Collaborating to Put SVG on the Map

Mike McCullough, Ph.D.
Citizen Computing Institute
647 E. 14th Street
New York, New York 10009
(212) 254-3582
mm@CitizenComputing.org

Keywords: SVG advocacy, SVG browser support, open source, data visualization, mapmaking, citizen computing

Abstract

SVG aspires to become an integral part of the daily Web experience. This raises the hurdle for its success very high. This presentation addresses challenges facing widespread adoption of SVG and discusses actions developers can take to help assure SVG's success. These actions include reaching out to developers at user groups and other gatherings, starting SVG user groups, letter-writing campaigns to promote native browser support, participating in online support forums and open source development projects, offering SVG classroom instruction, collaborating on potential SVG killer apps and promulgating SVG use among particular communities. Sample work is demonstrated from the Citizen Mapmakers project, an effort to promulgate SVG among environmentalists and citizen advocates.

I. THE ISSUE OF BECOMING A DE-FACTO STANDARD

I am from the Citizen Computing Institute (CCI) a new non-profit orgranization based in New York City (http://www.CitizenComputing.org). We conduct software research and development for non-profits. We believe that SVG holds great promise as a data visualization and mapmaking tool for citizen groups trying to call public attention to important issues.

I am going to address some of the big picture challenges facing SVG and discuss ways that we as developers can help assure SVG's success [1]. Many of us at this conference are investing time, energy and resources in projects whose ultimate success is bound up with the fate of SVG. What can we do together to help assure that SVG graduates from an official W3C standard to a de facto standard? And, over the long run, how can we facilitate successful open source development of SVG mapping code as well as SVG graphics? Also, how might we promulgate SVG among our own professional communities or special constituencies? This is what we are trying to do at CCI with a project called Citizen Mapmakers. We hope to convince environmental and other citizen action groups of the value of SVG mapmaking. I will present our "roadmap" for promulgating SVG in the hope that it may have broader use and briefly demo how we are trying to do this.

Let me start with a sample of the buzz in the press about SVG. We've all heard this kind of stuff before but it's always nice to hear again. One reviewer declares that SVG is "one of the most exciting developing technologies...The future of SVG not only seems bright, it seems certain to play a major role on the Web in the years to come [2]." "Mark our words," another analyst predicts, "SVG is a standard that will quickly become a key component of Web content [3]." A couple of developers writing in the XML Journal claim that SVG "is likely to revolutionize the way Web graphics are rendered, stored, manipulated, and associated with content [4]."

And, of course, there's always "the next big thing". I did a search on Google for "SVG" and "the next big thing". There were 125 hits. When you get labeled with that, it almost makes you worry. The graveyard of technical innovations is littered with "next big things". And SVG's ambitions are indeed very big. It aspires to nothing less than becoming a new graphics standard for the Internet. Certainly it was designed to be nothing less. SVG's day with destiny is supposed to be everyday. That is, it's supposed to become an integral part of the daily Web experience, one of the many indispensable tools of daily Web development and use.

All this raises the hurdle very high. While I personally believe that SVG will succeed, I don't think success is guaranteed. Perils range from insufficient browser support to too few applications to too little user demand. These perils feed on one another. If too few applications are developed, there will not be enough demand to enable the browsers. And if there are insufficient numbers of SVG-enabled browsers, there will be insufficient incentive to develop. Lack of browser support results now in what one reviewer calls a "hesitancy in the market [5]."

Consider just one notable example of such hesitancy. XML co-creator Tim Bray has developed software to visualize search engine results and it is reportedly able to employ SVG (http://antarcti.ca/index.html). Bray, however, says that he will not SVG-enable his system until there is native support from the "leading browsers" [6]. The fact that SVG does not yet make business sense for a co-inventor of XML is, I believe, a loud and clear signal about how important it is to achieve native browser support.

The Browser Question

What about plug-ins? Adobe has been a great patron for SVG. And Flash is proof that the plug-in strategy can succeed. It appears, however, to be far from an optimal solution for the long-term. In addition to the hassle of a download, there is the question of performance costs resulting from lack of integration with the browser.

