Thursday, June 23, 2005

Sakai, XOSIDS, and cross-language issues #

I have learned that the web services support in Sakai 2.0 is only an embryonic framework, not the working product for which I was hoping. The announcement of web services in Sakai 2.0 made me hope for something like Flickr Services, which has enabled interfacing to Flickr from Java, PHP, Python, Perl, Actionscript, Ruby, .NET, among other languages.  Maybe a full suite of web services  in Sakai will be forthcoming soon. On a related front, I've been wondering how the release of the XOSIDs will enable the development of "OKI-compliant" Python tools. Wilbert Kraan answers a few of my questions with his piece CETIS-OKI and IMS, wires and sockets revisited (emphasis mine): As a result, the time it takes to generate OSIDs for languages other than Java has been slashed considerably. Some OSIDs have already been ported to languages such as PHP, C# and Objective C, and are used in production systems. Note that implementing an OSID in such a language doesn't mean that it will then be interoperable with OSID adapters that have been written in other languages. Language bridges such as the Mac world's widely used Objective C to Java bridge needs to be used for such a purpose. Still, the XML based nature of the XOSIDs also opens the possibility of the creation of new XOSIDs as part of the new IMS specification workflow. In the latter, a new specification's datamodel and behaviour are specified in abstract UML first, and then transformed to the XML Schema definition and WSDL files that Web Service enabled systems use to interoperate. Deriving OKI style API descriptions (i.e. XOSIDs) at the same time should be perfectly possible, provided that the IMS spec fits with the existing OSIDs and how they are factored. Which of course raises the older question of how OSIDs relate to Web Services in practice. Part of the answer will be demonstrated at alt-i-lab in Sheffield, were a fair number of search tools will be coupled to an even larger number of repositories via the repository OSID. The not-so-secret sauce that makes this possible is the fact that each of the relatively simple and innovative search tools only has to implement the one OSID, and that each repository needs to supply only one OSID adapter to the search tool. I will write Wilbert for some followup to his last intriguing paragraph. (Thanks to Scott Leslie for pointing to Wilbert's article.)   Posted by Raymond Yee on 6/23/05; 5:05:26 PM from the Educational Technology dept. Discuss

A bit on why I've started taking so many pictures #

Yesterday, I wrote about how I want to begin my days with early-morning reflections. Because of the need to get out of the house early after a lot of sleep and an early morning meeting, I'm only now getting to my coveted reflection time. I have a strong preference to writing "in the open" whenver poosible, though clearly there are times for private meditation and writing. I do happen to have somethings to share with the world. Today, I am pausing at the same spot I did yesterday with the question of what to write when there is so much to write. I will take the same strategy of writing a bit about my picture-taking since it's become an integral part of my daily life. I will also cut myself off without exhausting the topic. One question I've been pondering is why I'm so into taking pictures and presenting them in Flickr. At the beginning, I was just playing around, experimenting with the features of Flickr. (There is, of course, a whole pre-history of my getting into digital photography in July 2000, a story for another time.) To experiment with image collections, I figured that I should have some of my own pictures with which to work. My early experiments with Flickr coincided with my getting a cameraphone. I started taking pictures with my Treo 600 on the premise that the cameraphone would be an easy way to create a collection of meaningful images, even if the images were not about grand themes but only about my own daily life. I did set out to create images along certain thematic lines (e.g., Milosz, garden plants around Berkeley, Chinese art.) But so many other things caught my attention that I essentially started "life logging/life blogging" (to echo the name of commercial products for managing a lot of mobile media.) Along the way, my relationship with my images and my camera evolved. At first, I was largely content with the poor VGA level Treo 600 pictures from my camera phone, but once I started carrying around a 5 MP Pentax Optio S5i, I would use my camera phone as a last resort. With my 5 MP camera, I started to generate too much data to sustain a daily process (overloading the processor time, bandwidth, and disk storage I had for daily picture taking). Consequently, I have set my 5 MP camera to take VGA pictures by default! (I'd like to come back to the issue of how many pixels are enough for my purposes....) Still, at the current rate, I am generating more data than I can process satisfactorily (i.e., at the very least tag). I continue taking pictures, thinking of the process as gathering raw data for further processing recontextualization at some later time. In midst of taking daily pictures, I have become a "daily media producer" (a concept I learned from Marc Davis' Garage Cinema Research Group.) For someone who never thought of working with images in any serious way, I'm astounded to learn how pictures have become a major part of my vocabulary for telling stories. I've had to take on technical challenges to produce images that begin to be along the lines that I want. (e.g., creating a rigorous time sequence for my images, calculating some amount of geo-referencing, writing my own uploading tools). Now, is any of my picture taking relevant to research and teaching? I think so, and I will want to articulate reasons why some other time. Many researchers are working with collections of their own digital images and need to use and reuse them in various contexts. So I believe a lot of what I'm trying to accomplish with my own picture taking, sharing, and archiving will help us to provide useful services to researchers and teachers on this campus.   Posted by Raymond Yee on 6/23/05; 4:42:54 PM from the Unclassified dept. Discuss

Biotagging our plant pictures #

For the Interactive University's City|Watershed Project, Doug Kern has uploaded over 1700 pictures of plants to his Flickr account. It seems that they are all tagged with genus/species information, which represents a substantial amount of detailed work. Currently, it is unclear which tags correspond to the genus and which to the species. For the pictures have been placed in sets labeled for the genus and species, it is possible to distinguish the genus from the species. I am considering indicating more specifically the genus and species information by tagging plant pictures with tags like bio:genus and bio:species (e.g., bio:genus=Spartina and bio:species=foliosa) Such tagging is akin to the use of geo:lon and geo:lat in the recent geotagging craze in Flickr. Hopefully, before there is massive amount of tagging, I'd want to do a careful study of taxonomic markup languages in use to propose a standard for genual use. Such a study would also include looking the conventions already being adopted in Flickr) To experiment with such tagging, I would copy over some of the images in Doug Kern's collection so that I can experiment with them in my own Flickr space. Now why should one go through the trouble of formally tagging plant imagery with the genus and species? One major motivation is the prospect of making precise links to sources of detailed information about the plant once one has a precise name. For example, consider one of Doug's images of Spartina foliosa: Spartina foliosa flower on Flickr One can link to the National PLANTS Database by using its advanced query capabilities to determine that the Spartina foliosa has been given the USDA symbol SPFO: Plant Profile for Spartina foliosa (California cordgrass). Moreover, this species has a taxonomic serial number of 41275: ITIS Standard Report Page: Spartina foliosa. I don't quite understand the importance of the connections I have laid out here. I don't know who uses the symbols or the taxonic serial number. Is the number akin to an ISBN for books (in some really rough way?). At the very least, the Jepson Herberaium has a system that makes use of the ITIS number: UCJEPS: Jepson Interchange: Spartina foliosa Trin. I figure that On-line Resources is a good place to look next to answer more of my questions. A sidenote: Calflora also provides information about Spartina foliosa: Calflora: Spartina foliosa. Note that Calflora demands registration for most of my readers. Calflora - Obtain An Account: "You must register to use Calflora unless you are under 13, in a K-12 school, or at a public library."   Posted by Raymond Yee on 6/23/05; 12:58:33 PM from the Unclassified dept. Discuss