| 05:22:25 | mib_t03vos: | hi there |
| 05:23:04 | mib_t03vos: | anybody here uses MapGuide ? |
| 05:23:26 | mib_t03vos: | I've got some questions about layers |
| 05:24:44 | ominiverdi: | mib_t03vos: have you tried #mapguide ? |
| 05:24:55 | mib_t03vos: | no |
| 05:25:00 | mib_t03vos: | now I try :- ) |
| 05:25:03 | ominiverdi: | :- ) |
| 05:31:40 | mib_t03vos: | nobody is listening there |
| 05:31:42 | mib_t03vos: | :-( |
| 05:31:53 | mib_t03vos: | anybody here uses MapGuide? |
| 05:32:01 | mib_t03vos: | or any other map server.. |
| 05:32:08 | FrankW: | I'm afraid the MapGuide user community is not very reachable by IRC. |
| 05:32:16 | FrankW: | I think the mailing list is your best bet. |
| 05:32:24 | mib_t03vos: | yes |
| 05:32:25 | FrankW: | The mapserver community on the other hand is reasonably accessable in #mapserver. |
| 05:32:38 | mib_t03vos: | I wrote on the mailing list too |
| 05:33:35 | mib_t03vos: | on #mapserver there is nobody |
| 05:34:02 | FrankW: | I'm on #mapserver, so I know that is wrong. |
| 05:34:32 | FrankW: | I see over two dozen nick's in the channel, though many not be listening closely. |
| 05:34:33 | mib_t03vos: | ok |
| 05:34:39 | mib_t03vos: | true |
| 05:34:50 | mib_t03vos: | i'm seeing |
| 11:19:45 | tbowden: | anyone know any free tools for building concave hulls ( not covex. Rather this: http://ubicomp.algoritmi.uminho.pt/local/concavehull.html )? |
| 11:19:47 | sigq: | Title: Concave Hull vs. Convex Hull ( at ubicomp.algoritmi.uminho.pt ) |
| 11:20:25 | tbowden: | there is a published paper on it here: http://repositorium.sdum.uminho.pt/handle/1822/6429 but it is restricted for some reason |
| 11:20:29 | sigq: | Title: Universidade do Minho: Concave hull : a k-nearest neighbours approach for the computation of the region occupied by a set of points ( at repositorium.sdum.uminho.pt ) |
| 11:21:23 | crschmidt: | hm |
| 11:21:29 | crschmidt: | sounds sort of like what schuyler's tool does |
| 11:21:44 | crschmidt: | http://code.flickr.com/blog/tag/clustr/ |
| 11:21:45 | sigq: | Title: Code: Flickr Developer Blog » clustr ( at code.flickr.com ) |
| 11:22:04 | crschmidt: | http://code.flickr.com/svn/trunk/clustr/ |
| 11:22:05 | sigq: | Title: Revision 461: /trunk/clustr ( at code.flickr.com ) |
| 11:25:26 | tbowden: | looks like it does this is one way or another. looking deeper |
| 11:25:36 | tbowden: | thanks crschmidt |
| 11:25:58 | : | * crschmidt loops funrollingly |
| 11:30:39 | crschmidt: | a quick example does: |
| 11:30:40 | crschmidt: | http://crschmidt.net/~crschmidt/clustr.png |
| 11:31:21 | tbowden: | yes, that's it. |
| 11:33:11 | crschmidt: | http://crschmidt.net/~crschmidt/clustr2.png |
| 11:33:14 | crschmidt: | is another example |
| 11:33:55 | crschmidt: | that's a kinda handy little tool |
| 11:35:29 | tbowden: | discussing it on the postgis user list |
| 11:35:42 | : | * crschmidt guesses he isn't on that |
| 11:35:50 | tbowden: | http://postgis.refractions.net/pipermail/postgis-users/2008-November/021981.html |
| 11:35:51 | sigq: | Title: [postgis-users] Get concave hull from an array of longitude inlatitude points ( at postgis.refractions.net ) |
| 11:36:38 | crschmidt: | yeah, Rob Tester's response is the right one |
| 11:36:44 | crschmidt: | Schuyler is just using CGAL for this, I think |
| 11:37:20 | tbowden: | oh, ok. I didn't see that capability from a quick look at the cgal stuff |
| 11:37:24 | tbowden: | must be a bit blind |
| 11:38:02 | tbowden: | saw convex hull stuff but not anything to weed out the "outside" triangles |
| 11:38:11 | tbowden: | worth another look then |
| 11:38:24 | Schuyler: | it's a bit complicated |
| 11:38:28 | crschmidt: | I could also be wrong : ) |
| 11:38:29 | Schuyler: | and my code doesn't work perfectly |
| 11:38:41 | tbowden: | Oh I don't doubt that Schuyler! |
| 11:38:45 | Schuyler: | tbowden: what do you mean by "outside" triangles? |
| 11:38:54 | Schuyler: | and what's it you're trying to do? |
| 11:39:25 | : | * crschmidt is glad cgal is in debian now |
| 11:39:46 | tbowden: | constrain triangles to the "natural" )for want of a better word ) shape |
