Re: [Qgis-developer] Bug with avoid intersection - willing to fund fix
Hi Andreas Of course I fully understand your point about the need for financing, and hopefully George`s offer will change things for topological correct editing. However, there could be several reasons for a lack of funding until now: Probably, many users are not really aware of the bug(s)? Back in 1.7.4 editing was no longer possible once invalid geometries were produced by "Avoid intersection" and users got an error message. Nowadays, you can continue digitizing without noticing that geometries are not topologically correct. The result looks good and you have to actively search (or check) for such digitization errors. I guess, not everybody does that, so I assume also a lack of awareness could be a reason for the lack of funding... Furthermore, some users who are aware may prefer such (of course not too convenient) workarounds over trying to raise some money for funding a proper fix. And finally, as it is unfortunately the case for me - for example -, even if I am seriously affected by the bug it is really very difficult (for not to say impossible) to raise a reasonable amount of money to get this (or another bugs) fixed, because: In my Institute ArcGIS is the "Standard" GIS application, and I am one of few "nerds" (that is at least how I feel I am sometimes being looked at) who use OSGeo products. I was and am actively promoting (amongst other OSGeo applications) QGIS in my institute because I am convinced that QGIS is in many cases much more suitable for our purposes than ArcGIS. One very important activity is mapping (and digitizing) habitat types for public authorities, who request data with proper topology. So, the topology problems are definitely an obstacle regarding promoting QGIS here. And since my bosses already paid for ArcGIS licenses (which does not have these problems) and a wider introduction of QGIS will also cause some extra transaction costs (e.g. learning) in short term, the consequence is that my institute will stick to ArcGIS. This makes it extremely difficult to convince my bosses to donate some money for bug fixing, even though QGIS offers (from my point of view) superior solutions for our problems and we will benefit from using QGIS in the long run... It is a bit of a witches circle in a way, and the only alternative would be to use private money... Cheers Stefan -Original Message- From: qgis-developer-boun...@lists.osgeo.org [mailto:qgis-developer-boun...@lists.osgeo.org] On Behalf Of Andreas Neumann Sent: 14. november 2013 10:54 To: qgis-developer@lists.osgeo.org Subject: Re: [Qgis-developer] Bug with avoid intersection - willing to fund fix Hi, So if there is so much interest in having this issue fixed I assume that there would be plenty out there who would fund the fix! The fix won't happen if noone pays for it - just complaining about this does not help. Get an estimate from some developer (e.g. Sandro Santilli oder Sourcepole) to know how much it would cost to fix. As much as I like GRASS - and it is very good to know that this workaround exists (thanks Stefan for mentioning it!) - I don't think it is reasonable to ask people to go through GRASS to get proper data out of their QGIS digitized data set. Who will take the lead to ask a developer? Andreas Am 14.11.2013 10:40, schrieb Blumentrath, Stefan: > Hi George > > I had a lot of trouble with those issues too, and I would really love to see > them fixed. In other words, I agree with Giovanni that these are pretty > important bugs for many and not really uncommon digitization use-cases and > would support his categorization as blocker... > > However, my current workaround is to ignore the untidiness in the topology at > first and then send the vector dataset through GRASS which is (at least for > not too big datasets) quite convenient using the Sextante/Processing > framework (see also discussion here): > http://osgeo-org.1560.x6.nabble.com/Running-v-clean-on-a-QGIS-vector-l > ayer-Polygon-through-Sextante-td5050418.html > > Cheers > Stefan > > > > ___ > Qgis-developer mailing list > Qgis-developer@lists.osgeo.org > http://lists.osgeo.org/mailman/listinfo/qgis-developer > ___ Qgis-developer mailing list Qgis-developer@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/qgis-developer ___ Qgis-developer mailing list Qgis-developer@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/qgis-developer
Re: [Qgis-developer] Bug with avoid intersection - willing to fund fix
