Re: [gdal-dev] Questions about projects of GSoC 2009
Hello Klokan, Thanks for encouraging me! And congratulations for your excellent job with gdal2tiles. It really saved me from a lot of unproductive hours... It would be great to help you with WMTS implementation, or work with PNG driver, or with WKT Raster driver. I'm going to apply for the three projects directly on the GSoC page. Frank Warmerdarm told me on the GSoC list that it was the correct way. Am I right? Best regards Jorge On Tue, Mar 31, 2009 at 7:33 PM, Klokan Petr Přidal wrote: > Jorge, > > I am glad you are interested in the PNG driver, which I suggested as a > project for Summer of Code... > It would be really great to have this in GDAL and I guess more > projects will profit from this... > > I am going to assist as much as I can, whether in gsoc program > timeframe or even otherwise. > Feel free to contact me with any questions related to this project... > > I guess for position of mentor there are better candidates then me > (like Frank, Even, Mateusz, ...) > > Because I am a PhD student I applied in this Summer of Code as well as > a student so I technically I can not be your mentor. I can work this > summer on the WMTS implementation and other GDAL2Tiles improvements... > but I don't want to block position of new candidates... > If you are interested in WMTS implementation and GDAL2Tiles then just > go ahead, this I can mentor for sure ;-) > > Best regards, > > Klokan > > 2009/3/30 Mateusz Loskot : > > Jorge Arévalo wrote: > >> The entire process could be improved by several ways: > >> > >> - I would like to have an enhaced PNG driver, to reduce the size of the > >> tiles. For this reason, I had a theoretical approach to NeuQuant > >> algorithm and I used pngnq. I discovered the relation between this > >> algorithm and Kohonen Neural Networks, that I studied during my degree, > >> and I liked a lot. Basically, for these reasons I'm interested in the > >> project about PNG Driver > > > > Nice idea. > > > >> - On one hand, at first I was using raster data, instead of vector data. > >> When the need of merging data arose, I had to change the point and use > >> vector data + Postgis. On the other hand, I would like to integrate my > >> layer to another ones in a bigger context, and I think that the raster > >> support on Postgis will be a really useful thing. For these reasons, I'm > >> interested in the project about WKT Raster. > > > > Good. > > > >> - Before using gdal2tiles, I made my own "map tiler", creating the tiles > >> needed "on-the-fly". As thought, it didn't work. I tried to use my own > >> "tile cache system". One problem was the low rate of cache hints. Then, > >> I discovered the WMS standard, and recently, the coming WMTS standard, > >> even better. For this reason, I'm interested in the project about WMTS > >> implementation. > > > > Good as well. > > > >> Now, this is my "minimum schedule". This is, the minimum amount of time > >> that I can use this summer. With high probability, I will be able to use > >> more time, but I prefer to put myself in the worst situation: > >> > >> May 23 - June 30: 22h per week (total 118h) > >> July 1 - July 31: 45h per week (total 207h) > >> August 1 - August 10: 22h per week (total 26h) > > > > It should be more than enough. > > > >> Finally, I have experience with C, C++, Java and Python languages. My > >> favourite one is C. I made the most of my practice work during degree in > C. > > > > Good. > > > >> Do you think, with this provisional schedule in mind, that I can apply > >> for one of the projects? > > > > Yes, I'm sure you are able to do it. > > > >> If not, would it be possible to find a mentor for one of them? > >> Without paying and GSoC deadlines, of course. > > > > I'm sure you can try, it won't hurt :-) > > > > Best regards, > > -- > > Mateusz Loskot, http://mateusz.loskot.net > > Charter Member of OSGeo, http://osgeo.org > > ___ > > gdal-dev mailing list > > gdal-dev@lists.osgeo.org > > http://lists.osgeo.org/mailman/listinfo/gdal-dev > > > > > > -- > http://blog.klokan.cz/ > http://www.maptiler.org/ > http://www.oldmapsonline.org/ > ___ gdal-dev mailing list gdal-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/gdal-dev
Re: [gdal-dev] Questions about projects of GSoC 2009
