Working Group Minutes/EWG 2023-08-23
- Salim Baidoun
- Andrew Hain
- Adam Hoyle
- Tobias Knerr
- Roland Olbricht
- Emerson Rocha
Elect new chair
There is no candidate for the position. The attendees unanimously defer the election by four weeks. The work of the secretary is for the moment considered to be sufficient.
Tobias Jordans reached out to a contact who is interested in doing Rails work. Sarah Hoffmann supports that and suggests to extend the mission to communication tools for the website at large. Her suggestion is to plan with a ballpark number of 10.000 EUR for the total work.
The attendees support unanimously the following motion: the EWG is indicating to be willing to pay for inspecting the code base if necessary, and a ballpark number for this is 2 000 EUR. It is understood that other contract work cultures put the burden for that on the contractor, but we are are coward that this drives away even the few people interested in the code base that we have. The deliverable for a codebase inspection is an effort estimation for the issues to work on. These shall comprise the mute user project plus other communication related issues that the contractor deems to have synergies.
The attendees approve Paul's plan as accomplished deliverable for the contract. The actual plan to implement shall be derived from that, exect that Paul's plan is missing an executive overview. This shall be amended.
The Engineering Working Group deems the implementation of the plan to be board matter because it is substantially over the budget of the working group.
Data Model Work
Roland presents the opposition he has fathomed:
Jochen rightly pointed to the fact that the interdependencies require paid staff support. Stereo on the board preferred to spend money on something user visible, and indeed the STF application would cover the Data Model Work only for management and only tangentially.
On the hackweekend in February two talking points have been that proper Rails Code plus the best possible indexes for the data as relational database might be fast enough for any forseeable future. Given a decade of not delivering way geometry, Roland doubts that.
Another reason to dispute had been that the existing data model works good enough and other (non-user visible) areas rather deserve attention. Basically this has been disproven enough in the responses in the data model repository.
It is only an annoyance in current data, and for historic data we could throw hardware on it for a wrapper have been talking points, too, but Roland does not remember the origin. See the fact that it does not work already since a decade.
Some early discussion for context.
The Engineering Working Group decides to proceed as the lack of vision is not a valid argument. There is a problem to solve here, and the data model change still looks like the best approach.
The idea to give suggested topics to channel ideas has been dismissed: As shown with the mute feature topic, the Engineering Working Group shall treat them as their won projects independently of any Microgrants scheme.
Salim suggests to use a different platform than the wiki. He volunteers to re-read the Microgrants report for lessons learned about the platform itself and to make a suggestion based on the findings there. An example for an alternative is GitHub.
The topic will be revisited in two weeks.
"Nominatim Special Phrases, Overpass-Turbo Wizard and beyond" / OSM data intent (delayed)
After discussion by mail, one "low hanging fruit" found common features between OSM tools which could be generalized. Emerson plans to receive comments on the document "Nominatim Special Phrases, Overpass-Turbo Wizard and beyond" / OSM data intent here on next EWG meeting.
Emerson volunteers to do a quick presentation of it on the next EWG meeting.
Any other business
Next meeting is 2023-09-06, 15:30 UTC