I already asked a few devs without response. Ive even dug under the cpp hood to see if I could find somethimg interesting, and I did, but Im not sure on how to resolve the issue, since cpp is kinda of a black art for me :-) Devs around the world! How much it will cost to fix this bug? Em 14/11/2013 07:53, "Andreas Neumann" escreveu: > Hi, > > So if there is so much interest in having this issue fixed I assume that > there would be plenty out there who would fund the fix! > > The fix won't happen if noone pays for it - just complaining about this > does not help. > > Get an estimate from some developer (e.g. Sandro Santilli oder > Sourcepole) to know how much it would cost to fix. > > As much as I like GRASS - and it is very good to know that this > workaround exists (thanks Stefan for mentioning it!) - I don't think it > is reasonable to ask people to go through GRASS to get proper data out > of their QGIS digitized data set. > > Who will take the lead to ask a developer? > > Andreas > > Am 14.11.2013 10:40, schrieb Blumentrath, Stefan: > > Hi George > > > > I had a lot of trouble with those issues too, and I would really love to > see them fixed. In other words, I agree with Giovanni that these are pretty > important bugs for many and not really uncommon digitization use-cases and > would support his categorization as blocker... > > > > However, my current workaround is to ignore the untidiness in the > topology at first and then send the vector dataset through GRASS which is > (at least for not too big datasets) quite convenient using the > Sextante/Processing framework (see also discussion here): > > > http://osgeo-org.1560.x6.nabble.com/Running-v-clean-on-a-QGIS-vector-layer-Polygon-through-Sextante-td5050418.html > > > > Cheers > > Stefan > > > > > > > > ___ > > Qgis-developer mailing list > > Qgis-developer@lists.osgeo.org > > http://lists.osgeo.org/mailman/listinfo/qgis-developer > > > > ___ > Qgis-developer mailing list > Qgis-developer@lists.osgeo.org > http://lists.osgeo.org/mailman/listinfo/qgis-developer > ___ Qgis-developer mailing list Qgis-developer@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/qgis-developer
Re: [Qgis-developer] Bug with avoid intersection - willing to fund fix
> In other words, I agree with Giovanni that these are pretty > important bugs for many and not really uncommon digitization use-cases and > would support his categorization as blocker… I can't remember how many times I blamed the user not to digitize the right way, and at the end discovered that is a qgis issue... that by the way affects not only postGIS, but Spatialite and shapes as well... This bug is critical because among the other things is misleading about how correct can be the work of digitizing topologically correct maps. > However, my current workaround is to ignore the untidiness in the topology > at first and then send the vector dataset through GRASS which is (at least > for not too big datasets) quite convenient using the Sextante/Processing > framework (see also discussion here): cleaning afterwards is certainly a possibility, especially now is easier with processing/sextante, but still it is not an easy solution for the average user. Moreover importing vectors into GRASS data model using the snapping can do miracles, but also can change slightly the geometries, and sometimes this is not an option. having a topological editing toolbox, that will allow to choose the rules beforehand, seems still a long shot so relying on basically the only rule available should be possible o should work. ___ Qgis-developer mailing list Qgis-developer@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/qgis-developer
Re: [Qgis-developer] Bug with avoid intersection - willing to fund fix
Hi, So if there is so much interest in having this issue fixed I assume that there would be plenty out there who would fund the fix! The fix won't happen if noone pays for it - just complaining about this does not help. Get an estimate from some developer (e.g. Sandro Santilli oder Sourcepole) to know how much it would cost to fix. As much as I like GRASS - and it is very good to know that this workaround exists (thanks Stefan for mentioning it!) - I don't think it is reasonable to ask people to go through GRASS to get proper data out of their QGIS digitized data set. Who will take the lead to ask a developer? Andreas Am 14.11.2013 10:40, schrieb Blumentrath, Stefan: > Hi George > > I had a lot of trouble with those issues too, and I would really love to see > them fixed. In other words, I agree with Giovanni that these are pretty > important bugs for many and not really uncommon digitization use-cases and > would support his categorization as blocker... > > However, my current workaround is to ignore the untidiness in the topology at > first and then send the vector dataset through GRASS which is (at least for > not too big datasets) quite convenient using the Sextante/Processing > framework (see also discussion here): > http://osgeo-org.1560.x6.nabble.com/Running-v-clean-on-a-QGIS-vector-layer-Polygon-through-Sextante-td5050418.html > > Cheers > Stefan > > > > ___ > Qgis-developer mailing list > Qgis-developer@lists.osgeo.org > http://lists.osgeo.org/mailman/listinfo/qgis-developer > ___ Qgis-developer mailing list Qgis-developer@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/qgis-developer
