On Sat, 07 Aug 2004 13:43:44 -0400, Matt Feifarek wrote: > It is CERTAINLY not my point of view that these other kits have no > value... it's merely my opinion that Webware/webkit should be pared > down to the essentials, with these other kits being "plug-ins".
It's difficult to say what is essential. That differs from person to person both in terms of the kits and also what each kit contains. > And/or, we could distribute all of the useful "kits" under the > Webware umbrella "brand" and do releases of each. If we use > distutils eventually, this will be transparently easy for users to > use and install. Is distutils easy? I don't know how to get it to do any of the following things: - list what has been installed including version number - remove a package - upgrade a package - downgrade a package (e.g., rollback to prior version) - install a package by name, downloading it if necessary and any of its dependencies - show me the diffs between the current installed files and the original If I'm missing something let me know, but this is why I still use CVS for some installations. "cvs diff" and cvs tags give me at least some of those abilities. And bundling kits reduces my installation workload as a consumer and reduces my packaging workload as a vendor. I use other popular libraries that install more libraries than I use: * egenix BASE * win32 extensions * wxPython * Python's standard library In all cases, I get much more functionality than I asked for, but that's okay. The unused libraries sit on my hard drive consuming space, but: 1. They never pester me. 2. Maybe someday I'll use them. 3. For all I know, some other package I use that "requires egenix BASE" is using mxStack. I could go to vendors of those packages and ask them to remove various parts, but they would probably say: 1. Your minimal set is not someone else's minimal set. 2. Breaking them apart increases our release work. ...which is what I say to breaking up Webware. Besides, WebKit (and maybe PSP?) is the only Webware component that is hard to release. The others have plenty of unit or regression tests and can be released on very short notice. One final note: One reason why Webware is released infrequently (besides the fact that I have no free time any more) is that when people to go to release they throw in a bunch of requirements that aren't really required. There hasn't been a release in a long time. If we simply shipped what we currently have in its current form as a 0.9, that would certainly be of value to the community. Now back to $contracting.... -Chuck ------------------------------------------------------- This SF.Net email is sponsored by OSTG. Have you noticed the changes on Linux.com, ITManagersJournal and NewsForge in the past few weeks? Now, one more big change to announce. We are now OSTG- Open Source Technology Group. Come see the changes on the new OSTG site. www.ostg.com _______________________________________________ Webware-discuss mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/webware-discuss
