At this point I think the main advantage to an open source ZIS (the xml message router) is that it will make it much easier for open source developers to experiment with SIF and make more likely the design of new applications with this technology in mind.Again true. However, the "existential reality" is that it doesn't exist...<grin>It does exist http://www.openzis.org/ http://openzis.tigris.org/ Should release first candidate based off some older code in the comming weeks that will work with the SIF 1.1 Spec.
I think this discussion thread misses some imortant issues regarding SIF. The most glaring issue is who said SIF was a standard in the first place?
I don't think anyone sat on an ANSI or ISO committee and produced a real standard. Perhaps they did. Is this an ISO or ANSI standard?
If this gets widespread adoption, it will become de facto and lack de jure blessings. We don't need any more proprietary standards unless they are promulgated by standard making bodies. I include the Open Group as a proprietary organization attempting to create stanards. They sold Microsoft their DCE license and allowed Microsoft to alter DCE to the point no one can use the protocols except Microsoft. DCE became know as Object Linking and Embedding (OLE), then ActiveX and onward.
The disinformation network has a new saying, " Starting with an open application is like having someone give you a free puppy".
If you haven't heard that one yet, expect to hear it in the future. I suggest that getting a Microsoft application is like getting a ROTTWEILER with parvo.
Open source doesn't ignore standards. Whoever said that needs a reality check.
Active Directory, for example, claims to be standards compliant with RFC 2251. Have you tried to work with it? It started with an RFC complaint standard and then zig-zagged all over the place. It's not RFC compliant.
I agree that we should ignore SIF and work on open source applications. As Les said:
The libraries have done a good job of building servers which have a backend mapped to their existing catalogue management systems. The servers provide the interconnectivity, IR, etc. and those are z39.50 ISO standards based servers.Let's say you're going to build a distributed system of small applets to solve educational problems. Would you build your system on top of a system of vapourware? (or on top of a complex and continously changing specification?) Since most of our projects are tiny we have very distinct limitations in terms of time and resources that we can use to write applications. I just feel that we should husband these resources very carefully.
I recommend that we turn to Universities to form an ISO commitee for the standards recommend originally by SIIA and allow them to provide funding for the standards and Open Source Sofware implementations.
But, that's just me. I admit to having a bias toward public domain solutions. I don't think the people (as in 'we the people') need to keep paying over and over again for the same software when we can just do it once and share it for the common good.