Selecting the tools before you know what your needs are
(blowing off some steam here)
Someone blogged earlier about selecting tools and platforms before you're sure about what your needs are. Looks like I'm going to be affected by that now. I'm one of the responsible architects for implementing a java based portal that must integrate with a document management system and a search system that knows nothing about each other, has no standard ways of plugging into each other and no one seems to have done this before. Yay! Besides, one of the requirements are to have automatic and seamless login for the windows users into the portal, which, as I said, is based on Java...
We fixed the automatic login requirement by using JCIFS open source stuff from samba, which implements the CIFS/SMB protocol, but IMHO it's sort of a hacky way of doing it. I'm looking at replacing that part with some 3rd party access management system though... For large enterprise sized portals I prefer using a supported way of doing access, which *is* a pretty important part. No access - no portal, right?
Well, only thing to do is just smile and dig into each product and make them talk to each other somehow. It will take much longer to implement and it won't be a standard way of integrating portals with content and search, that's for sure.
Some people complains about the risk of "getting stuck into proprietary Microsoft products", but personally I prefer that to getting stuck into a much smaller product that once sprung up from some open source portal framework initiative and now has evolved into a monster that needs a gazillion of bytes just to wake up and start chewing request! During the time it grew from baby monster to a huge monster the owner of the pet changed a number of times and the price went from zero buckazoids to a price relative to the amount of memory needed to get it running...
At least I got the power (in my dreams) to select the platform for building the Web Service layer we will need to integrate with other systems. I'll certainly raise my hand for .NET Web Services and use WSE 2.0 (or what there will be in ASP.NET 2.0) for security. I'll see how far AXIS and WSS4J has come...
I need to say though, that I don't mind being involved in this project because even though that it could have been easier with other products, it's pretty fun to try to solve problems with integration, security, where to store content and how to search for it :)
(I feel much better now, thanks)