| 11:39:49 | Schuyler: | huh that Portuguese code might be more sensible than what I'm doing with CGAL |
| 11:40:14 | Schuyler: | CGAL gives somewhat spurious results with its triangulation mesh |
| 11:40:17 | tbowden: | and not have any between two outlying points that are far apart |
| 11:40:37 | Schuyler: | because all it returns is the mesh |
| 11:40:51 | Schuyler: | so my code traverses the mesh, attempting to build polygons |
| 11:41:02 | Schuyler: | but sometimes it gets dangles and self-intersections |
| 11:41:11 | tbowden: | the page regina pointed to is the result I'm after |
| 11:41:22 | Schuyler: | which are hard to weed out without some kind of backtracking |
| 11:41:55 | tbowden: | building the initial mesh isn't the problem. It's stopping the large outlier triangles |
| 11:42:14 | crschmidt: | phttp://www.cgal.org/Manual/3.3/doc_html/cgal_manual/Alpha_shapes_2/Chapter_main.html is the CGAL docs on their Alpha shape stuff |
| 11:42:16 | sigq: | Title: 2D Alpha Shapes ( at www.cgal.org ) |
| 11:42:47 | Schuyler: | GRAH |
| 11:42:57 | crschmidt: | sorry? thought it might be related/helpful |
| 11:43:04 | Schuyler: | tbowden: yeah, that's part of the problem I had |
| 11:43:04 | tbowden: | yes, that's it crschmidt |
| 11:43:07 | tbowden: | thanks |
| 11:43:20 | Schuyler: | the Portugeuse code is implemented in Mathematica |
| 11:43:21 | Schuyler: | bluh |
| 11:43:24 | crschmidt: | heh |
| 11:43:30 | tbowden: | yes, and it's restricted access |
| 11:43:43 | Schuyler: | perhaps their algo is described in one of their publications. |
| 11:43:47 | tbowden: | so far I don't think anyone's asked them for it |
| 11:43:52 | Schuyler: | if it is, I would be willing to write a C++ implementation. |
| 11:43:55 | tbowden: | the publication is also restricted |
| 11:44:00 | Schuyler: | fools! |
| 11:44:25 | tbowden: | publication: http://repositorium.sdum.uminho.pt/handle/1822/6429 |
| 11:44:27 | sigq: | Title: Universidade do Minho: Concave hull : a k-nearest neighbours approach for the computation of the region occupied by a set of points ( at repositorium.sdum.uminho.pt ) |
| 11:44:51 | Schuyler: | anyway, tbowden, if you could help me figure out how to eliminate the dangles and self-intersections from the Clustr code, that would be *awesome* |
| 11:44:52 | crschmidt: | looks like they have an email based request form |
| 11:45:02 | tbowden: | I suspect someone with academic qualifications would have the best chance of convincing them to to release it |
| 11:45:10 | Schuyler: | I can deal with the implementation, I just haven't been able to come up with a solution |
| 11:45:30 | crschmidt: | some googling also recommends http://www.qhull.org/ |
| 11:45:31 | sigq: | Title: Qhull code for Convex Hull, Delaunay Triangulation, Voronoi Diagram, and Halfspace Intersection about a Point ( at www.qhull.org ) |
| 11:45:32 | Schuyler: | all my code does is mesh traversal plus reference counting ( as it were ) |
| 11:45:34 | crschmidt: | not sure if that's also relevant |
| 11:45:35 | tbowden: | looks like between you and them there is the full ( potential ) package |
| 11:46:32 | Schuyler: | crschmidt: qhull doesn't seem to do anything CGAL doesn't, fwiw |
| 11:46:32 | crschmidt: | ( qhull and a few others mentioned in http://ask.metafilter.com/65636/Join-some-of-the-dots-throw-the-rest-away ) |
| 11:46:35 | sigq: | Title: Join some of the dots, throw the rest away | Ask Metafilter ( at ask.metafilter.com ) |
| 11:46:36 | : | * crschmidt nods |
| 11:46:49 | tbowden: | qhull looks to be convex, but not concave. seems that's a much harder problem to solve |
| 11:47:11 | crschmidt: | oh, duh, srry, i can't read : ) |
| 11:47:30 | tbowden: | ok, neither can I most of the time! |
| 11:48:56 | tbowden: | all the hist seem to max out on convex hull. Concave hull isn't a ready solution |
| 11:49:25 | tbowden: | the Universidade do Minho crowd seem to be the leaders here |
| 11:49:54 | Schuyler: | yep |
| 11:50:08 | Schuyler: | it sounds like there's something useful to be built here |
| 11:51:48 | Schuyler: | well, I requested a copy of their paper. |
