> Adam R. B. Jack wrote:
> >So one can look at a repository of artefacts and select the "best" for
> >user automatically. Each user has a different view of "best", some want
> >latest (nightly/snapshot), some want latest "release" only.
> >
> >Having a repository with automated tools that require humans to
> >the contents kinda breaks the automation...
> >

"Michal Maczka" <[EMAIL PROTECTED]> wrote:

> Not necesserly the fact that version number is not parsable by machines
> determines that fact that versions are uncomparable.

That sounds more like a 20/80 rule, i.e. you can't for some -- so don't
bother for all. ;-) Who is to say we couldn't get some sort of agreement
amongst apache software? IT doesn't have to be today, maybe later, but I
feel it is critical that the artifacts are "understandable" in order to
allow intelligent processing.

Two versions only need to be comparable with others within their own
namespace, not against all others.

> Simply metdata atached to artifacts  (which can be denoted as URL)  is
> not  sufficient for  making smart tools and never be.

I know, I completely agree, it is not for all cases. But, I believe we set a
miniumum bar that a repository ought be able to do a mimimum set of things
even without metadata. If we can't build a tool to download the "right" jar
without metadata we have the bar too low.

> Artifact versioning is horribly complicated topic and I am simply afraid
> that attempts to e.g. guess version precedence  from version strings
> are not relaiable in the best case.

Yes, but there are some established standards -- it isn't all chaos:


Metadata could inform a tool what version standard it adheres to, or not,
and tools could then leverage. I think stating a versioning scheme (should
one exist) is much better than positional metadata. I might be hoping for
too much, but I think it is worth considering, especially within the
confines of an Apache repository.

Let's take this to Wiki [I'll figure out a way]. I (clearly) have a bias
towards trying to process/understand versions ( see
http://www.krysalis.org/version ) -- but I'm *NOT* trying to force this on
the repository, just trying to ensure that version processing doesn't get
overlook accidentally, only intentionally [if it doesn't fit].



Reply via email to