Weblog
Weblog
the second installment, this time building games upon the 'Shadow Graden' system, where the players shodow on a projection is analysed by a camera and sofware to make inovative gameplay.... wow
Finding Similar Images
Sometimes, an image travels multiple paths before it ends up on my disk, and thus gets saved under different names. But that's a waste of disk, so I want to eliminate duplicates where I can. At first I was using a simple tool that computed the MD5 hash of each image in my collection, and eliminated the duplicates easily.
But frequently, the image has been scaled or maybe even re-rendered at a different JPEG quality, or converted from JPEG to PNG. In that case, the actual bits are not the same, even though the image is nearly the same on my screen.
Notabene: this is Perl code, only a little Image::Magick to read the images...
And then there is pcomp (and the KDE GUI frontend kpcomp for it)
And let's not forget the stunning imgSeek. Awesome, if you happen to have Linux somewhere. Also a webinterface/deamon thingie is planned
Swampwolf - Howl
Howl is a cross-platform implementation of the Zeroconf networking standard. Zeroconf brings a new ease of use to IP networking.
Google's new ~ operator
Curious about what synonyms Google is using for its new synonym operator? Try searching for your synonym keyword and then exclude the same keyword from results, like so:
~help -help = guide, manual, faq, support, tutorial, helping, tips, problems
~search -search = finder, listings, searcher, database
... hmm that might a fine thing to combine with the 'relevat words' feature you see below each post herelll
SourceForge.net: - WideStudio
WideStudio is an Integrated Development Environment (IDE) for building window applications in C++, Ruby, Python an Perl for Windows, MacOSX, Linux, FreeBSD, SOLARIS and other unix. This is open and free (under MIT/X Consortiun Licence).
ongoing · Unflash
We hired a local designer to pitch in on the first Flash-ified version of the product. The results were stunning; areas on the map were shaded oval jelly-beans, it all moved when you clicked on it, descriptive text flickered in and out of existence in a fixed location on the screen. It made a terrific demo.
But we found that real customers doing real installations ended up going for the HTML version every time.
all the best
to you, Jeff, in your new full-time (if shared) job!
posting from Knoppix
I just booted my Laptop from the Knoppix 3.2 CD and everything works, just plain works. That was easier than SuSE or Mandrake or anything I tried before. Recommended cool
Another CSS image replacement technique
While the Fahrner Image Replacement technique that is used there is a very handy trick, it does require special meaningless markup as a hook.
Babylon 5 Books and Novels
as I'm currently deeply immersed into the Bybylon 5 universe - watching season 1 & 2 on DVD - and trying to get the full scope of the background...
- 190 Bester Place - a very comprehensive backgroundsite
- Crusade Writer's Bible Review - there seemes to be a limited edition of this in print at some time
- Aha. well, it was availeable around 2k sez JMS
more on Text Summarization
over at LaughingMeme Kellan compares some approaches (OTS, MACOS X built-in and Classifier4J)
Also contained in the comments: a webapp pointer for Classifier4J
Also contained in the comments: a webapp pointer for Classifier4J
motivator
nothing is as good a motivator to keep digging as a nice link to one's work. Thanks Maciej!
offline at home...
What started as a dying fan has grown into a severe case of data loss... My home PC is currently in a deep state of coma and I'm not sure when or if it will be revived. One big scare of yesterday has been averted tho: it looks as if the HD and with it three years of work and adresses, emails etc. might still be salvageable. Not sure if I will need a new HD, but it sure looks as if my trusty Athlon900 mainboard will need to be replyced. Silent upgrade through desaster ;)
Statistical Translation
A language-translation that uses statistical models has been developed by University of Southern California computer scientist Franz Josef Och.
The system uses matchup patterns to find the most likely translation, instead of explicit grammatical rules and traditional dictionary lists. Multiple human translations allow the computer to check its rendering by a scoring system.
Could this be something similar to whar MeaningfulMachines.com announced earlier?
Open Text Summarizer
OTS is both a library and a command line tool. Word processors such as AbiWord and KWord can link to the library and summarize documents while the command line tool lets you summarize text on the console. The program can either print the summarized text as text or HTML. If HTML, the important sentences are highlighted in context. The program is multi lingual and works with UTF-8 encoding.
DEVONagent
Automatically queries multiple search engines Collects, analyses, weights and summarises the search results Crawler auto-surfs your favourite web pages Presents the result as a dynamic digest based on an intelligent topic list Automatic filter eliminates junk and page-not-found errors Integrated web browser based on the Safari engine Stores results in a searchabe web page archive Flexible plug-in architecture Built to seamlessly integrate with Mac OS X and DEVONthink
To me this sounds a lot like the 'personal proxy' ideas that have been floating around for some time now...
DEVONthink
powerful search technology provides instant, ranked search results and lets you broaden your search by proving similar words based on spelling or context. For example, similar words based on context for "Hubble" could be "telescope", "space" and "astronomy", where as similar words based on spelling could be "hubble", "Hubbles" and "hubbles", but also "bubble". Dynamic keywords and a fully automatic see-also function based on content similarity are only some of the many features that make DEVONthink really "think".
Portable Full Text Engine Revisited
So the C-based engine that drives Feedster will not be released under an OpenSource license. Sad, but very understandable
Linguistic convergence?
I've noticed an odd phenomena. When searching Google I consistently see results from projects I'm involved in, people I know, and of course, myself.
Kellan raises some very interesting points in that post... esp if you take the 'similarity search' idea to the extremes....
evhead says:
Most of the people who I've heard, anectdotely, say that they think blogs sometimes get undue ranking, mention coming across their own blog for searches. They don't necessarily complain about about coming across other blogs. Could that in part be because: a) They're using phrases to search that are natural to them (and, therefore, the words they use to write) but that aren't quite as likely to be used by others as they might think? b) They're not going to learn anything from their own blog, so they consider that result useless, while another blog's contents they may find valuable?
HolaryHey!
Seems the ContextGraph similarity-engine is working like it should now... thanks to some valuable input from Maciej, we are now seeing really compatitive results from both approaches (VectorSpace anf ContextGraph)... and it seems in some cases CG is now even superior at finding stuff... Sadly my two scripts differ too far in the architecture and strategy to really compare results, but I can say that I will drop VectorSpace soon in favor of the ContextGraph... (One down point is the fact that the exact change to the code that made the big difference also is a performance hog... up from 23 sec to about 7 min for one run now... far better still than the 40+ min. the vs.pl runs)
alles Bild, Text und Tonmaterial ist © Martin Spernau, Verwendung und Reproduktion erfordert die Zustimmung des Authors