If native browser support is what's ultimately required, Microsoft's dominance looms as an issue. Estimates vary but analysts generally believe that Microsoft's Internet Explorer controls well over 70% of the market [7] with some estimates going as high as 89% [8].

It's hard to discern what Microsoft's intentions are. There are rumors and hearsay about Microsoft plans to incorporate an SVG viewer into IE Explorer but nothing substantial. There is also much talk about the death of Microsoft's Vector Markup Language (VML) but Microsoft's 2002 line of products for Front Page, Microsoft Publisher and Visio all include some VML features. Does Microsoft maintain any vested interest in VML [9]?

It is interesting to hear of SVG projects that dovetail with areas of strategic importance to Microsoft. Yesterday, Shane Aulenback noted in his presentation that SVG's integration with XML makes it "a compelling format for use as the front end for Web services" [10]. Hopefully, projects like this will succeed and help convince Microsoft to jump on the SVG bandwagon.

What if Microsoft does decide to integrate SVG in IE Explorer? Would it "embrace and extend" SVG for its own purposes? Would we be better off with a company like Adobe appears to genuinely committed to SVG? Or would native IE support respect the W3C standard and offer precisely the shot in the arm that SVG needs to take-off?

II. COLLABORATING ON SVG PROMOTION AND DEVELOPMENT

Whatever the dynamics of browser support may be, popular demand for SVG will play a decisive role in whether or not SVG succeeds as a standard. And SVG development will be the key to generating such demand.

SVG has a very good fighting chance to succeed largely due to the broad range of tools available to SVG developers and the amazing diversity of potential applications. As developers, we know that SVG is a tool that we can get our hands on easily, to mold it, to make it perform, to make it dance. Presenters at this conference have proven this point. SVG has been on full and glorious display here with presenters pushing the frontiers with SVG's toolbox: JavaScript, DOM, CSS, XML, XSLT, Java, Perl. The list goes on. It is a growing toolbox full of surprises.

We have also seen that, in the realm of applications, you can't put SVG in a box because it will keep on popping out. SVG is not just a mapmaking tool, a data visualization tool, or an animation tool. It is all of these. It is any one of these. And it is more. It is, as demonstrated here, a tool for scientific visualization, for mobile communications, for product design, for tracking trucks on European highways or tracking the genome, for controlling crime in the Paris metro and train system or managing a water supply network. We have even seen how SVG images of heraldic blazons belonging to knights of the Middle Ages can be produced from XML data. And this is no mere exercise in Nordic pride but a contribution to efforts to facilitate graphical content database searches.

As a database developer, I am especially enthusiastic about SVG's potential for data visualization. SVG can make many dead oceans of Internet text come alive and finally succeed as tools of public communication. And it is already clear that cartography is going to be one of the most productive areas of SVG data visualization. SVG mapmaking is a realm where the creative imagination knows no boundaries. The possibility of dynamically generating "smart" maps with SVG opens a new frontier for innovation. While the old pocket maps, wall maps and raster maps remain tremendously useful, programmable vector Web-based maps that, with a click, can morph into animations of geo-social trends, illustrate geo-economic statistics in a host of data visualizations or serve as hubs for Internet communications place cartography on the threshold of a new era.

And let's not lose sight of SVG's educational potential. SVG's full text syntax and basic simplicity let people get off to an easy start before setting sail in deeper waters. It's something that young children can use to publish pictures on the Internet and learning simple basics of coding in the process. It's something that students in an elementary school geography class can learn their first longitude and latitude lessons with.

We have witnessed here the amazing versatility and practical value of SVG - but we have to get the word out. Not enough developers know yet about SVG and not enough people have seen what SVG can do. We need to significantly expand the SVG developer community and the SVG audience. We need to help serious developers learn about SVG and how to get a handle on it.

SVG Advocacy

What are some of the steps we can take? In some cases all it takes is to get a graphics developer who doesn't know SVG to visit the site of Kevin Lindsey (http://www.kevlindev.com/) to have a career-changing Eureka experience. Cartographers can have a similarly transforming experience when visiting, for example, Andreas' Vienna map (http://www.karto.ethz.ch/neumann/cartography/vienna/) -- as geographer Brandon Plewe reported in the forum I set up before the meeting or as Marian de Vries noted in her presentation yesterday morning..

