geoLibro.org

James Fee posts some news from ESRI 2008 User Conference and one piece of it is half welcome:

Q: When will the metadata editor in ArcCatalog be improved?

“I’d say this is very welcomed.  The metadata editor hasn’t improved since 8.x was released and if we are to expect people to edit metadata, it should be made easier.”

–SpatiallyAdjusted.com

Improving the metadata editor is a fine idea, but how about integrating metadata management into the rest of the workflow? How hard could this be? When I add a new column, why don’t I have the option of documenting it? Why don’t I have the option (though as a librarian I might argue it should be mandatory) to at least write an abstract every time a new feature class is manually generated? And why is there no metadata toolbox in Toolbox? It has certainly not been made easy to write good metadata — by any GIS software project, by the way — so I’m not sure why ESRI is so sanctimonious about it.

I guess it’s no coincidence that just as ArcDex was ruminating on the gross and saddening similarities between Microsoft and ESRI the AGX Team gleefully released a preliminary look at a future ArcGIS Explorer interface that looked, well…like a Microsoft Office product. Which is really fortunate, because if anything is going to move AGX out of the realm of a pathetic joke of an afterthought of a farcical not-a-competitor-to-Google-Earth competitor to Google Earth, it’s a rotund, ornately-menued, visually complicated Office clone. Jesus Christ. Bravo, ESRI.

Tobin at Fuzzy Tolerance looked over the ESRI Developer’s Summit schedule and noticed that “the only ArcIMS session at the developer’s conference is about how to stop using ArcIMS.” And good riddance, friend. Now thousands of heavy, cumbersome, ugly sites will atrophy and be forced into an overhaul that might, if we’re all lucky, raise the floor of what geo* users will accept as a legitimate mapping site.

My favorite example of people trying to teach old dogs new tricks.

Hey, ESRI. I just thought of a really good idea for ArcGIS Explorer. First, launch a multi-step wizard every time somebody wants to add data to a map. Really you could stop there, because I would love that. But if you really want me to enjoy using AGX please go ahead and make it so that when somebody wants to add N layers, they’ll get N separate wizards and they’ll have to click through every fucking one of them rather than have a little box there that will “apply to all” or something. I know it’s only two steps, but maybe later you can add more. Yokay?

The National Geospatial Digital Archive project has a beta of an interface up at http://clients.alexandria.ucsb.edu/ngda and it’s got some fairly innovative things going on. Especially the inline display of download options and link-outs to metadata. They’re using a Google Maps framework, but they’ve done some nice customization and it sounds like their middleware will support alternatives. Not much data in there, but I don’t think that’s the primary focus yet.

This is going to work out fine, yes. I’ve been toying with the notion that the open source ka-Map might just be smooth enough (built on top of MapServer’s robust enough) to handle some of the web mapping projects we have coming up. ka-Map is a javascript and php-built front end that takes MapServer’s output and tiles it, precaches it, then renders it within any number of web environments, a number of which come pre-built with the source. Think Google Maps to MapServer’s ArcIMS. So after a couple of false starts, one of which was due to a mysterious incompatibility somewhere within the string of libraries required to run it on OS X (it set up with no problem on Ubuntu and one of the grad assistants got it set up on Windows with a little text editing), everything fell into place this past weekend. Now in just a couple of days I have map attributes (shapefile) query and search running, have those queries then throwing out to some other database.

…And back! What’s the point? Well, Google recently attached map searches to their Book Search results. MetaCarta has been leading the pack on the same thing. What we have going here is something a little like that, but without the great power of the former and the awesome natural language processing of the latter. Specifically, we’re supposed to have an old soil survey of Tippecanoe County, IN scanned, its map scanned and rectified and, in fact, digitized, and ka-Map is going to allow us to run searches back and forth between the text of the survey and the data layers extracted from the map. In our minds it will be a great combination of library work and GIS and ka-Map is getting us closer every couple of weeks.

I love this, love this:
The BBC is reporting a story about a letter that was addressed with nothing but a map and delivered as such. Article here.

This story renews my desire to compile user-built, hand-drawn, fake, or otherwise alternative. I intend to call the collection “Punk Maps.” Maybe I should copyright that right now. Anyway, I just came across a book that may have done this, already, anyway. I’m sure there are others (feel free to submit).

MetaCarta’s GeoTagger apparently does excellent, impressive, cool work on unstructured documents. The result of its awesomeness is geotagged xml with which one can do whatever one pleases. Make sense? I spoke to a sales rep from MetaCarta at the Indianapolis “What’s New with ArcGIS 9.2″ seminar about GeoTagger and grilled him about what could be done. Not too many specifics from that, but then I got a call from the great MetaCarta themselves since I expressed interest. Unless they’re mean-spirited liars, GeoTagger does do natural language processing and does put out very useful xml and does handle versioned data and does include user-specified structured data (think MySQL tables) and does…do more still.

But today’s post is a leech off of Fantom Planet’s post about another slick product built on top of GeoTagger. It’s PageMapper, and you can think of it as a more invasive, more thorough GeoTagThings. Sounds like it’s a little lab-y still, but have you built such a system?

So the most innovative GIS company extant today really is MetaCarta. Their GTS (Geographic Text Search) makes me very jealous and rue the day I didn’t somehow learn how to develop computer applications and run a GIS company and then think to combine both of them. It’s built on what they call a Geographic Reference Engine, which uses natural language processing to identify geographic references and their contexts within existing text documents. The output is legible from within Google Earth and ArcMap, but my guess is that it could be made to be read by almost anything that’s hip these days.

This is a big deal, if not in practice (I don’t know anybody who has purchased and/or used their GTS), then certainly as a proof-of-implementation that should be of interest to those institutions who are collecting massive amounts of text content electronically and then developing interfaces and other connectors. Say, like, Purdue University Libraries’ Distributed Institutional Repository.

MetaCarta also made GeoTagger, which is essentially a geocoder of existing text collections that outputs xml for use in other contexts.

Anyway, on to the buried lead:
FantomPlanet notes that MetaCarta now has posted an online georeferencing utility that allows anybody with an electronic, unreferenced map document to reference it to its proper place on earth using a series of user-selected control points. There are copyright considerations here, of course (uploading anything means anybody can use it), but at its simplest it’s further proof that MetaCarta gets what’s happening: geospatial information is infiltrating. It’s becoming a common point of access, a common perspective from which we access our documents of any kind. So we’re not imagining, we’re seeing our world of information become mapped, and become accessible by using those maps as interfaces. Not even analysis, which seems like the hangup for so long. That is, GIS for so long was a tool for analysis and data visualization. Now it’s a user interface as well, and librarians who have an interest in normalizing and logicalizing (uh, “to make logical”?) access to information should be pleased. I am, if you take “pleased” to mean “jealous.”

We first saw MetaCarta’s Geographic Text Search back in 2005. There must be some new functionality or something (document density map, maybe?), because the All Points Blog and others are covering it again. I’m jealous of MetaCarta’s effort more than its existence. It’s the sort of business that a library-based GIS might attend to (especially as more and more library faculty are ushered toward doing interdisciplinary research [with, say, computer scientists]) and it is the sort of project I would really like to develop as a module or other functional component of Purdue’s Institutional Repository project, wherein GIS isn’t analytical but rather an information application almost in and of itself. MetaCarta is using a map interface for non-map, non-GIS document clusters and we should all be interested (but librarians especially) in how these documents are ingested and indexed. If they can do automated geographic indexing of 10,000 documents per day, why don’t more libraries have map interfaces or at least some other geographic utility for locating materials?