Jorge, I am glad you are interested in the PNG driver, which I suggested as a project for Summer of Code... It would be really great to have this in GDAL and I guess more projects will profit from this... I am going to assist as much as I can, whether in gsoc program timeframe or even otherwise. Feel free to contact me with any questions related to this project... I guess for position of mentor there are better candidates then me (like Frank, Even, Mateusz, ...) Because I am a PhD student I applied in this Summer of Code as well as a student so I technically I can not be your mentor. I can work this summer on the WMTS implementation and other GDAL2Tiles improvements... but I don't want to block position of new candidates... If you are interested in WMTS implementation and GDAL2Tiles then just go ahead, this I can mentor for sure ;-) Best regards, Klokan 2009/3/30 Mateusz Loskot : > Jorge Arévalo wrote: >> The entire process could be improved by several ways: >> >> - I would like to have an enhaced PNG driver, to reduce the size of the >> tiles. For this reason, I had a theoretical approach to NeuQuant >> algorithm and I used pngnq. I discovered the relation between this >> algorithm and Kohonen Neural Networks, that I studied during my degree, >> and I liked a lot. Basically, for these reasons I'm interested in the >> project about PNG Driver > > Nice idea. > >> - On one hand, at first I was using raster data, instead of vector data. >> When the need of merging data arose, I had to change the point and use >> vector data + Postgis. On the other hand, I would like to integrate my >> layer to another ones in a bigger context, and I think that the raster >> support on Postgis will be a really useful thing. For these reasons, I'm >> interested in the project about WKT Raster. > > Good. > >> - Before using gdal2tiles, I made my own "map tiler", creating the tiles >> needed "on-the-fly". As thought, it didn't work. I tried to use my own >> "tile cache system". One problem was the low rate of cache hints. Then, >> I discovered the WMS standard, and recently, the coming WMTS standard, >> even better. For this reason, I'm interested in the project about WMTS >> implementation. > > Good as well. > >> Now, this is my "minimum schedule". This is, the minimum amount of time >> that I can use this summer. With high probability, I will be able to use >> more time, but I prefer to put myself in the worst situation: >> >> May 23 - June 30: 22h per week (total 118h) >> July 1 - July 31: 45h per week (total 207h) >> August 1 - August 10: 22h per week (total 26h) > > It should be more than enough. > >> Finally, I have experience with C, C++, Java and Python languages. My >> favourite one is C. I made the most of my practice work during degree in C. > > Good. > >> Do you think, with this provisional schedule in mind, that I can apply >> for one of the projects? > > Yes, I'm sure you are able to do it. > >> If not, would it be possible to find a mentor for one of them? >> Without paying and GSoC deadlines, of course. > > I'm sure you can try, it won't hurt :-) > > Best regards, > -- > Mateusz Loskot, http://mateusz.loskot.net > Charter Member of OSGeo, http://osgeo.org > ___ > gdal-dev mailing list > gdal-dev@lists.osgeo.org > http://lists.osgeo.org/mailman/listinfo/gdal-dev > -- http://blog.klokan.cz/ http://www.maptiler.org/ http://www.oldmapsonline.org/ ___ gdal-dev mailing list gdal-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/gdal-dev
Re: [gdal-dev] Questions about projects of GSoC 2009
Jorge Arévalo wrote: Hello, Thanks for the link Mateusz. OK, I need your opinion. I've been working during 1 year (still working) in a project that uses gdal library (developing some code), and Postgis, to transform vectorial data (MIF files) on PNG tiles to show over google maps. I use Postgis to perform a previous merge of some vectorial data. Then, I transform the vectorial data in raster files, and "burn" the polygons over them using given colors. Finally, I transform these raster files in PNG tiles with gdal2tiles. Guys, Maybe this interests you. The GMT tool "ps2raster" (CVS version) is now able to create KML code from any -- GMT-complicate-as-much-as-you-want -- images. Joaquim Luis ___ gdal-dev mailing list gdal-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/gdal-dev
Re: [gdal-dev] Questions about projects of GSoC 2009
