Date: Tues, 20 Feb 96 12:13:42 From: jburka@geog.umd.edu Subject: NSDI_L SUM: metadata tools To: nsdi_l@fgdc.er.usgs.gov Cc: dt11@umail.umd.edu, march@oriole.umd.edu A few weeks ago, I made the following posting: >I'm involved in a clearinghouse building effort and am currently >examining tools for metadata compilation. > >So far, I've had a look at Document.aml 7.0.4, the currently availble >version of BLM's rewrite of Document, Montana State Library's Document, >the NOAA powerbuilder app, and I'm in the process of getting xtme to >compile on my system. I've also looked over the documentation to Geolineus >3.0, though I don't have the actual software. > >I'm curious as to which of the available tools folks are actually using and >how satisfied you are with your choice. I'd also like to know to what >extent folks are using mp to verify files and then generate html to serve >from their clearinghouse nodes. > >Any comments? This generated a reasonable response, approximately 10 messages, ranging from "Please summarize" to "I use brand x." Below, I'll summarize the meatier responses and throw in some of my own impressions of the tools I've worked with to date. Amongst the folks who shared their own findings, I was somewhat surprised by the mostly universal distaste toward the bulk of the tools. One reason for this appears to be that much testing was done before the current crop of tools and/or versions was available. For instance, someone complained that the blmdoc tool only worked with Arc/Info 6, though I've been quite successful running it under Version 7.0.3. A number of folks are using mp for validation of their metadata (actual files or output by one of the database systems), while others felt that it's simply *too* particular and that it's not worth the effort to clean up output files. To an extent, though, that's as much a problem with the need for machine, not human readable output. For instance, mp will produce a flurry of warnings and errors when it process output from Montana State Library's (MSL) "datadictionary.aml" (dd) because that tool assumes that a couple of contact data elements will have the same values, and thus inserts "see above" type entries. This is fine for human-readable output, but in a situation where an IR system is searching specific tags (reference the recent nsdi_l/isite-l thread regarding hierarchical tag searching), will fail. Let's see. I had a couple of respondents tell me they used a word processing template. Another used xtme. Another used a unix rewrite of corpsmet (a simple pc-based metadata "editor") called "makemeta". I had no responses from folks who said that they were actually making use of Document.aml, blmdoc, or the MSL tool, though I would hope that at least *someone* at BLM is using blmdoc, and that the Montana State Library is using data dictionary. I also had no response from people using the NOAA tool, and I was never able to actually get this package to run on either of the two PCs I tried it on. If anybody has successfully installed and run the NOAA PowerBuilder tool straight from the zip file on the NOAA ftp site, I'd love to hear how you configured everything, especially the ODBC drivers. I also heard nothing from users of GeoLineus, though one respondent commented that it looked neat, but too expensive to consider using at this point. --- I've been playing around with the Arc/Info based tools a fair amount over the last couple of weeks, and feel that the tools have definite advantages and disadvantages. After getting the newest version of Document.aml from ESRI (7.0.4), I can say that the tool and its level of compliance has been greatly improved. Still, I was disappointed to see how poorly the output passes through mp. It starts generating errors almost from the start, due to the lack of a "Citation_Information" tag between "Citation" and "Organization" (the first data field of the compound Citation_Information). You can't entirely blame this on ESRI -- if you look at, for instance, the _Content Standards for Digital Geospatial Metadata Workbook_ (3/24/95), you'll see in the examples that the Citation_Information tag is not used between Citation (compound) and Organization (data element). While it may not be necessary for output to pass mp, it would be nice to see all of the necessary tags included. Still, there's nice flexibility on such things as contact and distribution information, as well as the new feature for direct generation of HTML files (which are structured much like those generated by mp). Data Dictionary has, by *far* the best designed GUI of these three tools, as well as some features not included in other programs, such as the ability to digitize a G-POLYGON for the dataset, and the avoidance of using a text editor such as emacs for entry of a long list of items. Other nice touches are the main menu (which allows the user to graphically select which coverages to work with) and the facilities for small thesaurae for themes and sources, allowing the user to easily insert commmonly used information. The package has two major drawbacks. The first is a fairly long list of unsupported tags (which may or may not be important to most users; I believe they all fall into the optional category). The second is a very limiting method of storing site information, as well as a non-compliant method of outputting some of this information, which can cause confusion in examining warning and error messages generated by mp. I'd like to point out that I'm very grateful that a list of unsupported tags was provided! The single best thing I can say about the BLM tool is: jburka@gis [240]mp -e blm.err -h blm.html blm_test.txt jburka@gis [241]ls -l blm.err -rw-r--r-- 1 jburka geog 0 Feb 15 13:42 blm.err which is to say that the output passes through mp without generating so much as a warning bit. The GUI is basically the same as Document's, but has been cleaned up a bit, as have the help files. My biggest complaint is that the automation of Entity_and_Attribute_Information has been replaced by a few template lines in the editable text file (giving the user only the Overview_Description to fill out). I'm curious as to whether there's going to be another public release of the BLM tool. Sol Katz recently sparked some discussion with his questions about how to deal with a few specific metadata fields, so he appears to still be working on it, but the most recent version available dates back to last August. --- I've also fiddled around a bit with two stand-alone tools, Peter Schweitzer's xtme and NOAA's Metadata Toolkit. xtme is a fairly basic tool which does what it needs to -- allows the user to enter or remove tags as necessary, and enter data for data element tags. The largest advantage to xtme, which is not the most user-friendly package, is that its output is guaranteed mp compliant. Compilation under solaris was a snap. NOAA's Metadata Toolkit was a bit harder to get working, though my problem was solved by Larry Claflin at NOAA. The tool is quite easy to use, and has a pretty good interface. To an extent, it feels like what so many folks were looking for last August at the Charleston CCAP metadata workshop -- a series of straightforward forms to be filled in, clearly following the order used in the metadata workbook. The current output format is a bit too simplistic (a straight text file in which all lines are flush left), but thanks to the inclusion of element numbers, it would be fairly easy to write a script which would strip out the numbers and replace them with the requisite number of tabs/spaces. If you're trying to run the Toolkit and getting ODBC errors, you might want to have a look at the important bits from Larry's e-mail: >In the c:\winloc\odbc.ini file insert the following: > >[META_JOEL] >Driver=c:\winloc\system\wod40w.dll >Database=c:\wsql40\win\metadata.db >Start=db32w -d >DatabaseFile=c:\wsl40\win\metadata.db >DatabaseName=Metadata >Autostop=yes >UID=dba >PWD=sql > > >In the c:\winloc\odbcinst.ini file make sure the following lines are there:> > >[WATCOM SQL 4.0=Installed] > >[WATCOM SQL 4.0] >Driver=c:\winloc\system\wod40w.dll >Setup=c:\winloc\system\wod40w.dll --------- Many thanks to those who responded to my initial post: Chuck Stein Mike Scott Ken Shaffer Reid Campbell Sol Katz Andrea Musone Hugh Phillips Bruce Westcott Bruce Potter as well as Cherie Griffith at ESRI who helped ensure that I had the most recent version of Document.aml. Thanks are also due to the folks who are making metadata editing/ creation tools available to the public. If you have any questions, feel free to write me. Jeff |Jeffrey C. Burka | Pithy, insightful quote to be inserted when one | |jburka@glue.umd.edu | occurs to me. *If* one occurs to me. | |http://www.wam.umd.edu/~jeffy/html/home.html |