With regard to block deployment and its similarity to general package management systems (deb, rpm, msi, etc.), should there be any discussion about central repositories a la Debian's stable, testing, and unstable?

A self-contained unit of code (component, package, whatever) is great, but there's something to be said for:

apt-get install foo

It is so convenient in fact that many people used Debian or Debian-based distributions solely as a way to avoid the hunt you used to do for rpm files.

How does a user find a package that does SVG if they've never heard of Batik? They know that they want to use SVG and serialize, but then they must go to a web site, find the list of COBs, read the docs on each one, and download the appropriate one? Are we expecting people to do a google search but with a site:cocoondev.apache.org modifier (which of course assumes they know to constrain the search there)?

My question is how should block directories and discovery be handled? I guess it could be considered an implementation detail, but I think it should be brewing on the back burner in peoples' mind at least. Having useful repositories requires more forethought than just a download directory.

Thoughts?

- Miles



---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, email: [EMAIL PROTECTED]

Reply via email to