|
LibraryCogit's like, you know, library systems and stuff... |
|
All systems seem to have a trade-off between utility and efficiency at some level. There are many applications where the trade-off is irrelevant, especially in an era of phenomenally cheap processing, but there can be tricky metrics for achieving throughput that can sometimes make or break a system task. I have seen three big bottlenecks in the last year that brought me back to the days when mainframe solutions were common in libraries, and a misaligned application could literally throttle your budget because of the costs of operating it. One of these bottlenecks involves an unnamed library automation vendor where the solution has been to muster an insane amount of processing capacity to mediate the problem. In this case, the cost implications are not that different from the darkest days of the mainframe, but I won't talk about this one, at least for now. One bottleneck concerns the use of Lucene for a number of large scale indexing jobs, and the slowdown here is much less concerning and I think has some fairly simple solutions. The third is related to image searching, and illustrates why the other two are neither particularly suprising nor without learning potential, and why scaling software solutions will continue to help ensure that the world's supply of caffeine will remain in high demand for years to come. Read More... (1 comment, 7622 bytes in body) I am slow on the uptake these days, in part because of this announcement, and my laptop is becoming littered with half finished essays, programs, and other remnants of too much travel and not enough time at home. For someone who never left Canada's East Coast until I was well into my twenties, and only ventured out of North America for the first time last Spring, I am quickly making up for lost time. This is great on a lot of levels and I am involved in projects that are close to my heart, but one of the most rewarding activities has been to have a chance to work on software in the strange land that exists between airport connections, overnight sojourns, and train trips. I have written on WebDAV and desktop integration too many times already, and some of my grumbles from the past are being addressed nicely by others (see this comment on a web clipboard from two years ago as compared to the work that Dan describes). Technologists tend to overuse the term convergence but I really do get the sense that a lot of the desktop barriers to the Web are crumbling faster than ever before. Here is yet another description of the advantages of benevolent Web-desktop interaction, this time found while probing how to use image searching for digitized historical collections while sitting in a hotel in the booming city of Calgary. Read More... ( 8254 bytes in body) I have spent the last week in Beirut, Lebanon, and had a chance to to explore both the city and the country a little. It was my first trip to the Middle East, and I had jumped at the opportunity to go, despite a formidable traveling schedule to get there and back, and some timing that meant that the trip was taking me out of MPOW (really two MPOWs for me) at a less than ideal time. Still, it was a great experience and well worth facing the timing issues on my return. One of the areas that I visited in Lebanon was Baalbeck, in the Bekaa valley, a UNESCO World Heritage site because of the amazing temple ruins from the Roman period. Here are a few notes about Beirut, road trips in Lebanon, shaky Internet connections, and the high potential for epiphanies while standing in the shadow of the Temple of Jupiter. Read More... ( 7410 bytes in body) Before the code4lib conference, I was looking around for some advice on how to unify audio with presentation software in connection to podcasting. I wasn't interested so much in narration support within a set of slides, but instead wanted to record a session while it was happening, and somehow have it synced with the slides automatically. Peter Binkley suggested on the code4lib channel that I use timings, and his approach, like all tips emanating from Peter, made a lot of sense. Peter's concept was to capture the time when each slide changes (using S5, of course), and reconcile the timings with the recorded audio. As long as the clocks on the machine where the recording takes place, and on the server where the slides are hosted, are reasonably close, this method would allow sorting them out afterwards. I started to work on an audio recorder on the trip to the code4lib conference since I figured Peter's suggested process might be enhanced by controlling the recording stream. It also seemed like something I could work on with limited network access between the airport and the plane. I managed to get a recorder of sorts running along the way, but the conference itself made me realize that I wanted more than audio support be worked in. Stir in some ideas from Lawrence Lessig, Dan Chudnov, and a couple of more bits added on the way home, and I think I have the pieces coming together for a different kind of presentation software. Read More... (6 comments , 8722 bytes in body) I am sitting in a hotel beside the airport in Portland after having just attended the wickedly cool code4lib conference. I just uploaded my presentation, which I would have done before the conference if I had finished the edits before getting on the plane. There has been lots of conference coverage captured at the code4lib planet, but here are a few scattered notes before I grapple with the impending reality of a seriously early morning flight. Read More... ( 3293 bytes in body) Blame Ross Singer. Read More... ( 2286 bytes in body) In 1999, I wrote an article for InsideOLITA called The End of the ILS. At the time, I was in the thick of writing my own ILS alternative and had studied what seemed to be emerging trends in enterprise systems, particularly those used by organizations with deep pockets like banks. I did a lot of data modeling, learned a lot about component software, but didn't really appreciate that I was looking in the wrong place for community uprisings against monolithic and inflexible library systems. Herein lies what might be a more effective approach, concerning ways to push library data out on a wider scale and setting up opportunities for it to interact with other systems more frequently. Read More... ( 6116 bytes in body) I am one of the organizers for the code4lib conference in Corvallis, Oregon, and I have been amazed at how well it has come together. I have been on the back end of conference activities more than once, and probably devoted one of the most intense periods of my life to Access 2002. It was fun, and I had a ton of help, but it was also a bit daunting. Maybe that's why the code4lib conference has been such a pleasant surprise. It's true that Jeremy Frumkin has quietly and brilliantly done the logistics pieces that are often the parts of conference organization that can drive you crazy, but code4lib has been remarkably free of stress compared to most of the conferences I have known. The conference lineup and schedule was a true grassroots initiative. Proposals came in, the vetting process was open and democratic, Ross Singer pulled off a stunning last minute feat of web magic to set up the polling, and the results are phenomenal. I agonized over the Access program in 2002, yet the code4lib program practically built itself. I am also perversely proud that one of the organizers, yours truly, just barely managed to get on the program at all, and only made it because of some shifts in the availability of speakers. This is a strong sign that the process was truly open to the participants, and it makes me wonder how far this model of conference organization could go in revitalizing conference events in general. Read More... ( 3650 bytes in body) In our last episode, we figured out how to insert a custom image in Google Maps, and used the world's silliest example of how the Google Maps interface could be adapted and extended. I was convinced that SVG, the XML standard for 2 dimensional graphics, could be really handy for working with Google Maps, but I couldn't figure out how best to make this work in practice. Several iterations and a couple of wrong turns later, I can now report why I think SVG could be your best friend for working with Google Maps, while at the same time possibly providing further evidence of why I could be the worst graphic designer to ever draw breath. Read More... (2 comments , 13401 bytes in body) I love maps. It could be a feeling shared by all directionally challenged people like myself, but maps are amazing creations. A visual tool that can make sense of spatial realities in an instant, maps seem to me to exist at some cosmic intersection of intuition, artistry, and logic. Google Maps make this magic come alive on the web. I have seen other mapping tools that accomplish similar wizardry but Google's interface pushes the limits of what a web browser can accomplish with maps in a widely accessible manner. It's also hard not to be impressed by the many hacks that have been put together using Google Maps, but here's a quick example of a modest hack that seeks to suggest a possible relationship between mapping and library collections, as well as to reveal a secret about some of the participants in the recent Access conference. Read More... ( 4966 bytes in body) |
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
More Search Options 
|
"There are of course many problems connected with life, of which some of the most popular are `Why are people born?' `Why do they die?' `Why do they spend so much of the intervening time wearing digital watches?'" -The Hitchhiker's Guide to the Galaxy, by Douglas Adams | |
| All trademarks and copyrights on this page are owned by their respective organizations. Comments are owned by the Poster. | ||