Thanks for this Gary.
I agree with the fundamentals of the proposal as it makes sense to me to
use the existing database model as a start for the new Bloodhound. I think
it would be much easier to make changes to a fully formed database model
than to design this from the ground up.

I will try to take part in this debate going forward.

Regards

John

On Wed, 28 Apr 2021 at 23:41, Gary Martin <[email protected]> wrote:

> Hi everyone,
>
> Sorry it has been so long. I have been putting in some time to look at
> various things around the Apache Bloodhound project to try to push things
> along a bit.
>
> One of these things was to attempt to fix a problem with accessing the
> issue tracker. I am not yet happy with this fix as, at least temporarily,
> it leaves me as the only person with an account. This may not matter too
> much if we are willing to go through re-registration (and, yes,
> registration is still disabled at the moment so this may be fun to
> organise) but as there may be a need for a further re-registration event
> when we migrate to a Django based Bloodhound, I am reluctant to force this
> upon people. Hopefully I can report on this further soon.
>
> Anyway, with that sorted temporarily, albeit just for me, I have been
> using my access to begin putting some work into describing my current ideas
> for Bloodhound at [1]  - it is currently very rough but I'm hoping that it
> will evolve fairly quickly with input from others. I also didn't want to
> put it in place fully formed as I would prefer that others would be able to
> get their ideas into the proposal.
>
> One of the bigger changes from what I have suggested in the past is that
> it may be more pragmatic to start from database models that are closer to
> those that we already have and look to change them where we think this is
> appropriate, supported by Django's migration mechanism.
>
> I have included a diagram that I have generated from a possible
> translation of the current Bloodhound database to Django models which can
> be found in the proposal (or see ref [2] below for a direct link to that
> attachment.) I have not proved that I have got the model equivalent in
> enough detail at this point but I am planning on getting these changes
> prepared for our bloodhound-core git repo shortly.
>
> I'm looking forward to hearing opinions on any of this.
>
> Cheers,
>     Gary
>
> [1]
> https://live.bloodhound.apache.org/bloodhound/products/BHD/wiki/Proposals/BEP-0016
> [2]
> https://live.bloodhound.apache.org/bloodhound/products/BHD/attachment/wiki/Proposals/BEP-0016/bh_models.dot.png
>

Reply via email to