[QGIS-Developer] Processing 3.0: Possible change to the Convex Hull algorithm

[Matthias Kuhn]

Thanks Giovanni and sorry for keeping my last message so short that it
was hard to follow.

Yes, indeed, the proposal was that the user experience should be the
same as now. A dedicated entry (algorithm) in the toolbox with an option
for "group by attribute".

That this "algorithm" internally is a model that calls two core
algorithms should be totally transparent to the user.

Matthias


On 6/28/17 12:12 AM, G. Allegri wrote:
> I understand and agree on having "atomic" algorithms but I'm not sure
> if we can always consider single features as the unique operating
> level. I know this would bring benefits to the code engineering but we
> have to keep in mind the end users expectations.
>
> I agree with Mathias, probably we should have atomic/feature level
> core algorithms and "processing algorithms" using them for higher
> level structures (datasets) in case they fit the task better.
>
> my 2 cents.
> giovanni
>
> Il 27 giu 2017 21:06, "Anita Graser" <anitagraser at gmx.at
> <mailto:anitagraser at gmx.at>> ha scritto:
>
>
>
>     On Tue, Jun 27, 2017 at 7:45 AM, Matthias Kuhn
>     <matthias at opengis.ch <mailto:matthias at opengis.ch>> wrote:
>
>         On 6/27/17 7:19 AM, Paolo Cavallini wrote:
>
>         > Hi Nyall,
>         >
>         > Il 27/06/2017 01:41, Nyall Dawson ha scritto:
>         >
>         >>> This discussion relates to the "Convex Hull" algorithm.
>         I'd like to:
>         >>>
>         >>> 1. Drop the "Field (optional, only used if creating convex
>         hulls by
>         >>> classes)" option and the accompanying method choice used
>         to set the
>         >>> convex hull to 'create convex hulls based on field'.
>         >
>         > I understand the rationale behind this; I'm just a bit
>         worried it will
>         > be more complicated and less understandable for users.
>         > Perhaps adding another module that does both commands
>         (collect + convex
>         > hull) in one shot would be useful?
>         > All the best.
>
>         Maybe there is the need to ship some often-used models by
>         default then?
>         Thinking of code complexity and maintainability, it makes much
>         sense to
>         modularize algorithm code as much as possible
>
>
>
>     ​I absolutely understand the reason from an engineering point.
>     From the user point - and as someone answering a lot of user
>     questions - it is predictable that end users will be confused. 
>
>     Idea: Have a warning if a user tries to run the new "convex hull"
>     on a dataset with single part geometries?
>
>     Best wishes,
>     Anita​
>
>      
>
>
>     _______________________________________________
>     QGIS-Developer mailing list
>     QGIS-Developer at lists.osgeo.org <mailto:QGIS-Developer at lists.osgeo.org>
>     List info: https://lists.osgeo.org/mailman/listinfo/qgis-developer
>     <https://lists.osgeo.org/mailman/listinfo/qgis-developer>
>     Unsubscribe:
>     https://lists.osgeo.org/mailman/listinfo/qgis-developer
>     <https://lists.osgeo.org/mailman/listinfo/qgis-developer>
>
>
>
> _______________________________________________
> QGIS-Developer mailing list
> QGIS-Developer at lists.osgeo.org
> List info: https://lists.osgeo.org/mailman/listinfo/qgis-developer
> Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-developer

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/qgis-developer/attachments/20170628/8b850b39/attachment.html>