Hi Murray,
In speaking about priorities, you hit the nail on the head. Both Kimbro and I had priorities, paying the bills being the most important.
The company had died, and as a last attempt to keep the project alive, we donated it to the ASF, after that, for many months, both of us continued to try to maintain the project, living on our savings, and trying to scrape money together however we could, simply because we wanted to see the project succeed. But when you're out of cash, and starting to realize you may go hungry soon, finding some income becomes a priority. Personally, I was offered a job with BEA, which is where I worked for about 6 or so months until it was proposed to me that dbXML could be brought back to life, but only if it were to be a commercial product. My love for the product forced me to quit BEA, and take a substantial pay cut to return to dbXML. In neither situation did I have the bandwidth to dedicate to Xindice.
And then there was the whole divorce thing.
It has been my intention since returning to dbXML to ultimately either (a) donate major portions back to the ASF, or (b) open source the whole thing. The fact that the two projects have diverged substantially led me to the latter option.
Personally, I care about both projects, and it would be nice to see Xindice succeed. Obviously there is an ego thing involved if it does succeed, but I'm just as bandwidth constrained now as I have been for the past year, and my focus has to be on the company's product, rather than other major projects.
-- Tom Bradford - http://www.tbradford.org/ CTO - The dbXML Group - http://www.dbxml.com/ Project Labrador - http://www.dbxml.com/labrador/
On Dec 3, 2003, at 4:03 PM, Murray Altheim wrote:
Tom Bradford wrote:dbXML, which served as the original codebase for Xindice 1.0, has recently been re-released as a GPLed project. For over a year, dbXML was developed as a proprietary product and licensed commercially. In the process, it has benefited from real world customer feedback, and has become rather robust. It is very different from Xindice, and in a sense was nearly completely rewritten from scratch after version 1.0.
It may be worth checking out if you can stomach the GPL.
Now onto my personal opinion about Xindice being stalled and retired. The reason that I decided to open source dbXML again was mainly because Xindice was stalled, and I didn't see that situation changing significantly in any reasonable amount of time.
The biggest problem I see with it is that nobody has stepped up to jump into the guts of the system and try to stabilize or build out the database core. Unless that's done, the project is very well doomed. Unfortunately, hard core database skills are very hard to come by, and so it may be quite impossible to find even one person willing to put the effort into maintaining the database core.
Tom,
As someone who was beginning to use dbXML just as it switched over to Xindice, I've been both a beneficiary and a victim of its situation. Given that you and Kimbro were the people with "hard core database skills" who absented themselves from the Xindice project, you two as much as anyone were the ones who doomed Xindice to oblivion. I don't know why dbXML ended up getting your attention, and perhaps that's none of my business, but the damage is done. There was nobody available to step into your shoes, and the project has languished. You apparently are either unable or unwilling to step into those shoes at this time -- I was surprised to see your message stating the status of Xindice so flatly, since it was your lack of involvement as much as anyone's that is the cause.
I don't state this with any blame or anger (maybe some frustra- tion), really, just as a matter of record.
In the past week I finally started to fix the problems I have been having with my own project's use of Xindice 1.0 (not daring to upgrade to 1.1) by trying out eXist. I would have liked to use Ozone, but I don't think I have the time and the latitude to experiment in whether it would work okay or not, and the new version's support for XML:DB is still nascent.
So, I tried out eXist; I spent about a day swapping over so that my project includes both Xindice and eXist 0.9.2, the latest version. It wasn't too difficult given XML:DB, but eXist had a few bugs. eXist relies to a small degree on Xindice's class CollectionManager, which in Xindice 1.1 is now an interface. The eXist implementation doesn't follow Xindice's API exactly, nor does the DOM implementation follow the DOM exactly either. So I had to make some minor modifications just to make it conform. But other troubles loomed.
If you look at the code of org.exist.dom.ElementImpl you'll see that the namespace URI is being totally ignored, and the method returns null rather than an empty string. And the XML namespace attributes are being completely lost.
So I've sent a message off to the eXist project head, including the bug fixes I'd done. After two days, no reply. He may be on holiday. But it's a very unsatisfying start to my eXist experience. These kinds of things are reminiscent of the early days of dbXML and Xindice. I've now shelved my eXist efforts until I get some feeling on any real activity happening within that project, whether there's any bug fixing going on, or if the project is going to do another Xindice number on me.
As for my XNode API, I tried farming it out at Apache's Incubator, joining their mailing list upon Brian Behlendorf's advice, but quit in frustration. Waaaay too complicated and involved a process to simply publish an API. I would have had to commit major amounts of time just to do so.
So in the end, there's no magic bullet except people with demonstrable skills pouring them into projects. Given that my priority is my Ph.D. project and that I don't have any particular database chops, there's little I can do here. But I must admit my frustration, exasperation, and curiosity in the fate of Xindice. If such a project can't survive at Apache, I don't think it will survive anywhere else. I'm sad to see it go, as had 1.1 been viable and on time I'd be using it. I can't stomach a GPL project (LGPL would have been okay) but I'm also not willing to switch to dbXML given that the effort in my opinion could have been spent on Xindice instead. They're both your babies, one was just left on a doorstep. The big question remains, that of all people who could have stepped up, why not you? And if you weren't willing, why didn't you tell us all about a year ago?
Murray
...................................................................... Murray Altheim http://kmi.open.ac.uk/people/murray/ Knowledge Media Institute The Open University, Milton Keynes, Bucks, MK7 6AA, UK .
[...] all matters of authority and responsibility are ultimately matters of social practice, and never matters of ontology (that is, never just a matter of how things in fact are in the nonhuman world). [...] just as we should not look to ground our moral judgments in the nonhuman authority of a god, so we should not look to ground our empirical judgments in the nonhuman authority of an external world. -- Robert Brandom http://www.tilgher.it/brandom.html
