- From: Danny Ayers <danny@panlanka.net>
- Date: Sun, 22 Apr 2001 11:08:45 +0600
- To: "RDFInterest" <www-rdf-interest@w3.org>
<- But I don't have any feeling for what it is that you want to <- actually DO. What are you trying to do with those documents and <- metadata? What problem are you trying to solve? I want to be able to make the documents available online in a form that other can access with ease. <- A very common problem is to search a collection of documents <- by author and publisher. In that case, why not have a RDBMS <- that gets loaded up with author, title, and publisher info? <- That is not RDF, but the point is to solve problems, not <- exercise technologies gratuitously. I should have made myself clearer - I'm talking in the Semantic Web context with the data & metadata will be exposed. If I just want the stuff in a database then I'll just put the stuff in a database. <- Another scenario might be that you want to provide those 10k <- documents to someone else. In that case, you might send the <- documents and an RDF description of each (generating the RDF <- descriptions out of the database mentioned above). The receiver <- can then parse the descriptions and import their content into <- their own database. Wasteful of bytes being transferred? Yes. <- Simple to implement and understand? Yes. <- So, unless I am working on a problem where the cost of <- transferring bytes is higher than the cost of programming, <- I'd do it the simple way. Yes, fair enough. So the simple way has a high cost in bytes. The system is at the early stages of development, why not reduce the cost of the simple way at this stage. Why build a wasteful system? <- > The model for what's ideally needed is multiple inheritance, <- or at least <- > something approaching this. <- <- "Needed" to do what? It sounds to me like you want to <- be able to describe things using the minimum number of statements. <- That is a fine thing to want to do, but it seems like a <- "nice to have" rather than a "must have" for any applications <- that I am currently looking at. I just think it might be worth considering. We don't *need* any of this IT stuff. Or taking an intermediate line - there are a great many applications that can be built using the BASIC programming language, why should we bother with these object things? We can solve problems one at a time, or we can make extensible tools that will allow us to solve a whole load of them, and be better prepared for tomorrows problems - using good design methodologies means this needn't be any extra effort. <- The hardest part about RDF is clearly identifying the problem <- you want to solve, and distinguishing it from related problems <- that you don't need to solve. RDF is a means to an end, not an <- end in itself. In my analogy there were 2 approaches, one of which would mean potentially a 1000-fold reduction in the number of bytes wasted. Ok, let me add a factor to that analogy - I've got limited bandwidth. There, now the efficiency has a bearing. It may be that bandwidth is a consideration in the real world too.
Received on Sunday, 22 April 2001 01:13:39 UTC