Hello, Thanks for the link, and for your opinion! I'll apply for the three projects. Best regards Jorge 2009/3/30 Mateusz Loskot > Mateusz Loskot wrote: > > Jorge Arévalo wrote: > >> The entire process could be improved by several ways: > >> > >> - I would like to have an enhaced PNG driver, to reduce the size of > >> the tiles. For this reason, I had a theoretical approach to > >> NeuQuant algorithm and I used pngnq. I discovered the relation > >> between this algorithm and Kohonen Neural Networks, that I studied > >> during my degree, and I liked a lot. Basically, for these reasons > >> I'm interested in the project about PNG Driver > > > > Nice idea. > > By the way, here is project developed by a friend of mine > > C++ Kohonen Neural Network Library > > http://knnl.sourceforge.net/ > > Perhaps you will find it useful in your work. > > Best regards, > -- > Mateusz Loskot, http://mateusz.loskot.net > Charter Member of OSGeo, http://osgeo.org > ___ gdal-dev mailing list gdal-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/gdal-dev
Re: [gdal-dev] Questions about projects of GSoC 2009
Mateusz Loskot wrote: > Jorge Arévalo wrote: >> The entire process could be improved by several ways: >> >> - I would like to have an enhaced PNG driver, to reduce the size of >> the tiles. For this reason, I had a theoretical approach to >> NeuQuant algorithm and I used pngnq. I discovered the relation >> between this algorithm and Kohonen Neural Networks, that I studied >> during my degree, and I liked a lot. Basically, for these reasons >> I'm interested in the project about PNG Driver > > Nice idea. By the way, here is project developed by a friend of mine C++ Kohonen Neural Network Library http://knnl.sourceforge.net/ Perhaps you will find it useful in your work. Best regards, -- Mateusz Loskot, http://mateusz.loskot.net Charter Member of OSGeo, http://osgeo.org ___ gdal-dev mailing list gdal-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/gdal-dev
Re: [gdal-dev] Questions about projects of GSoC 2009
Jorge Arévalo wrote: > The entire process could be improved by several ways: > > - I would like to have an enhaced PNG driver, to reduce the size of the > tiles. For this reason, I had a theoretical approach to NeuQuant > algorithm and I used pngnq. I discovered the relation between this > algorithm and Kohonen Neural Networks, that I studied during my degree, > and I liked a lot. Basically, for these reasons I'm interested in the > project about PNG Driver Nice idea. > - On one hand, at first I was using raster data, instead of vector data. > When the need of merging data arose, I had to change the point and use > vector data + Postgis. On the other hand, I would like to integrate my > layer to another ones in a bigger context, and I think that the raster > support on Postgis will be a really useful thing. For these reasons, I'm > interested in the project about WKT Raster. Good. > - Before using gdal2tiles, I made my own "map tiler", creating the tiles > needed "on-the-fly". As thought, it didn't work. I tried to use my own > "tile cache system". One problem was the low rate of cache hints. Then, > I discovered the WMS standard, and recently, the coming WMTS standard, > even better. For this reason, I'm interested in the project about WMTS > implementation. Good as well. > Now, this is my "minimum schedule". This is, the minimum amount of time > that I can use this summer. With high probability, I will be able to use > more time, but I prefer to put myself in the worst situation: > > May 23 - June 30: 22h per week (total 118h) > July 1 - July 31: 45h per week (total 207h) > August 1 - August 10: 22h per week (total 26h) It should be more than enough. > Finally, I have experience with C, C++, Java and Python languages. My > favourite one is C. I made the most of my practice work during degree in C. Good. > Do you think, with this provisional schedule in mind, that I can apply > for one of the projects? Yes, I'm sure you are able to do it. > If not, would it be possible to find a mentor for one of them? > Without paying and GSoC deadlines, of course. I'm sure you can try, it won't hurt :-) Best regards, -- Mateusz Loskot, http://mateusz.loskot.net Charter Member of OSGeo, http://osgeo.org ___ gdal-dev mailing list gdal-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/gdal-dev
Re: [gdal-dev] Questions about projects of GSoC 2009
