417275543432noreply@blogger.comBlogger10125tag:blogger.com,1999:blog-6406593750327945950.post-33551449873523382792008-10-21T22:55:26.035-07:002008-10-21T22:55:26.035-07:00Charles, thanks for the comments.<BR/><BR/>My main point is that 3.0 is when you can make major changes that will break many legacy stuff and you'll get away with it.<BR/><BR/>I'm glad to hear about the code for processing javascript and css.<BR/><BR/>wrt "standards compliance" your business case can be decreasing the page size (and with it simplifying the DOM), which results in faster page loads and better javascript performance.Igor Minarhttp://www.blogger.com/profile/03520548417275543432noreply@blogger.comtag:blogger.com,1999:blog-6406593750327945950.post-73293914895052817652008-10-21T18:08:48.420-07:002008-10-21T18:08:48.420-07:00I'm pretty sure Confluence runs on current versions of Glassfish, but there's a pretty significant investment of effort standing between "we're pretty sure it works" and "we actively support it". We simply don't get enough demand for Glassfish support to justify that investment. (more info <A HREF="http://confluence.atlassian.com/display/DOC/Supported+Platforms+FAQ" REL="nofollow">Confluence supported platforms FAQ</A>)<BR/><BR/>The main focus of the Confluence 2.9/2.10/3.0 trilogy has been on user experience. To respond to Guy, most of the work for these releases has focused around improving existing features -- the basic UI, search, the rich text editor, the plugin system, performance and security -- rather than adding new stuff.<BR/><BR/>A lot of the suggestions on this list are pretty high up on my personal list of priorities: improving the database schema, the renderer, clustering, developer documentation, all things I agree we need to put time into over the next few releases.<BR/><BR/>Some we've already been working on: for example we have an implementation for auto-minifying and batching javascript and CSS resources that just didn't get done in time for the 2.10 feature freeze.<BR/><BR/>Other suggestions are a little harder to make a business case for. It makes sense for us to follow our existing plan that when we work on a piece of the system, we improve the standards compliance of the HTML code as we work on it. It's a lot harder to make a case for overhauling all 535 velocity templates in one fell swoop, just so we can say in the release notes "Confluence: now does exactly what it did before, but without offending your office web standards guy"<BR/><BR/>Similarly, it's hard to make a business case for cleaning up the URL space for a single release, but there are things we can do to improve it incrementally.<BR/><BR/>(The "/display" prefix is there because if we didn't have the prefix, the app's root URI namespace would clash with space keys. We'd have to arbitrarily disallow keys that clashed with reserved symbols ("users", "images" etc) and constantly worry that anything new we put into the namespace was going to clash with some existing user's space key)Charles Millerhttp://www.blogger.com/profile/04469280378310928436noreply@blogger.comtag:blogger.com,1999:blog-6406593750327945950.post-67032314931324554862008-10-15T11:34:10.038-07:002008-10-15T11:34:10.038-07:00@eduard/o You are right. Glassfish is <A HREF="http://confluence.atlassian.com/display/DOC/List+Of+Supported+Application+Servers" REL="nofollow">*not* supported</A>. The linked issues are reported against Glassfish v1 and SJSAS8.x. I bet that Glassfish v2 works just fine. I recall seeing some blogs about people who run confluence on glassfish already.<BR/><BR/>@guy confluence markup has it's own cons and pros. The point is that users (or admins) should have an option to select the markup the want to use. I'm not saying that confluence markup should be replaced with Creole markup, I think that it should be possible to make them run side by side.<BR/><BR/>@Peter Raymond :-) I feel with you. Clustering is *very* important for us too to achieve high availability, but with Confluence it's been a painful experience so far.Igor Minarhttp://www.blogger.com/profile/03520548417275543432noreply@blogger.comtag:blogger.com,1999:blog-6406593750327945950.post-84285740333515850362008-10-14T11:20:06.716-07:002008-10-14T11:20:06.716-07:00Tossing another "agree wholeheartedly" on the pile. Our Enterprise requires us to run a cluster even though Confluence stability dropped through the floor after doing so. Our Enterprise Application support people looked at how clustering is implemented and alternated between laughing and cursing...Peter Raymondhttp://www.blogger.com/profile/05580686826671545620noreply@blogger.comtag:blogger.com,1999:blog-6406593750327945950.post-16920224332529024172008-10-14T08:38:50.061-07:002008-10-14T08:38:50.061-07:00Hi Igor, I wholeheartedly agree. Confluence is a very good product, but could definitely be better and I think 3.0 is the time to do it. I'll also try and throw some suggestions on here soon.David Hergerthttp://blog.hergert.usnoreply@blogger.comtag:blogger.com,1999:blog-6406593750327945950.post-68075978190425216782008-10-14T04:15:33.175-07:002008-10-14T04:15:33.175-07:00thank u r information <BR/><BR/>it very usefulMatthewhttp://www.creativewebsitedesigner.comnoreply@blogger.comtag:blogger.com,1999:blog-6406593750327945950.post-57533668966854309252008-10-14T01:57:00.536-07:002008-10-14T01:57:00.536-07:00Hi,<BR/><BR/>Looks like you've run in to many of the pain barriers that Adaptavist encounter on a daily basis. Totally agree that 3.0 needs to be a major update with significant improvements to _existing_ functionality (not just a bunch of new features for marketing purposes).<BR/><BR/>As for Creole - it's 'effing hideous. Since when does "==" instinctively mean "heading one"? I'd rather climb the Eiffel Tower with nothing but the suction power of my own lips, or have my kneecaps stapled together, than have to work with nasty Creole-like syntax.<BR/><BR/>And as for making HTML output better that's going to be a very hard task indeed - first, we have the joy of widespread MSIE6 infestations on corporate networks. IE6 can barely render HTML, yet alone XHTML (which even modern browsers like Firefox are only just getting to grips with). Then there's all the plugins out there that are written in Java and, uhm, well, Java developers are awesome at server-side code. But not that hot when it comes to the browser... ;)Guyhttp://www.blogger.com/profile/07474655406155302803noreply@blogger.comtag:blogger.com,1999:blog-6406593750327945950.post-41635630866144209982008-10-13T21:00:39.943-07:002008-10-13T21:00:39.943-07:00I thought that glassfish was already supported, no?Igor Minarhttp://www.blogger.com/profile/03520548417275543432noreply@blogger.comtag:blogger.com,1999:blog-6406593750327945950.post-76383367984647079402008-10-13T20:16:03.808-07:002008-10-13T20:16:03.808-07:00Plus support for GlassFish; please? - eduard/opelegrihttp://www.blogger.com/profile/16469447400885151834noreply@blogger.comtag:blogger.com,1999:blog-6406593750327945950.post-72666266863681607752008-10-13T19:17:04.519-07:002008-10-13T19:17:04.519-07:00I like where you're going with this; going to swing back around soon and put my two cents in. At the moment, however, I am buried in code. Good read. I at least wanted to say that for now.Cody Burlesonhttp://www.burlesontech.comnoreply@blogger.com