I did not mean that any of them are blocking.

Keeping two documentation up to date is a growing pain. It's my weakness
though that it's still not done. Yet, if you plan to do a bigger reworking,
it's better done on the DocBook version.

On Sun, Oct 24, 2021 at 6:58 PM Siegfried Goeschl <
siegfried.goes...@gmail.com> wrote:

> Hi Daniel,
>
> Both topics are non-blocking in my opinion
>
> Regarding documentation:
>
> A complete Maven website including Markdown content is generated. And
> documentation will be updated, extended and moved to Docbook but that can
> be done any time - no need to introduce additional dependencies
>
> Regarding backward compatibility:
>
> * The code is mostly written by one person and that's me - so it is not a
> mature code base
> * There are hardly any users out there and new user will detect bugs,
> suggest improvements or will tell you that some parts are simply broken by
> design - enforcing backward compatibility will do some harm here
> * I consider backward compatibility important assuming that you HAVE many
> users out there
> * It is a command-line tools mostly used by developers and they know what
> a 0.x release means - some things are in motion and need time to settle
>
> Thanks in advance,
>
> Siegfried Goeschl
>
>
>
> > On 24.10.2021, at 15:16, Daniel Dekany <daniel.dek...@gmail.com> wrote:
> >
> > Yes, I guess we can get away with the Maven generated site, if the
> standard
> > ASF footer and that conference ad thingy can be added. It would be more
> > rational to push through with the conversion to DocBook though. The main
> > cause of the slowdown is that I had this idea that we actually run
> > everything that we show, and never copy-paste sources and output. That
> was
> > proven to be tricky in many cases, and is still unsolved in some (like
> > where the example uses Linux shell features). I should just let that go
> for
> > now and push through with the conversion with copy-pasting where we still
> > have problems. On that note, I wonder if you want to rework the content
> > anyway, like we want to move most examples outside the documentation, and
> > then people can open them in IDE, modify them to play around, etc. If you
> > do such reworking, or any serious reworking really, that should be
> already
> > done in DocBook.
> >
> > The warning about no backward compatibility needs to be apparent from
> > whatever documentation we release (the DocBook version has it). Backward
> > compatibility is really the main pain with the release. As we promise not
> > backward compatibility, we basically release software without promising
> > later support. People can still decide to use it (or they just don't
> > realize what this means). But, you may feel pressure to keep backward
> > compatibility instead of doing the right thing, which at this stage is
> > maybe not wise. (Also no support can be tricky when there's a security
> > issue with an old release. Although that's probably less relevant for a
> > tool like this.)
> >
> > On Sun, Oct 24, 2021 at 1:02 PM Siegfried Goeschl <
> > siegfried.goes...@gmail.com> wrote:
> >
> >> Hi Daniel,
> >>
> >> There is still the Maven-based site which can be created using
> >>
> >>> mvn clean site site:stage
> >>
> >> I will look into the source release packages ...
> >>
> >> Thanks in advance,
> >>
> >> Siegfried Goeschl
> >>
> >>
> >>> On 24.10.2021, at 11:38, Daniel Dekany <daniel.dek...@gmail.com>
> wrote:
> >>>
> >>> Still no site for example. Note sure about the others, had to review
> last
> >>> time's list.
> >>>
> >>> Can we build a source release package with all the necessary
> >>> NOTICE-s/LICENSE-s and signing? For this kind of project we will also
> >> want
> >>> a binary release package.
> >>>
> >>> On Sat, Oct 23, 2021 at 6:59 PM Siegfried Goeschl <
> >>> siegfried.goes...@gmail.com> wrote:
> >>>
> >>>> Hi folks,
> >>>>
> >>>> What stops us from having the first release? Any blockers we are aware
> >> of?
> >>>>
> >>>> Thanks in advance,
> >>>>
> >>>> Siegfried Goeschl
> >>>>
> >>>>
> >>>>
> >>>>
> >>>
> >>> --
> >>> Best regards,
> >>> Daniel Dekany
> >>
> >>
> >
> > --
> > Best regards,
> > Daniel Dekany
>
>

-- 
Best regards,
Daniel Dekany

Reply via email to