On 3/4/07, Johan Sundström <[EMAIL PROTECTED]> wrote: > > > It makes Exhibit based mashup writing substantially easier too. > > > > Will you help me test that? I think it's easier to have a real > > "customer" for that feature, as I myself don't use it much, yet. > > You can count on that; it's my personal main interest in Exhibit. > Expect resistance to ideas that make that more difficult than it need > be. ;-)
As an aside, the actual mashup part of it is actually almost amusingly difficult at the moment, due to the well guarded item id attribute. Mashup writing, at heart, is about adopting, reconciling and interlinking two (or more) data sets from different descent, environments and/or maintainers. The means are closely tied to linkage, which is orchestrated through item id:s and exhibit expressions, which in turn find items by id. We are presently left to inspecting data sets in advance in a white room environment (which is not always possible, probably even "rarely", given mashups of live data sources), checking for id clashes, or to run it and inspect the whole set. In the latter case, we can not inspect the id:s themselves, though, (because they are hidden and can't be exposed), but are left to finding the symptom, of disparate items joined into one, mainly detected by something having excessively many tags, long names or similar. Even given these friction losses, Exhibit remains a very powerful tool for mashups, but more for lack of competition than for being especially cut out to the task. An excellent mashup writing tool needs base support for working with, rather than against, item id:s. -- / Johan Sundström, http://ecmanaut.blogspot.com/ _______________________________________________ General mailing list [email protected] http://simile.mit.edu/mailman/listinfo/general