| 11:52:36 | Schuyler: | tbowden: Anyway, the problem with the mesh traversal is that my code decides to traverse an edge based on how many times it's already touched the vertex on the other side. |
| 11:53:10 | Schuyler: | It doesn't retain any state beyond that. The problem is that if traversing the vertex yields a non-simple polygon, there's no way to back up and undo. |
| 11:53:39 | tbowden: | rob tester on postgis-users has also pointed to http://www.cgal.org/Manual/3.3/doc_html/cgal_manual/Alpha_shapes_2/Chapter_main.html |
| 11:53:42 | sigq: | Title: 2D Alpha Shapes ( at www.cgal.org ) |
| 11:53:53 | Schuyler: | tbowden: do you personally need this for something? or are you just trying to answer someone else's question? |
| 11:54:01 | tbowden: | so maybe we already have good code without knowing it |
| 11:54:10 | Schuyler: | no, trust me |
| 11:54:14 | tbowden: | more a problem I've had in the past and someone else has asked |
| 11:54:19 | Schuyler: | I spent a couple weeks fighting with CGAL and trying to make sense of it |
| 11:54:33 | tbowden: | oh, so cgal not so easy then |
| 11:54:40 | Schuyler: | no. but it does work. |
| 11:54:46 | Schuyler: | FSVO "work" |
| 11:54:53 | Schuyler: | I've done a lot of thinking about this |
| 11:54:55 | tbowden: | ah! |
| 11:55:10 | tbowden: | proly in a large crowd there... |
| 11:55:15 | Schuyler: | I'm not on that list but if you would be so kind as to chime in and point them at the code that crschmidt pasted and perhaps send them my way |
| 11:55:20 | Schuyler: | that would be appreciated |
| 11:55:25 | tbowden: | np |
| 12:27:06 | tbowden: | pdf on alpha shapes by one of the cgal dev's: http://www.inf.ethz.ch/personal/fischerk/pubs/as.pdf Might be of interest to anyone wanting to work on concave hulls |
| 12:32:26 | CameronShorter: | jmckenna, I'm reviewing foss4g budget. We have an item for 10 computers for delegates to log into during the conference. Was this provided at other conferences? I'm inclined to suggest that people will have their own laptops. |
| 12:38:38 | tbowden: | CameronShorter: I don't think so, at least I can't recall there being any, except what may have been at the osgeo both which certainly wasn't 10! |
| 12:39:09 | CameronShorter: | ok, thanks tbowden. |
| 12:39:27 | jmckenna: | CameronShorter: i agree with tim. nice-to-have but not needed |
| 12:40:16 | tbowden: | I suspect one or two at most for that purpose would be plenty. If we get our target audience we may have a number of govt bodies who don't normally come equiped with build in laptops who may want to use such a facility though |
| 12:40:17 | CameronShorter: | cool, jmckenna, I'll reccommend removing from the budget. ( it is costing us $20/delegate ) |
| 12:40:27 | tbowden: | way too much cost! |
| 12:40:59 | CameronShorter: | We can set up an unofficial access to computers at various stands. |
| 12:41:04 | CameronShorter: | for free |
| 12:41:35 | tbowden: | if we have a demo pc ( big screen ) at the osgeo booth like in Cape Town then that should fill that need |
| 12:41:42 | CameronShorter: | I'm offline for a bit - running for a bus. |
| 12:41:49 | tbowden: | k |
| 12:43:13 | tbowden: | ppl without a laptop can check webmail etc on the osgeo linux computer. Might be a revelatory experience for some! ( What, no windows?! ) |
| 12:43:33 | tbowden: | running the live dvd of course |
| 12:45:14 | TylerM: | in 2007 we had about 4 PCs set up |
| 12:45:17 | TylerM: | plus 2 in the booth |
| 12:45:24 | : | * tbowden remembers speculating at Cape Town that one difference between there and Victoria was that in Victoria there were probably more wifi enabled devices than people. |
| 12:45:36 | TylerM: | they were just extras ordered from the workshop PCs I think |
| 12:54:45 | FrankW: | CameronShorter: were you willing to be on the OSGeo conference committee? |
| 12:55:06 | FrankW: | Venka is suggesting you for a voting position. |
| 13:00:50 | jmckenna: | darn |
| 13:02:08 | jmckenna: | i will respond on list |
| 13:02:25 | jmckenna: | but, anyway, seems i'm too late |
| 13:05:36 | Schuyler: | tbowden: you see how the alpha complex diagrams in that PDF don't actually correspond to simple polygons :/ |
