Re: FW: [OSGeo-Discuss] Proposal: OSGeo Cartographic Library
Hi, this mail was forwarded to me, and I immediately subscribed to the mailing list to set something straight. Rushforth, Peter wrote: In the XML world, the answer to graphics rendering is SVG + XSL-FO, generated by XSLT scripts according to styling rules. XSL-FO is then rendered via a formatting objects processor, like Apache FOP. OK, but... you do know that this is only true for small datasets. For large datasets FOP is a bad choice performance wise. For instance: if you have to convert an XML to a PDF and you expect the end result to be 50 pages, Apache FOP might be a good choice. If you expect 10,000+ pages, then it's a bad choice. A document with 10,000+ pages is not as unusual as most PDF newbies think. I've just created a PDF with 28,500 pages; each page is a letter to one of the 28,500 students of Ghent University. I've heard from banks that create PDFs with 100,000+ pages. In this case, I use iText (obviously: I'm the developer of iText): http://www.lowagie.com/iText/ iText is used by Google, NASA, DoD,... It's purpose is to create all kinds of PDF files as fast as possible. I'm of the opinion that a pipeline like: GML +/- (any XML data) + SLD/FE + WMS graphics = XSLT = SVG+XSL-FO - pdf, (?geopdf anyone?) XSLT? XSL-FO? What a performance killer! would make for a killer map scripting environment. Plus, it has the added benefit of being based on standards or de facto standards across the board, with open source solutions available in each area. Use Cases: -- I. Interactive Browsing (e.g. web mapping) 1. Good-looking web maps (more control of grid/graticule labeling) FOP doesn't support all PDF features. For instance: I've made a map in PDF that shows a raster image when you first open it. Then when you zoom in, the PDF automatically switches to vector data. FOP doesn't support OCG! iText does! OCG = Optional Content Groups: meaning you can turn on and off layers in a PDF file. For instance: you could add street names to a map in English en French and define the OCG so that only one language at a time is shown. II. Ad-hoc Authoring (one-time GIS style layout using GUI) 1. Good-looking printed (ps,pdf,etc) maps automatically providing grid/grat, scalebar, legend, north arrow, SRS description - provide an API to exiting GIS apps? FOP isn't fast enough to create detailed PDFs on the fly. III. Automated Mapping (script driven) 1. Map Series (single page, identical layout) 2. Map Atlas (mostly map with some text, multi-page) 3. Map-centric documents (mostly text with some map, multi-page) 4. Route Alignment Sheets (rotated (non 90 deg) to fit page) With iText you have direct access to all PDF operators and operands. you can draw every shape you want. Please download chapter 1 of my book: http://1t3xt.com/docs/book.php#free There's an example of a map of a (fictive) city. I hope you'll find it very interesting! br, Bruno ___ Discuss mailing list Discuss@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/discuss
Re: [OSGeo-Discuss] Proposal: OSGeo Cartographic Library
Dave Patton wrote: Actually, dealing with the legend, or the contents of the entire map collar, is non-trivial, however, a library that can render the content within a map's neat line has all the tools necessary to be able to render the map's collar. The difference may be that there would need to be some additional functionality, or different methods of calling the same underlying functionality, in order to make the tasks involved in composing the collar easy. If you make the map in PDF, you could provide the legend as a floating annotation that can be moved around by the user. If you are using OCG, you also have the Optional Content panel that can be used as a legend and that can be used to make certain layers visible/invisible. All this is fairly easy to achieve in PDF. br, Bruno ___ Discuss mailing list Discuss@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/discuss
Re: [OSGeo-Discuss] Proposal: OSGeo Cartographic Library
