incubator-odf-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jeremias Maerki <>
Subject Re: Choice of JDK version (earlier -- Re: Source code checked in, what next?)
Date Tue, 20 Sep 2011 10:29:57 GMT
On 20.09.2011 12:15:36 Svante Schubert wrote:
> Our project is meant as a software tool for other developers and is not
> addressing the end user market.

By users, I meant developers using ODFDOM, i.e. the users of the library.

> I have observed the laziness of user, similar to the desire of stability
> of user that are not changing any software of a running system at all.
> The latter makes sense, but if there is sufficient test coverage there
> is no problem.

...or "probably no problem". People still have to re-integration-test
their systems after a version upgrade.

> I fully agree on the early adoption as acceptance killer and that the
> EOL is a decision of the manufacturer.
> Nevertheless there are no further security patches and there should be
> some date to move on, as we are saving time in our development, becoming
> more efficient.

Of course, there is a time to move on. The trick is to find the right

> I am curious why people stick with an old Java version and taking the
> security risk?

Because they don't expose their system to the evil internet. Because
they may not be able to migrate to a newer platform because their
application server doesn't support it, yet, or because some legacy
software (that can no longer be upgraded) doesn't work with a new JDK.
Or because the boss doesn't allow it or doesn't want to allocate the
resources for the upgrade of the whole platform. All more or less lame
excuses but it seems to happen surprisingly often. That's why I was
talking about "realities". Of course, you could argue: why would they
then want to upgrade ODFDOM or doing new development with ODFDOM? I'm
not saying anything like you should do this or that. I just advocate to
make an informed and balanced decision.

> Am 20.09.2011 11:58, schrieb Jeremias Maerki:
> > Just my 0.05 CHF as a bystander...
> > Reasons not to change could be:
> > - Developers want to use the latest toys but may sometimes forget about
> > the users.
> > - The reason for a higher JDK may be triggered only by some special
> > feature while the majority of the code would work just fine by itself
> > with a lower JDK. Maybe some core edition could use JDK X while an
> > extension component with the additional feature could require JDK X+1.
> > - Too early adoption of a new JDK (currently JDK 7) could be a serious
> > acceptance killer (and therefore an incubation risk here). The lower the
> > JDK, the larger your possible audience.
> > - JDK EOL is a decision by the manufacturer of the JDK. It does not
> > always mirror the realities in the wild.
> >
> > Polling the community is always a good way and gives you interesting
> > data. See [1] what we did in Apache XML Graphics Commons some way back
> > as an example. We switched from JDK 1.3 to 1.4 as a consequence, but
> > only switched to JDK 5 about a year ago.
> >
> > [1]
> >
> > On 19.09.2011 23:45:55 Svante Schubert wrote:
> >> Am 19.09.2011 23:23, schrieb Nick Burch:
> >>> ..
> >>>
> >>>> What is Tika's reason to stick with JDK5?
> >>> I think the most recent discussion on it was:
> >>>
> >>>
> >>>
> >>>
> >>>> ODFDOM's reason to use JDK 6 is its need for the ODF package
> >>>> encryption & signature feature and even JDK 7 comes along with
> >>>> improved ZIP access of NIO2, therefore I will from my current point
> >>>> of view I will vote to switch to JDK in July 2012 when JDK 6 is EOL.
> >>> If you have a reason to change, then you can, I was just suggesting
> >>> that you consider a major version bump along with it to make it clear
> >>> to everyone
> >>>
> >>> POI switched to needing JDK 5 back in 2009, and we still sometimes get
> >>> people asking about letting it run on JDK 1.4!
> >> The pure existence of some people desiring some thing, is not already a
> >> reason for an offering. There is always desire for everything and there
> >> might be even fools out there. It is quite the opposite, there should be
> >> a good reason to stick with the old JDK version.
> >>
> >> Neglecting new features, keeping multiple JDK versions on development
> >> machines, multiple JDK test runs, time that could be spend on
> >> enhancements, the risk of security breaches on all involved machines
> >> with outdated JDKs.. all this for what benefit?
> >>
> >> The question should not be why should we change, it should be why should
> >> we NOT change.
> >>
> >> But perhaps it is just a different way of thinking...
> >>
> >> Regards,
> >> Svante
> >>
> >
> >
> >
> > Jeremias Maerki
> >

Jeremias Maerki

View raw message