In terms of developer outreach, we can learn an important lesson from Bill Gates. One thing Gates has always understood well is the importance of getting through to developers. In the mid-1980s, Gates used to show up at our NYPC User Group meetings and other large gatherings of developers around the U.S. We should also be proactive in getting through to developers about SVG.

A few months ago, I attended a meeting of a GIS users group and spoke briefly about SVG. Out of about 40 or 50 people, only two, as far as I could tell, had heard about cartography with SVG. We need to go to current user groups, to make presentations and demos.

In the spring of 1975, when I was a student at Stanford, I attended one of the first meetings of the world's first computer user group, the now legendary Homebrew Computer Club. While I had no idea at the time how important that particular user group would be, I have been a member of many other computer user groups since and have learned how critical they are in getting people over what might otherwise be insurmountable problems. What about SVG user groups? Does anyone here know of the existence of an SVG users group [11]? I did a search on Google for "SVG Users Group". Zero hits. I did a search for "Flash Users Group". It came up with 181 hits. When we go back home, let's seek out other SVG enthusiasts and form SVG user groups!

We can take SVG advocacy further with letter-writing campaigns. Take a look at the letters proposed by SVG author Micah Laaker on his site svgnow.com. His letter to Microsoft's IE Product Manager makes the case for native browser support. He also offers a letter encouraging Adobe to keep up its broad support for SVG [12].

Another way to support SVG is by frequenting and participating in online SVG development support forums and, if we can, to help out with questions there. In addition to the Adobe forum (http://www.adobe.com/support/forums/main.html) and carto.net mailist (http://www.carto.net/papers/svg/mail_e.html), there are helpful forums at Yahoo (http://groups.yahoo.com/group/svg-developers/messages) and svg-spot.com (http://www.svg-spot.com/forum/) - and probably some others that I have missed.

In order for SVG to take its place alongside so many other computer development tools, potential SVG developers need to have the opportunity to take classes in SVG. Does anyone here know of an SVG class offering anywhere [13]? As members of the first wave of SVG development, we should be thinking about how we might offer classroom instruction in SVG or help such classes get started.

Collaboration on open source development is another key means by which we can help SVG take flight. On Monday the developers of Apache Batik (http://xml.apache.org/batik/) let us know how we can get involved in that project. As many of you know, SourceForge.com is a major online collaboration system assisting open source development projects (http://sourceforge.net/). I counted over 30 SVG projects on the site. It appears that these projects could use more participation by developers so I encourage you all to go and search out the SVG projects on the site and, if you can, to offer some development assistance.

SVG Killer App?

Is there a killer app for SVG? I believe so. A simple but compelling way to visualize search engine results would, I think, be the application SVG needs to achieve a popular breakthrough. What could be more a powerful way to advance SVG than by bringing to life the now deadbeat text-based heart of the Internet -- search engines -- and reaching millions of people every minute in the process?

This is not a novel idea. As I alluded to earlier, Tim Bray clearly has this objective in mind with his endeavor. It is possible to get started with large or small scale projects of this sort using the RDF-formatted search engine data of the Open Directory Project at http://dmoz.org/rdf/. Open source efforts along these lines could help existing efforts like Bray's reach their goals.

I stumbled across some intriguing news in this regard. On webreference.com, there is a three-part series entitled "Google SVG Search" that documents how to use perl's Soap::Lite package to access the Google Web Services API for the purpose of displaying search results with SVG. During the beta phase of this project, Google is permitting anyone who registers at their site to gain access to their Web Services API for this purpose [14].

III. CITIZEN MAPMAKERS

Another way that we can help promote SVG is to demonstrate the value of SVG to our own professional communities or special constituencies. At Citizen Mapmakers, for example, we are trying to promulgate SVG mapmaking among environmental and citizen groups. We are building maps that, along with the district boundaries of elected officials at various levels, show locations or concentrations of issues that have already galvanized advocacy groups into political action [15]. We hope to convince these groups of the value of SVG maps for Web-based communications and to help adopters keep their maps technically up-to-date.

In the hope that it may be of broader use, here is a "roadmap" for promulgating the use of SVG among a particular community based on our plan at Citizen Mapmakers.

