vendredi 11 mars 2005

Infrae answers: XML will be available for Zope 3 In response to the previous post, asking about Zope 3 and XML, Martijn gives an answer: "The answer is of course, yes, it should, and, what's more, yes it will, and Infrae will be one party working on it. In fact, we've already worked on it." Martijn's post is indeed great news. Combined with Infrae's co-announcement today of collaboration with Nuxeo, this will mean Zope 3 CMS's can anticipate having a good XML infrastructure. I've been reviewing, with Martijn's guidance, his ongoing efforts for adding this technology into Silva, both the software linked to above and lxml. It's already impressive, and as the first releases come out, I already have a loooooong list of ways I plan to start using this facility in Silva. For me, his work in Silva has the "stop everything you're doing, clear out a day, and take a look" aspect. Zope 3 and XML, in Zope 2, via Silva. Thanks, Infrae. I'm a happy camper. 5:26:13 PM   comment []

Pareto and POV join ZEA With the addition of Pareto, ZEA gains a Dutch company with long-time Zope experience, but also a larger profile in project management and sales capability. POV, located in Vilnius, Lithuania, has a growing reputation as a strong group of architects specializing in Zope 3. 5:10:34 PM   comment []

"Office Linebacker" videos and Zope 3 popes Heimo Laukkanen sent a link to some hilarious mockumentary videos about "Terry Tate: Office Linebacker". This really is a "stop what you're doing and watch it" kind of thing. Limi and I were joking in IRC that open source should be like that. Since I posted recently about Zope 3's pope, I had a mental image of Jim tackling someone, yelling: "Don't hardwire that crap, you gotta adapt! Don't be a complainer, make a custom container!" 2:51:13 PM   comment []

Should Zope 3 support XML? Update: Martijn pointed out that I got the wrong story. I was told that the Zope 3 released packages would only contain Python and Zope-repo code. Martijn said that it isn't out of the question for lxml and libxml2 to be in Zope 3. If that's true, then I'm thrilled to say, I was wrong! This is an open question. I know for the Zope 3 developers, XML isn't a high priority. I wonder, though, for Zope 3's customers, how they feel. Note: I'm not asking if Zope 3 should convert to XML as a first-class citizen. I'm just asking if Zope 3 should upgrade XML to second-class status, making the basics possible for those that want it. Right now, Zope 3 doesn't really have a quality XML facility. (IMO, minidom does not count as a serious XML parser.) Also IMO, an architectural re-write that anticipates the next 3-5 years of lifetime in the market, should have a legitimate XML parser available. One resolution is to say, "Zope 3 is extensible, people that want XML can just add it." However, saying "You can add it yourself" is the same as saying "We don't think it is legitimate." Fortunately, Python has a good story for XML. There are a number of mature choices, some lightweight (ElementTree) and some more industrial strength (libxml2 + Infraes's excellent lxml work). And Zope 3, as one of its top positioning statements, says it wants to get closer to Python and the Python community. So there's hope. Unfortunately, the ZPL is a barrier to this. Zope 3 can't use Python software under the Python license(s). It can only use ZPL. This is a barrier, no ifs-ands-or-buts about it, and to some degree, it blunts the effort Zope 3 has made in being more Pythonic. It's possible that we can convince a serious XML engine to relicense under the ZPL, but I'd say the odds are against it. I also doubt that the Zope community can produce a new, ZPL'd, quality XML engine. IMO, Zope 3 shouldn't go it alone. We should find ways to remove barriers between Zope and Python software. I realize there are issues, but they might be solvable, given sufficient desire to do so. So here's my vote. First, it's not wise to launch a new application server in 2006 (when Zope 3 ships) that only includes a toy XML parser. Bundle an ElementTree engine and make lxml an option w/ a similar interface. Second, Zope 3 should use Pycon as a forum to solve the Zope/Python IP barrier. 9:48:20 AM   comment []