Tyler Mitchell (OSGeo) wrote: For me the question of PDF is not limited by number of pages, or options that PDF offers, but whether or not it is an efficient format for sharing cartographic information - for which, so far, I'd say it fails miserably. How many times have I went to a municipal mapping site only to find their maps are all in PDF - what a pain! It might just be me though :) It depends on how they are made. Do they have a street index that allows you to jump to the exact location when you click on a streetname? Are they made out of raster images converted to PDF or are they drawn using vector data? Also, in the operational/industrial GIS map production environments I've been in, we've needed easy ways to print and re-print maps without having to open a viewer (and our printers didn't support PDF natively - though I assume some do now). To print, I've focused on native plotter file formats and/or Postscript since most plotters can support it. Of course this isn't going to be good for web distribution. There are ways to work around that print problem ;-) For delivering digital files, I've often converted the PS files into PDF but it's been far from ideal. But then you get a 'flat' PDF without any interactivity. I don't see any added value when you convert PS to PDF. I increasingly believe that web-based tools are going to be the only option. So what about off-line delivery? A CD or USB runable system is an interesting and more effective way than a PDF in some cases, though of course both have some memory overhead issues. So, is it just me or does stuffing a 1:20,000 topo map into a PDF makes a huge file that is virtually unusable unless you have gigabytes of RAM and dual processors. Delivering a 40MB PDF to client who is running an old computer doesn't bode well for your service ;-) Er... stuffing a 1:20,000 topo map. The fact that you mention 1:20,000 indicates that you are probably talking about raster images, not about vector data. If you write the vector data to a PDF, all the data is compressed. You get really small file sizes when compared to other solutions. In short: the major problem with PDF in the GIS world is a lack of understanding of the Portable Document Format by people who are specialized in GIS. Of course PDF sucks if you just stuff if with raster images or use a PDF that was converted from PS. Even a FOP generated PDF has no added value. As soon as I have the time, I'll make you some examples. br, Bruno ___ Discuss mailing list Discuss@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/discuss
Re: [OSGeo-Discuss] Proposal: OSGeo Cartographic Library
Jachym Cepicky wrote: www.openstreetmap.org The site isn't working for me. While trying to retrieve the URL: http://wiki.openstreetmap.org/ The following error was encountered: * Read Error The system returned: (104) Connection reset by peer An error condition occurred while reading data from the network. Please retry your request. All I need is a zip file with limited data of, for instance, a city. Any city will do. As long is I just get the raw vector data and a small explanation of how to use it. With Open Street map, I don't have access to the help/wiki and my first impression of the (blank) home page is that I'll have to compile my own data set. I don't have time for that. I don't ask you being 'PDF newbies' to read the PDF Reference (1,200+ pages), but I'll be happy to explain every feature that could be useful in a GIS application. In return I ask only one favor: please give me some data I can use without doing any extra GIS effort. I'm willing to write some Java code for free, but I'm not a GIS developer (haven't done any GIS development in the last 9 years). best regards, Bruno ___ Discuss mailing list Discuss@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/discuss
Re: [OSGeo-Discuss] Proposal: OSGeo Cartographic Library
Paolo Cavallini wrote: All this does not show off in evince, however. Evince is as they promote it simply a document viewer. It's not a full blown PDF viewer (well, maybe it is for most users, but for me as an avid PDF user it isn't). Evince doesn't support all the features that are explained in the PDF Reference manual. OCG was introduced in PDF 1.5 (dating back from 2003). You will see the OCG layers in Adobe Reader starting with version 6.0. br, Bruno ___ Discuss mailing list Discuss@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/discuss
Re: [OSGeo-Discuss] Proposal: OSGeo Cartographic Library
Paolo Cavallini wrote: So the question is: is there a free (as speech) and Open Source full blown PDF viewer? I think OSGeo is about open source software. I'm a F/OSS developer, so I'm going to use my standard answer: What's keeping you from adding OCG support to Evince? *LOL* My own library iText is 'free' as in 'free beer'; although it is available under the MPL/LGPL it's supposed to be free as in 'free speech' too, but I don't agree with the FSF argument that Licensed Software (any license: GPL, LGPL, MPL,...) is free as in 'free speech'. If it were, companies using iText wouldn't keep on harassing me about the license (and about the fact that they don't like the MPL/LGPL). If iText were really Free, they wouldn't have any reason for not using iText. I interpret the 'Free' in 'Free Software' as 'Free to pay for value' (Google for 'Voluntary Economies'). But than again: companies like Google for instance use iText in many different project but they've never paid anything for using it. That's why when people lecture me about 'Free Software' annoy me: most of them USE F/OSS for free (as in Free Beer), but the number of people CONTRIBUTING (*) source or PAYING for value is very limited. (*) By the way: this is the list of people who contributed code to the iText project: http://www.1t3xt.com/about/acknowledgments/index.php br, Bruno ___ Discuss mailing list Discuss@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/discuss