Hello, Thanks for the link Mateusz. OK, I need your opinion. I've been working during 1 year (still working) in a project that uses gdal library (developing some code), and Postgis, to transform vectorial data (MIF files) on PNG tiles to show over google maps. I use Postgis to perform a previous merge of some vectorial data. Then, I transform the vectorial data in raster files, and "burn" the polygons over them using given colors. Finally, I transform these raster files in PNG tiles with gdal2tiles. The entire process could be improved by several ways: - I would like to have an enhaced PNG driver, to reduce the size of the tiles. For this reason, I had a theoretical approach to NeuQuant algorithm and I used pngnq. I discovered the relation between this algorithm and Kohonen Neural Networks, that I studied during my degree, and I liked a lot. Basically, for these reasons I'm interested in the project about PNG Driver - On one hand, at first I was using raster data, instead of vector data. When the need of merging data arose, I had to change the point and use vector data + Postgis. On the other hand, I would like to integrate my layer to another ones in a bigger context, and I think that the raster support on Postgis will be a really useful thing. For these reasons, I'm interested in the project about WKT Raster. - Before using gdal2tiles, I made my own "map tiler", creating the tiles needed "on-the-fly". As thought, it didn't work. I tried to use my own "tile cache system". One problem was the low rate of cache hints. Then, I discovered the WMS standard, and recently, the coming WMTS standard, even better. For this reason, I'm interested in the project about WMTS implementation. Now, this is my "minimum schedule". This is, the minimum amount of time that I can use this summer. With high probability, I will be able to use more time, but I prefer to put myself in the worst situation: May 23 - June 30: 22h per week (total 118h) July 1 - July 31: 45h per week (total 207h) August 1 - August 10: 22h per week (total 26h) Finally, I have experience with C, C++, Java and Python languages. My favourite one is C. I made the most of my practice work during degree in C. Do you think, with this provisional schedule in mind, that I can apply for one of the projects? If not, would it be possible to find a mentor for one of them? Without paying and GSoC deadlines, of course. Many thanks in advance, anyway Best regards Jorge 2009/3/30 Mateusz Loskot > Jorge Arévalo wrote: > > "Difficult" is a very subjective thing, totally agree. Sorry for the > > approach of the question. > > [...] > > All concepts sound familiar (because > > I've studied or worked with them), but I think that I need some > > previous time to go deeper. I have to include this in my schedule. > > Here are some helpful notes: > > http://code.google.com/opensource/gsoc/2008/faqs.html#0.1_student_time > > GSoc lasts for fixed time period and no more. > > Best regards, > -- > Mateusz Loskot, http://mateusz.loskot.net > Charter Member of OSGeo, http://osgeo.org > ___ gdal-dev mailing list gdal-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/gdal-dev
Re: [gdal-dev] Questions about projects of GSoC 2009
Jorge Arévalo wrote: > "Difficult" is a very subjective thing, totally agree. Sorry for the > approach of the question. > [...] > All concepts sound familiar (because > I've studied or worked with them), but I think that I need some > previous time to go deeper. I have to include this in my schedule. Here are some helpful notes: http://code.google.com/opensource/gsoc/2008/faqs.html#0.1_student_time GSoc lasts for fixed time period and no more. Best regards, -- Mateusz Loskot, http://mateusz.loskot.net Charter Member of OSGeo, http://osgeo.org ___ gdal-dev mailing list gdal-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/gdal-dev
Re: [gdal-dev] Questions about projects of GSoC 2009
Hello, "Difficult" is a very subjective thing, totally agree. Sorry for the approach of the question. Maybe I can raise it in that way: What amount of theoretical/practical knowledge do I need to start working on the project? For example, the first project (PNG driver) suggests me the following initial knowledges (most of them are specified in the web http://trac.osgeo.org/gdal/wiki/SummerOfCode): - PNG format (of course) - C / C++ fluently - Raster, bands - Basic GDAL architecture - NeuQuant algorithm - Kohonen Neural Networks - pngnq - Several tools that optimize PNG The project about WKT raster (the list that you provided me on another mail): - raster - band - pixel and how pixel type influences raster structure - recognize BIP, BIL and BSQ layouts - reading and parsing binary data in C/C++ - encode/decode raw binary data to/from hex in C/C++ - understand idea behind Well-Known-Text and Well-Known-Binary - understand basics of architecture of GDAL - understand basics of PostGIS - how to write good unit test for GDAL driver - how to use GDAL, PostgreSQL and PostGIS manuals and websites All concepts sound familiar (because I've studied or worked with them), but I think that I need some previous time to go deeper. I have to include