All our source code and scripting is open source. We are opening an online forum where we will post notice of our production efforts, including requests for assistance. We invite all of you to check into the site to download files or to offer development assistance.

One of the initial foci of Citizen Mapmakers is community-level toxic chemical emissions. The Community Right to Know Act of 1986 required that industries annually report to the Environmental Protection Agency detailed information about toxic chemicals they release into the environment as well as their geo-coordinates [16]. By overlaying political districts on maps of facilities with toxic releases, it is possible to identify the elected officials responsible for districts where emission sites are located.

Both the toxics release and political district boundary data are freely available over the Internet [17]. Here is a map that demonstrates what we are trying to do with such data at Citizen Mapmakers (http://www.CitizenComputing.org/mapmakers.html). In terms of map quality, I cannot imagine having two tougher acts to follow - the latest path-blazing work from Andreas Neumann and Navimap's marvelous on-the-fly map of Paris! By contrast, I shall present something rather simple [18]. Simplicity is, however, a virtue when trying to promulgate the use of SVG. Simple, useful and easily replicable applications are probably the best ways to start drawing others into their own creative SVG adventures.

As a demonstration of what can be done with these public data sources, we have produced a map of the state of Louisiana displaying the location of facilities that report to the Toxics Release Inventory as well as layers for the boundaries of Louisiana's U.S. Congressional districts and State Senate districts. When hovering the mouse over a facility and a district, information about the facility and its toxic emissions appear as well as contact information for the elected official in that political district.

The map also conveys information about a problem uniquely associated with Louisiana called "Cancer Alley". A term coined in the 1980s by environmental activists, "Cancer Alley" refers to a pollution-ridden industrial corridor that is home, by some counts, to over 300 major industries. You can see this cluster standing out conspicuously on the map in the southeastern portion of the state, starting in the area south of New Orleans along Lake Pontchartrain, running along or parallel to the Mississippi in a northwesterly direction toward Baton Rouge. Louisiana development officials and industry supporters dispute the cancer association, claiming the area's cancer rates are typical of the rest of the country. The Cato Institute, for example, has called Cancer Alley an "environmental myth". For residents of the area, however, physical ailments are anything but mythical. As Barbara Koeppel noted in her piece in The Nation, "People living nearest the factories and waste dumps are sick and dying. Clusters of asthma, stillbirths, miscarriages, neurological diseases and cancers have mushroomed". Undoubtedly, scientifically reliable health analyses confirming the existence of Cancer Alley would help residents to advance their claims. According to Koeppel, such studies have been hampered, in part, by deficient data classification and sloppy record keeping at the state's Tumor Registry ("Cancer Alley, Louisiana", The Nation, November 8, 1999).

We clearly sympathize with the residents and advocates for whom the term "Cancer Alley" has acquired popular legitimacy. By putting Cancer Alley on the map, we aim to give voice to local complaints, visually associating health experiences with specific geographic locations. Moreover, we wish to demonstrate to the many environmental groups and citizen advocates for whom Cancer Alley is an important issue how they can use SVG to publicize their concerns. More broadly, we hope to show that SVG mapmaking can be an educational and informational tool in many sorts of citizen initiatives.

IV. CONCLUSION

I believe SVG will succeed - but it can use an extra push. Go out and spread the word about SVG. Let educators know about SVG. Go to user group meetings where you live and help more developers find out about SVG. Better, start an SVG users group where you live. Visit the online forums when you need help and, if you can, offer help to others there and in open source collaboration projects. Let's stay in touch and maybe we can even figure out some killer apps for SVG that we can all collaborate on.

Endnotes

[1] Several portions of the following text were, for the sake of brevity, not included in the actual presentation. Also, minor additions were added subsequent to the talk (such as footnotes recounting audience participation and text from the Citizen Mapmakers homepage that was displayed during the talk).

[2] "Scalable Vector Graphics: The Art is in the Code", E. Traversa, 21 November 2001. http://www.webreference.com/authoring/languages/svg/intro/.

[3] "Wrapping up SVG", M. Gibbs, 9 July 2001. Available at http://www.itworld.com/AppDev/328/NWW010709svg/.

[4] "Data Driven Web Graphics with SVG: A Declarative Approach" in XML Journal, P. Jain , S. Komatineni, May 2002, Volume 3, Issue 5.

