On 10/13/07, Bertrand Delacretaz <[EMAIL PROTECTED]> wrote:
> On 10/12/07, Keith R. Bennett <[EMAIL PROTECTED]> wrote:
>
> > ...I would like to propose as another criterion for releasing 0.1 that the
> > features we expose in the API work correctly (or are documented to not work
> > correctly), and are verified by unit tests....
>
> I (respectfully) disagree on having to be "perfect" in these areas for
> our 0.1 release.
>
> In my view this is clearly a "release early, release often" release.

+1

until tika has a release, other projects have to depend on SNAPSHOTs
which limits usage

> As long as people can experiment with it on a reasonable number of use
> cases, and hopefully give us feedback, doing the release (with all
> required disclaimers about alpha quality) has value.
>
> Having open issues like TIKA-56, for example, is fine IMHO. They are
> just that: open issues that will hopefully be fixed before the next
> release, but don't prevent people from experimenting with Tika 0.1.

but keith's proposal is definitely worthwhile: it's very easy to add
new features without ensuring that existing ones work. perhaps 0.2
tasks could be created along the lines suggested...?

- robert

Reply via email to