On 09/01/13 18:17, Simon Helsen wrote:
Thanks Andy,
right now, we are not yet planning to upgrade from 2.7.4. I don't know at
this point when we'll plan to adopt the next upgrade, but I have to work
around changed release cycles on our end. Usually, we make the decision
based on a compelling new feature, but our biggest go/no-go is determined
by the amount of adoption work by our internal clients. They tend to be
more conservative and if they fear additional work, they usually want
something for it. So that is what I have to juggle. The one thing that
stands out right now in 2.10.0 that relates to our use-cases is the fix
for aborting bad queries which made it in 2.10.0. But not sure if it is
enough. I also like to coordinate with another internal group and I still
have to talk to them about what they think in terms of going to the next
release
It is a tricky balance between adopting every release, and jumping
across release but with the potential for compounded changes. The costs
of migration will always be there in a later release.
Even aside from adoption of a particular release, there is also the
feedback we could useful use on the changes before they are committed to
a formal release. Before release, migration support is mutable, after a
release, and it's the "new legacy" (a wonderful phrase I came across
recently).
btw, my comment below as intended to be more general, not just my own or
IBM's interests. Any Jena user benefits from consolidating incompatible
changes in as few steps as possible
Understood, and I didn't read it as an IBM-only point.
That's why I hope having an up-to-date development build is useful
externally, as well as for the developers, to get poked and prodded so
the consequences (unintended) don't get baked into a release.
Andy
Simon
From:
Andy Seaborne <[email protected]>
To:
[email protected],
Date:
01/09/2013 08:54 AM
Subject:
Re: Towards Jena 2.10.0
On 08/01/13 19:41, Simon Helsen wrote:
I have no specific requests/opinions other than that I rather deal with
one bang than many small ones. Continuous refactoring is appealing for
development, but a serious pain for adopters. And while I realize great
care has been taken to keep the disruption as minimal as possible, I
fear
that some of our internal clients will have unexpected lower-level
dependencies.
Simon,
The development snapshots now follow the new organisation and have the
compatibility placeholders. Once a release is done, it becomes the new
legacy. As it may take you more than one month (elapsed) for internal
client testing, I strongly suggest starting now.
Andy
