[Qgis-user] Node Tool - proposing improvements

Wed Nov 28 03:13:14 PST 2018

Hi, it's time for me to join the discussion. Sorry for being late.

I have been involved in the new node tool stages along with others.

First, I agree there are issues with the new node tool, and I think we can
improve this a lot. I myself was not convinced at start it was a good
change. I had to practice a lot, forget my old habits, to finally recognize
this change opens a lot of new possibilities. I also asked at start if we
coudn't keep the old edit mode, and add an optional 'CAD-like' one. ( Under
the hood, the snapping engine does not allow that simply, and this would be
really a UX nightmare, so I didn't fought)

On the rationale and the process of changing the node tool, here is what I
am aware of:

- the old node tool did not allow the use of advanced digitizing
constraints.
- the old node tool did not allow multi layer editing
- Users and contributors worked (an funded) a lot to propose enhancement to
the node tool. A QGEP was opened
https://github.com/qgis/QGIS-Enhancement-Proposals/issues/69.
- A proof of concept plugin was even made.
https://github.com/wonder-sk/CadNodeTool
- In the project I was involved in, which mainly concern editing of
networks (maybe the issue is here), we tested intensively the plugin, and
asked for changes until it can be ported to C++
- We asked quite few times to many users to test the tool during the 3.0
pre release and "early adopters" stages. *We received almost no feedback
then. *

To sum-up things, I think that quite a lot of people have been involved
here, the process has been transparent as much as possible.* Still the use
cases that pushed this changes did not cover enough intensive digitizing of
massive polygons enough*. Moreover users /contributors involved in this
kind of use* did not raise their hand during 2.99, 3.0, 3.2 releases and
feature freeze period.   *

So I think we just need to just recognize that *we miss some early testers*
and
*I would like that every company using QGIS just tries to identify some
"early adopters" among their staff and dedicate some time to anticipate
what in coming by letting them test, read the QEPS, and so on.  *

*And now, we just need to polish the vertex tool. *

As far as my company is concerned, we were not satisfied by two points:

- There was no way to select one particular polygon. Martin Dobias added
just some minutes before the 3.0 release a "selected feature priority" as a
workaround. *Many thanks Martin. *
- Multi layer editing was the default behavior. This didn't play well with
transaction group and triggers in database. My company funded and
developped directly the "single layer vertex editor tool" to avoid that.

We would be glad to help in polishing the tool, but someone needs to
support this now I think. QGIS is user and developper driven. Release
early, release fast process is just made for that.

Best regards
Régis

Le mer. 28 nov. 2018 à 09:59, Antongeo76 <antoviscomi at gmail.com> a écrit :