| 13:20:40 | tbowden: | not really tried to follow it... |
| 13:21:54 | tbowden: | I figure I've pointed the right people at the right pieces and I'll see what washes up for now. Enough probs already on my list... |
| 13:23:02 | CameronShorter_: | FrankW, I think it is appropriate that I am involved in the OSGeo conference committee, based on my foss4g 2009 committments. I'm prepared to get involved in 2009 activities, but will have little time for anything else. |
| 13:24:08 | strk: | FrankW: have you had a look at the WKT raster idea from Pierre Ravine ? |
| 13:36:02 | jmckenna: | CameronShorter_: i assumed that you have enough on your plate for now, until 2009 is done..and then you would have time for the osgeo conference committee. is that also what you were thinking? |
| 13:38:24 | CameronShorter_: | jmckenna, I will provide low level support and advice to future conferences based on our experiences here in Sydney. but probably won't contribute as much as I would expect from a committee member. |
| 13:39:00 | CameronShorter_: | I'm already spreading myself more thinly than I can keep up with regarding OSGeo. |
| 13:40:10 | jmckenna: | amen. i would like to use your feedback, but i think it's best if this position goes to someone else NOT running the next event : ) |
| 13:40:22 | jmckenna: | due to time constraints |
| 13:40:50 | jmckenna: | i will mention this on the conference_dev list, if i have your blessing |
| 13:42:08 | jmckenna: | CameronShorter_: is this ok with you? |
| 13:43:31 | CameronShorter_: | jmckenna, yes, I'd be more comfortable acting as a list monitor than in an official capacity. Noting also that I'm probably not going to be at the next couple of foss4g conferences. |
| 13:43:57 | jmckenna: | ok, understood. thanks for letting me know. |
| 13:44:33 | CameronShorter_: | pleasure. |
| 15:16:16 | pramsey: | hrm? |
| 15:42:16 | tbowden: | pramsey, was that little outburst an indication you're looking to solve the concave hull problem? ;- ) |
| 15:43:12 | pramsey: | not even a little, tbowden, talk about the ultimate heuristical problem |
| 15:43:23 | pramsey: | everyone's idea of what hte "right" answer is will be different |
| 15:43:34 | tbowden: | yup, but sooo needed! |
| 15:44:32 | pramsey: | funny, TIN boundaries is the same place I ran into that problem in real life too |
| 15:44:45 | pramsey: | I'm trying to remember why a simple sweeping algorithm fails. |
| 15:44:53 | tbowden: | I suspect everyone who has done tin's has the same problem... |
| 15:45:23 | pramsey: | mmm, right, just doesn't work : ) |
| 15:45:56 | tbowden: | there is no definitive length delineation between good sides and bad sides |
| 15:46:27 | tbowden: | way too much subjectivity in the answer, but a very time consuming problem that always has to be solved |
| 17:14:18 | Kapivara: | If I have a database w/ spatial and non-spatial data, and I need to render this data AND use the non-spatial data as "semantic anotations" in a map viewer, It makes sense to set up a WFS to access this database? |
| 17:14:42 | Kapivara: | I'm quite lost with "when I should use WFS" |
| 17:31:09 | Kapivara: | guess I solve this point, bye = ) |
| 21:49:23 | CameronShorter: | Is there a OSGeo marketing meeting here in 10 mins? ( as per my calendar reminder ) |
| 21:59:22 | tbowden: | CameronShorter: not sure but I'm here anyway |
| 22:00:05 | CameronShorter: | I don't think anyone is prepared. ( no email was sent out ). |
| 22:02:03 | tbowden: | Maybe we could just propose the entire marketing budget be spent on us, pass the motion and start spending. |
| 22:02:09 | tbowden: | Just an idea ;- ) |
| 22:02:55 | CameronShorter: | Yes, think we need lots of beer. I'll order a slab. |
| 22:05:00 | tbowden: | Oh good, I'm thirsty. Could do with a nice cold king brown about now. |
| 22:11:11 | CameronShorter: | tbowden, not enough action here, I'm heading off home. |
| 22:11:47 | tbowden: | k |
| 22:47:51 | sigq: | osgeofeed: Ticket #319 ( task closed ): please delete private list archive <http://trac.osgeo.org/osgeo/ticket/319#comment:4> |