Re: [OSGeo-Discuss] Your open source career
Landon Blake wrote: The lack of good user documentation is a weakness of many open source projects. The problem is that most of us like to code, but few of us like to write! Speak for yourself! ;-) Am I the exception to the rule? *LOL* Please don't regard the following as shameless promo. I just want to share a very interesting experience with future F/OSS writers. I'm an Open Source developer ( http://www.lowagie.com/iText/ ) but I also like to write. In 2004, I took some time off from my day job to write a free online tutorial for iText because the lack of proper documentation was a problem that had to be addressed (you're right about that, Landon). Once the first pages of the free tutorial were online, I immediately received an offer to write a book about iText, first from O'Reilly, later on also from Manning. After long consideration, I decided to try writing a book for Manning Publications Co. because they have the reputation that they are very demanding. You may think I'm a masochist, but I thought that would be the best guarantee to write a good book. And it was! I talked with some authors who claimed that writing their first book for Manning was a good choice. In hindsight, I agree, although I might choose for O'Reilly next time ;-) I spent 3 months writing the book proposal (full TOC included). 6 months writing the manuscript. After these 9 months of hard labor, another 9 months were needed to get the book ready for production (copy editing, proof reading, making the index,...). The result is: http://www.1t3xt.com/docs/book.php Want to know how much I earned? No problem! Have a look at my Quarterly overviews here: http://www.lowagie.com/maand.php?year=2008month=4#806 The revenue listed is limited to the Royalties. You don't get rich from writing a book, but I also have indirect revenue from sales (when people buy the book after clicking a link on my site). I get 10% Royalties and if you study the Quarterly Overviews, you'll see that the sum I get for each book varies depending on many factors (time, location,...). I get between 5% (Amazon) and 15% (Manning) for selling the book using a link on my site. But it's not only about the money: the product has gotten much more attention and it has really boomed! Having a book is (almost) a guarantee for success for every F/OSS project. If you are planning to write a book, and you want an introduction at Manning Publications; or if you just want to talk about starting such a venture, let me know, and we'll chat. I know plenty of people who dream of writing a book, how I would like to persuade them that they should just start writing. I like to quote Henry David Thoreau: If you have built castles in the air, your work need not be lost; that is where they should be. Now put the foundations under them. best regards (and please pardon my enthousiasm), Bruno Lowagie ___ Discuss mailing list Discuss@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/discuss
Re: [OSGeo-Discuss] Open Source development metrics
Jody Garnett wrote: Try for a couple of the metrics you use when evaluating open source projects If a project or F/OSS developer has been around for a while, there are bound to be online metrics about it/him/her, for instance on Ohloh. Some examples: http://www.ohloh.net/projects/305 http://www.ohloh.net/accounts/blowagie Sure, this is just an indication based on an algorithm that takes many parameters into account, but indications such as increasing year-over-year development activity and maturity can't be faked. br, Bruno ___ Discuss mailing list Discuss@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/discuss
Re: [OSGeo-Discuss] Open Source development metrics
Jody Garnett wrote: Just a word of warning; those sites get tripped up; you can see the history for geotools is missing because we just cleaned up our subversion repository a bit. True: in my personal history, it's as if I have been inactive for years, but in reality all the CVS data disappeared when we switched from CVS to SVN. br, Bruno ___ Discuss mailing list Discuss@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/discuss