Re: [Qgis-developer] Bug with avoid intersection - willing to fund fix
Hi George I had a lot of trouble with those issues too, and I would really love to see them fixed. In other words, I agree with Giovanni that these are pretty important bugs for many and not really uncommon digitization use-cases and would support his categorization as blocker... However, my current workaround is to ignore the untidiness in the topology at first and then send the vector dataset through GRASS which is (at least for not too big datasets) quite convenient using the Sextante/Processing framework (see also discussion here): http://osgeo-org.1560.x6.nabble.com/Running-v-clean-on-a-QGIS-vector-layer-Polygon-through-Sextante-td5050418.html Cheers Stefan ___ Qgis-developer mailing list Qgis-developer@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/qgis-developer
Re: [Qgis-developer] Bug with avoid intersection - willing to fund fix
Any comments from the QGis crowd? Thanks On Sat, Nov 9, 2013 at 9:18 AM, Sandro Santilli wrote: > On Thu, Nov 07, 2013 at 04:26:32PM -0200, George Silva wrote: > > QGis does the explicit noding of the first adjancent feature, but the > > results are inconsistent. > > Noding should involve all nearby features, updating each of them > to include the new node. > > > Sometimes it works, sometimes it doesnt. I've tried many things to see > if I > > could find a reproducible series of steps, but without success. > > I bet, it's not easy to "manually" put points in critical spots > of a floating point grid ! > > --strk; > -- George R. C. Silva SIGMA Consultoria http://www.consultoriasigma.com.br/ ___ Qgis-developer mailing list Qgis-developer@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/qgis-developer
Re: [Qgis-developer] Bug with avoid intersection - willing to fund fix
On Thu, Nov 07, 2013 at 04:26:32PM -0200, George Silva wrote: > QGis does the explicit noding of the first adjancent feature, but the > results are inconsistent. Noding should involve all nearby features, updating each of them to include the new node. > Sometimes it works, sometimes it doesnt. I've tried many things to see if I > could find a reproducible series of steps, but without success. I bet, it's not easy to "manually" put points in critical spots of a floating point grid ! --strk; ___ Qgis-developer mailing list Qgis-developer@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/qgis-developer
Re: [Qgis-developer] Bug with avoid intersection - willing to fund fix
On Thu, Nov 7, 2013 at 6:26 PM, George Silva wrote: > QGis does the explicit noding of the first adjancent feature, but the > results are inconsistent. > > Sometimes it works, sometimes it doesnt. I've tried many things to see if I > could find a reproducible series of steps, but without success. I confirm. In one of the tickets I attached a screencast that shows that is seems pretty random https://www.dropbox.com/s/6elzwbwxcml1sck/postgis_editing_error.mp4 > Right now we are comparing the results ArcGIS is giving us against QGIS > topology checker. It looks like QGIS topology checker is finding a lot more > errors (I think QGIS is correct), but the main difference is visualization > of the errors (which is another deal). I also would like to stress that this is a really serious issue. In QGIS we don't have (yet) tools to define the topology rules *before* starting digitizing as other software does (see the other thread on this matter). If the "avoid intersection" rule fails, we loose the basically the only weapon we have and vectors need to be cleaned with the "usual" tools (GRASS, postgis). ___ Qgis-developer mailing list Qgis-developer@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/qgis-developer
Re: [Qgis-developer] Bug with avoid intersection - willing to fund fix