> Saber Razmjooei wrote
> > I agree with Bernhard. Last time, it was done through QEP and no real
> > feedback from users. The change to the node tool behaviour requires input
> > from experienced users who do digitizing in their day-to-day works. I
> > suggest to have the draft proposal as a blog post on QGIS.org to get more
> > users involved in the way forward.
> >
> > Currently, there are some bugs and several contradicting preferences and
> > requests. Any new changes to the current behaviour without consulting
> with
> > users will probably take us back to square one.
> >
> > Regards
> > Saber
> >
> > On Tue, 27 Nov 2018, 06:51 Bernhard Ströbl <
>
> > bernhard.stroebl@
>
> >  wrote:
> >
> > > Hi Nyall,
> >>
> >> the reason to post this issue to the users' list was to get feedback
> >> from users (and developers, too). I am totally aware that a QEP would be
> >> appropriate but I doubt that many users are taking part in the
> >> discussion there.
> >> Anyways I am going to sum up the results and create a proposal later on,
> >> will that do?
> >>
> >> Bernhard
> >>
> >> Am 27.11.2018 um 01:23 schrieb Nyall Dawson:
> >> > On Mon, 26 Nov 2018 at 21:39, Bernhard Ströbl <
>
> > bernhard.stroebl@
>
> > >
> >> wrote:
> >> >>
> >> >> Dear all,
> >> >>
> >> >> I want to apologize, this has become an extremely long mail...
> >> >>
> >> >> A lot of discussion has been going on about the recent
> >> reimplementation
> >> >> of the node tool in QGIS 3.
> >> >
> >> > Hi Bernhard,
> >> >
> >> > Thanks for the detailed proposal, and for taking the time to come up
> >> > with a constructive way forward here.
> >> >
> >> > Can I encourage you to copy the contents of your proposal across to
> >> > the "QEP" repository:
> >> > https://github.com/qgis/QGIS-Enhancement-Proposals/issues ? I think
> >> > that's the correct channel to discuss these ideas so that the
> >> > conversation is kept central and recorded permanently for reference.
> >> >
> >> > Nyall
> >> >
> >> >>
> >> >> I see two reasons for this:
> >> >> 1) The way the tool works is very different from the way it used to
> >> work
> >> >> in QGIS 2:
> >> >> a) in QGIS 2 it was choose feature - choose node(s) - do something
> >> >> whereas in QGIS 3 it is choose node - do something
> >> >> b) to move a node was click - (keep pressed) - move - release which
> >> was
> >> >> changed to click - release - move - click - release
> >> >> 2) There have been several bugs (as could be expected with a new
> >> >> implementation). These, however, were used to question the whole
> >> feature
> >> >> (most of the bugs are fixed now).
> >> >>
> >> >> Apart from bugs and people not being used to the new tool the critics
> >> >> concentrate around:
> >> >> 1) Probably most annoying: Mouse movements for choosing an existing
> >> >> vertex/adding a new vertex at the middle of the line are not well
> >> >> defined (especially important if features and its vertices are very
> >> >> close together). Imagine a polygon surrounded (or even worse:
> >> >> overlapped) by others, now try to click a particular vertex. You will
> >> >> fail if always the neighbouring polygon is highlighted. The
> workaround
> >> >> is to select the vertex with a mouse window while its feature is
> >> >> highlighted (or previously select the feature e.g. with the
> >> >> Select-Feature tool). For overlapping features it is even harder.
> >> >> 2) Highlighting effect is disturbing.
> >> >> 3) Adding a new vertex puts the new vertex on the mouse ready to be
> >> >> moved but user has no intention to move it but to leave it where it
> >> has
> >> >> been added (e.g. for topological reasons).
> >> >> 4) Vertices are not visible all the time but only when the feature is
> >> >> highlighted. This can make choosing a particular vertex difficult
> (see
> >> >> also 1).
> >> >> 5) It is easy to accidentally add or move a vertex because the most
> >> >> intuitive way of selecting something is to click it. But when you
> >> click
> >> >> a vertex/segment/cross while being highlighted you are already on
> your
> >> >> way to move it.
> >> >> 6) The behaviour is not consistent:
> >> >> a) If you want to move _one_ vertex, try to grab it when it is
> >> >> highlighted then move it; if you want to move several vertices, first
> >> >> select them, then click them, then move them.
> >> >> b) If you want to delete a vertex, try to grab it when it is
> >> >> highlighted; if you want to delete a segment, do not try to grab it
> >> when
> >> >> it is highlighted but select its two vertices and delete them.
> >> >>
> >> >> The German user group thinks that the points raised above are valid,
> >> >> even if all pending bugs are fixed. The node tool is still not
> perfect
> >> >> (it wasn't in QGIS 2 either).
> >> >>
> >> >> So we propose a change in the way the node tool works and would like
> >> to
> >> >> hear other users' s opinions.
> >> >> While the click - move - click change has valid reasons (use advance
> >> >> digitizing panel, pan or zoom map while moving vertices) the dropping
> >> of
> >> >> the first choose feature - then choose vertex workflow to immediately
> >> >> move a vertex has no real reasoning (except that it is less clicks in
> >> >> those cases where the node can be easily grabbed in the first place).
> >> >> But this new workflow is the main cause for the problems listed
> above.
> >> >>
> >> >> Therefore we would like to see the old workflow reimplemented with
> >> some
> >> >> of the new ideas added plus some new ingredients. This is the draft
> >> for
> >> >> how the tool might work in the future:
> >> >>
> >> >> 1) While the mouse is moved around features are highlighted to
> >> indicate
> >> >> they can be edited. Furthermore eventual vertices, segments or
> crosses
> >> >> are highlighted, too (as in QGIS 3).
> >> >> 2) While a feature is being highlighted, it can be chosen with a left
> >> >> click (new).
> >> >> 3) If the user makes a right click instead, another (adjacent)
> feature
> >> >> is highlighted and can be chosen with a subsequent left click and so
> >> on
> >> >> (new).
> >> >> 4) As soon as a feature is chosen, the Vertex-Editor panel is opened
> >> (as
> >> >> in QGIS 2).
> >> >> 5) A chosen feature is the only feature whose vertices can be edited
> >> at
> >> >> that point* (as in QGIS 2). The chosen feature and all its vertices