this in my schedule. I'll finish it tomorrow, and we'll be able to go further on this issue, if you want :-) Thanks for your response Best regards Jorge 2009/3/29 Mateusz Loskot > Jorge Arévalo wrote: > > Hello everybody, > > > > I'm interesed in these projects: > > - PNG Driver > > - WMTS implementation > > - WKT Raster Driver > > Great! > > > I would like to know a little more about them. > > Do you have any specific questions or concerns? > > > Basically, their estimated difficulty. > > Difficulty is a subjective thing. > What is difficult to me, might be easy to you, so estimation > may be lead misleading. > > > I'm very exciting about the possibility of being one of the selected > > OSGeo students for GSoC 2009, but I think that I need to be realistic > > and totally comitted with the project. I'm making a schedule, and I > > would like to know if all projects fit into it. > > And, what's your schedule? > > > For example, the last > > two projects seem to be longer that the first one, but maybe I'm > > wrong. > > Exactly, "seem" is the right word. > > We don't know what's your experience and what you consider as > easy/difficult. So, it's very hard to help you in making your > decision unless you have specific / technical questions. > > Remember, that GSoC projects are not managed in the commercial way. > It's you and mentor who agree that a project is completed or not. > See Q9 and Q10: > > http://code.google.com/opensource/gsoc/2008/faqs.html#0.1_use_code > > So, I'd not worry much about making a very detailed timeline plan. > > For instance WKT Raster driver, as I mentioned in the proposal, the idea > is to develop a *prototype*, something that proofs (or not) the concept > is feasible and implementation may perform well, something that will be > helpful and usable for further development. > > Best regards, > -- > Mateusz Loskot, http://mateusz.loskot.net > Charter Member of OSGeo, http://osgeo.org > ___ gdal-dev mailing list gdal-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/gdal-dev
Re: [gdal-dev] Questions about projects of GSoC 2009
Jorge Arévalo wrote: > Hello everybody, > > I'm interesed in these projects: > - PNG Driver > - WMTS implementation > - WKT Raster Driver Great! > I would like to know a little more about them. Do you have any specific questions or concerns? > Basically, their estimated difficulty. Difficulty is a subjective thing. What is difficult to me, might be easy to you, so estimation may be lead misleading. > I'm very exciting about the possibility of being one of the selected > OSGeo students for GSoC 2009, but I think that I need to be realistic > and totally comitted with the project. I'm making a schedule, and I > would like to know if all projects fit into it. And, what's your schedule? > For example, the last > two projects seem to be longer that the first one, but maybe I'm > wrong. Exactly, "seem" is the right word. We don't know what's your experience and what you consider as easy/difficult. So, it's very hard to help you in making your decision unless you have specific / technical questions. Remember, that GSoC projects are not managed in the commercial way. It's you and mentor who agree that a project is completed or not. See Q9 and Q10: http://code.google.com/opensource/gsoc/2008/faqs.html#0.1_use_code So, I'd not worry much about making a very detailed timeline plan. For instance WKT Raster driver, as I mentioned in the proposal, the idea is to develop a *prototype*, something that proofs (or not) the concept is feasible and implementation may perform well, something that will be helpful and usable for further development. Best regards, -- Mateusz Loskot, http://mateusz.loskot.net Charter Member of OSGeo, http://osgeo.org ___ gdal-dev mailing list gdal-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/gdal-dev
[gdal-dev] Questions about projects of GSoC 2009
Hello everybody, I'm interesed in these projects: - PNG Driver - WMTS implementation - WKT Raster Driver I would like to know a little more about them. Basically, their estimated difficulty. I'm very exciting about the possibility of being one of the selected OSGeo students for GSoC 2009, but I think that I need to be realistic and totally comitted with the project. I'm making a schedule, and I would like to know if all projects fit into it. For example, the last two projects seem to be longer that the first one, but maybe I'm wrong. Thanks in advance Best regards Jorge ___ gdal-dev mailing list gdal-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/gdal-dev