QGis does the explicit noding of the first adjancent feature, but the results are inconsistent. Sometimes it works, sometimes it doesnt. I've tried many things to see if I could find a reproducible series of steps, but without success. Right now we are comparing the results ArcGIS is giving us against QGIS topology checker. It looks like QGIS topology checker is finding a lot more errors (I think QGIS is correct), but the main difference is visualization of the errors (which is another deal). Any ideas? On Thu, Nov 7, 2013 at 3:36 PM, Sandro Santilli wrote: > On Thu, Nov 07, 2013 at 06:28:47PM +0100, kimaidou wrote: > > Hi > > > > I have seen another "bug" or misbehaviour IMHO : I would expect QGIS to > > automatically create a node at the intersection of 2 linestring when > > topology editing is ON, with no need to do it manually. This would help a > > lot too. > > +1 > > GEOS-3.4.0 introduced a GEOSNode interface that could be handy for that. > > Explicit noding would also probably help fixing other issues, > which I suspect are due to the conceptually wrong assumption > that you can snap a point to a line: > http://lists.osgeo.org/pipermail/geos-devel/2013-November/006720.html > > --strk; > ___ > Qgis-developer mailing list > Qgis-developer@lists.osgeo.org > http://lists.osgeo.org/mailman/listinfo/qgis-developer > -- George R. C. Silva SIGMA Consultoria http://www.consultoriasigma.com.br/ ___ Qgis-developer mailing list Qgis-developer@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/qgis-developer
Re: [Qgis-developer] Bug with avoid intersection - willing to fund fix
On Thu, Nov 07, 2013 at 06:28:47PM +0100, kimaidou wrote: > Hi > > I have seen another "bug" or misbehaviour IMHO : I would expect QGIS to > automatically create a node at the intersection of 2 linestring when > topology editing is ON, with no need to do it manually. This would help a > lot too. +1 GEOS-3.4.0 introduced a GEOSNode interface that could be handy for that. Explicit noding would also probably help fixing other issues, which I suspect are due to the conceptually wrong assumption that you can snap a point to a line: http://lists.osgeo.org/pipermail/geos-devel/2013-November/006720.html --strk; ___ Qgis-developer mailing list Qgis-developer@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/qgis-developer
Re: [Qgis-developer] Bug with avoid intersection - willing to fund fix
Hi I have seen another "bug" or misbehaviour IMHO : I would expect QGIS to automatically create a node at the intersection of 2 linestring when topology editing is ON, with no need to do it manually. This would help a lot too. Regards Michael 2013/11/7 Giovanni Manghi > > We have an important QGis digitizing issue that is blocking and > generating > > a lot of work for us. > > > > The issue has been described in more then 3 bugs ( > > http://hub.qgis.org/issues/8174, http://hub.qgis.org/issues/4880, and > > others) and it really hurts QGis image as a software. > > > > This is something that cannot happen. In our current setup we have 10 GIS > > Analysts working on a large area, doing land use maps. This problem is > > killing us! > > > > I'm cross posting in GEOS dev list to see if there is interest in fixing > > this bug, one way or the other and our company is willing to fund part > (or > > depending how much, all the costs) of the development of this fix. > > > > Please, let me know. > > Hi, > > thanks for raising this issue. I have tried to do so weeks ago and > didn't get much feedback even if this bug is a real show stopper for > certain use cases (that after all are pretty common). > > cheers! > ___ > Qgis-developer mailing list > Qgis-developer@lists.osgeo.org > http://lists.osgeo.org/mailman/listinfo/qgis-developer > ___ Qgis-developer mailing list Qgis-developer@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/qgis-developer
Re: [Qgis-developer] Bug with avoid intersection - willing to fund fix
> We have an important QGis digitizing issue that is blocking and generating > a lot of work for us. > > The issue has been described in more then 3 bugs ( > http://hub.qgis.org/issues/8174, http://hub.qgis.org/issues/4880, and > others) and it really hurts QGis image as a software. > > This is something that cannot happen. In our current setup we have 10 GIS > Analysts working on a large area, doing land use maps. This problem is > killing us! > > I'm cross posting in GEOS dev list to see if there is interest in fixing > this bug, one way or the other and our company is willing to fund part (or > depending how much, all the costs) of the development of this fix. > > Please, let me know. Hi, thanks for raising this issue. I have tried to do so weeks ago and didn't get much feedback even if this bug is a real show stopper for certain use cases (that after all are pretty common). cheers! ___ Qgis-developer mailing list Qgis-developer@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/qgis-developer