> >> stay
> >> >> visible until the end of this feature's editing session (as in QGIS
> >> 2).
> >> >> If a vertex was highlighted it is immediately selected, if a segment
> >> was
> >> >> highlighted, its two vertices are immediately selected, if the middle
> >> of
> >> >> a segment was highlighted with a cross a new vertex is created there
> >> and
> >> >> immediately selected (new).
> >> >> 6) A vertex is selected by left clicking on it (as in QGIS 2).
> >> >> 7) Several vertices are selected by using a mouse window (as in QGIS
> 2
> >> >> and 3).
> >> >> 8) A click on a segment selects its two vertices (as in QGIS 2).
> >> >> 9) Vertices can be selected via the Vertex-Editor panel, too (as was
> >> in
> >> >> QGIS 2)
> >> >> 10) Every new vertex selection clears any previuos selection except
> if
> >> >> Crtl is pressed which adds the new selection to the current if
> >> vertices
> >> >> were not selected or subtracts them from the current selection if
> they
> >> >> were selected (as in QGIS 2).
> >> >> 11) Selected vertices can be deleted (as in QGIS 2 and 3).
> >> >> 12) Selected vertices can be moved by clicking either of them, thus
> >> the
> >> >> vertex (the vertices) are attached to the mouse and can be dropped
> >> with
> >> >> another left click (as in QGIS 3). A right click aborts the moving
> >> >> operation but the vertices are still selected (new).
> >> >> 13) A new vertex can be created by double clicking on any segment (as
> >> in
> >> >> QGIS 2 and 3), the new vertex is immediately selected (new).
> >> >> 14) A right click ends the chosen feature's editing session, if
> >> another
> >> >> feature is within reach it is immediately highlighted, continue as in
> >> 1)
> >> >> (new).
> >> >> 14b) To be discussed: Left clicking another feature immediately
> >> chooses
> >> >> this feature (as in QGIS 2).
> >> >>
> >> >> * topological editing is still posssible, of course
> >> >>
> >> >> We see the following advantages in this approach:
> >> >> 1) The new quick editing possibilities are preserved, there is just
> >> one
> >> >> additional click (compared to current QGIS 3) to move an existing
> >> vertex
> >> >> or segment or to create and immediately move a new vertex.
> >> >> 2) It is easier to create a new vertex in place.
> >> >> 3) It is a lot easier to graphically choose the feature to edit.
> >> >> 4) The highlighting indicates, which feature will be chosen by an
> >> >> immediate click, no surprises for the user.
> >> >> 5) It is easier to graphically select the vertices to edit.
> >> >> 6) The workflow is always the same: Choose feature - select vertices
> -
> >> >> do something with them.
> >> >> 7) The workflow is consistent with creating a new feature (right
> click
> >> >> to finish editing).
> >> >> 8) The application is more consistent because graphically selecting
> >> >> vertices works the exact same way as graphically selecting features
> >> with
> >> >> the Select-features tool.
> >> >>
> >> >> I appreciate your feedback
> >> >>
> >> >> Bernhard
> >> >>
> >> >>
> >> >> __________ Information from ESET Mail Security, version of virus
> >> signature database 18443 (20181126) __________
> >> >>
> >> >> The message was checked by ESET Mail Security.
> >> >> http://www.eset.com
> >> >>
> >> >>
> >> >> _______________________________________________
> >> >> Qgis-user mailing list
> >> >>
>
> > Qgis-user at .osgeo
>
> >> >> List info: https://lists.osgeo.org/mailman/listinfo/qgis-user
> >> >> Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-user
> >> >
> >> >
> >> > __________ Information from ESET Mail Security, version of virus
> >> signature database 18446 (20181126) __________
> >> >
> >> > The message was checked by ESET Mail Security.
> >> > http://www.eset.com
> >> >
> >> >
> >>
> >>
> >> --
> >> Bernhard Ströbl
> >> Anwendungsbetreuer GIS
> >>
> >> Kommunale Immobilien Jena
> >> Am Anger 26
> >> 07743 Jena
> >>
> >> Tel.: 03641 49- 5190
> >> E-Mail:
>
> > bernhard.stroebl@
>
> >> Internet: www.kij.de
> >>
> >>
> >> Kommunale Immobilien Jena
> >> Eigenbetrieb der Stadt Jena
> >> Werkleiter: Karl-Hermann Kliewe
> >>
> >>
> >> __________ Information from ESET Mail Security, version of virus
> >> signature
> >> database 18447 (20181127) __________
> >>
> >> The message was checked by ESET Mail Security.
> >> http://www.eset.com
> >>
> >>
> >> _______________________________________________
> >> Qgis-user mailing list
> >>
>
> > Qgis-user at .osgeo
>
> >> List info: https://lists.osgeo.org/mailman/listinfo/qgis-user
> >> Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-user
> >
> > _______________________________________________
> > Qgis-user mailing list
>
> > Qgis-user at .osgeo
>
> > List info: https://lists.osgeo.org/mailman/listinfo/qgis-user
> > Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-user
>
>
> Hi Saber,
> the discussion on node tool seems like a fight against windmills,
> where few users raise doubts about the actual improvement of the node tool
> making them look like isolated cases or local and private matters.
>
> I agree with you even if it seems that there are few users interested in
> the
> subject in this mailing list I among others I tried to expose my concerns
> in
> QEP but I have not any feedback.
> I think there is a big gap between the development of this very important
> part and the end users!
>
> Then I would suggest the following
> could you create a survey directly at the opening of QGIS?
> For example a link (maybe in the welcome dialog) to a dedicated forum that
> gives the possibility to ask users what they believe is better to speed up
> and optimize node tool or other features in QGIS.
>
> Many of the users are not subscribed to the mailing list and don't even
> know
> that there is a mailing list!
> We should find a more incisive way to publicize the possible interactions
> between users and developers
>
> Greetings
>
> Antonio
>
>
>
> -----
> Antonio
> --
> Sent from: http://osgeo-org.1560.x6.nabble.com/QGIS-User-f4125267.html
> _______________________________________________
> Qgis-user mailing list
> Qgis-user at lists.osgeo.org
> List info: https://lists.osgeo.org/mailman/listinfo/qgis-user
> Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-user
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/qgis-user/attachments/20181128/1fd301ba/attachment-0001.html>