From: Sean O'D. <se...@ce...> - 2004-09-02 17:47:01
|
On Thursday 02 September 2004 10:31, Clark C. Evans wrote: > Sean, > > While I don't agree with the implementation of your proposal; I do really > like the 'gist' of what you are proposing -- a mechanism to fetch all > kinds of information, such as the owner, user documentation, schema, and > other information about a YAML tag. Could we work towards a concrete > mechanism that provides some sort of tag catalogue? > <snip> > > This would eventually become a _huge_ file, but it can be compressed > and/or represented in a database. And what ever sort of schama > information can be included as a sub-document or as a URL referencing > something external, it just doesn't matter. YAML distributions could > ship with the full file, or a subset, and we could have a nice browser > and search feature to find items you like. Large and difficult to maintain and distribute. I don't think one big catalog is wise. It would okay the first time someone authored it, because everyone would be interested and would help out, and it would work fine at first. But when everyone stopped helping provide updates, and it became the lone task of one or a couple people, it would start to fall behind. Eventually, I think it would become so unreliable as to be useless. And it would grow...filled with ancient schemas hardly used. If a schema is noted in the header of a document, it should provide the location of the schema as well. That way it scales. If 10,000 documents are asking for a single schema, that one schema has to be provided for appropriately. Posted across several hundred geocities accounts, of course. Schemas that are private and used once or twice only can be locally referenced, or placed on a temporary web server. Sean O'Dell |