[5] "The Next Wave for Graphics", B. Trippe, April 2002. Available at http://www.transformmag.com/db_area/archs/2002/04/tfm0204xm.shtml

[6] "Information Insider - Visualizing Data on the Web", R. Boeri, January 2002. Available at http://www.emedialive.com/r19/2002/insider1_02.html.

[7] "Netscape and Internet Explorer battle for the title", R. Baldazo, D. Tynan, 30 October 2001. Available at http://www.cnet.com/software/0-3227883-8-7614087-1.html.

[8] "Explorer Dominates Brower Market, Netscape Dwindles", Staff of SiliconValley.Internet.Com , 23 February 2001. Available at http://siliconvalley.internet.com/news/article/0,2198,3531_597681,00.html.

[9] Chris Lilley, chair of the W3C's committee on SVG, says that the presence of VML in some current products reflects Microsoft's past and not future commitments to VML. Lilley believes that Microsoft has terminated development of VML. The designer of VML represented Microsoft on the SVG 1.0 committee and has since left Microsoft. The W3C is seeking to involve Microsoft in the SVG 2.0 committee. (Personal conversation with the author, July 16, 2002.)

[10] "SVG - The visual interface for Web services", S. Aulenback, 2002, SVGOpen pre-conference abstract. Available at http://www.svgopen.org/abstracts/aulenback_williamson__svg_as_visual_interface_for_webservices.html.

[11] In response to this question, no one reported the existence of an SVG Users Group. Ronan Oger, however, spoke of a meeting of SVG users at Goldman Sachs in New York. It was unclear whether this was actually an ongoing public users group. I noted plans by Micah Laaker and I to form an SVG users group in New York City and challenged members of the audience to beat us in holding a first meeting. Now, we can report that on July 24, 2002 we did in fact hold a meeting of an SVG Users Group at the Cedar Tavern in Greenwich Village and have plans for future meetings. Lacking news to the contrary, we claim to have held the world's first SVG User Group meeting. Others claimants? Please make your case.

[12] Micah Laaker's SVG advocacy letters are available at http://www.svgnow.com/resources/evangelism/evangelism.shtml.

[13] There were four or five affirmative responses to this question from members of the audience.

[14] "Google SVG Search I", M. Classen, 29 April 2002. Available at http://www.webreference.com/xml/column55/. "Google SVG Search II" , M. Classen, 13 May 2002. Available at http://www.webreference.com/xml/column56/. Available at http://www.webreference.com/xml/column57/

[15] Please note that political districts in the United States tend to play a different role in the United States than they do in Switzerland and most other European countries. Rather than the multi-member districts common under proportional representation, in the U.S. a single elected official represents the geographic districts of political bodies like the House of Representatives, state legislatures and city councils. This means that a single official can be held accountable for problems or conditions within a given geographic district.

[16] The 1992 Earth Summit called for international efforts to promote similar Pollutant Release and Transfer Registers (PRTRs). Other countries that have or plan to have PRTRs are Canada, the UK, Norway, the Czech Republic, Japan, and Australia.

[17] The political district map data came from the U.S. Census Boundary Files and the TRI data from the Right-to-Know Network or RTK Net.

[18] By "simple", I mean that this demo map does not push the limits on state of the art SVG cartography as does Andreas Neumann's work on user interaction and Yvonne Isakowski's interactive map browser (http://www.svgopen.org/papers/2002/isakowski_neumann__svg_for_interactive_topographic_maps/index.html) or Navimap's PostreSQL-generated map of Paris (http://www.svgopen.org/abstracts/duquesne_vignard_georges__navimap.html). I do not, however, wish to overstate the simplicity of this example for SVG beginners. The data preparation required basic database programming skills. After storing the ungenerate/ASCII map files in Foxpro tables, Foxpro coding was used to generate an SVG document. Some JavaScript skills were also required to manage mouse-related events and to line-wrap lengthy portions of text. At Citizen Mapmakers, we are exploring simpler, more direct ways to convert ungenerate/ASCII files into SVG documents.

Mike McCullough is co-director of the Citizen Computing Institute, a non-profit organization based in New York City.


Valid XHTML 1.0!