Weblog


Weblog

Huh?

24.07.2003, 00:04 · Martin · Permalink

They wanted to include some information in their feed that isn't defined by the RSS spec, so they created an extension, also known as module or namespace, with information about the search results. An aggregator that was tuned to work with Wired search results could use the extra info, but the feed will still work with aggregators that don't know about their namespace (like all the ones out there now). Modularity. Coolness. Marc Canter asked me to keep stressing this point. He was right. People should know that RSS is extensible.

Now who was recently complaining about 'funky' feeds 'n stuff? Weird... Something must have changed.

ContextGraph similar entries

22.07.2003, 16:36 · Martin · Permalink

Are here! Yes, I finally dug into the 'Contextual Network Graph' Search::ContextGraph stuff and came up with a replacement for the VectorSpace based 'similar entries' engine that was running here for some time now. It tool a lot of tweaking of parameters till it gave similar results to the VS setup I had...
The real fun part about it is this: for much the same result, it's blazingly FAST. Where my VS thingie would take about 40+ min. of CPU time to complete a scan, the CG thing takes about 23 sec. !! And that'S w/o using the cool 'store' functionality that let's you store a ContextGraph for later use (a thing tha isn't technically possible with the VS approach)... And, this Search::ContextGraph is a pure Perl module with very little dependancies... VectorSpace needs a C-based extension (PDL)

So much for now, more details when I get back to breath ;)

[update:] something is weird with it on this server... shouldn't it find VectorSpace? Posts with that word in it should be similar to this one, no? Back to the drawing board, I guess...

[update:2003-07-23] ... it appears the results for VectorSpace and ContextGraph are rather similar if the posts are rather short... but with the size of the post CG fails... digging..

My AOL Journal...

22.07.2003, 10:51 · Martin · Permalink

Well, I thought it time to cash-in on all thos I've been paying my AOL fee w/o ever even as much as reading my mails... so I went and created Dreamtank...

Mini-XML

21.07.2003, 10:48 · Martin · Permalink

Mini-XML is a small XML parsing library that you can use to read XML and XML-like data files in your application without requiring large non-standard libraries. Mini-XML only requires an ANSI C compatible compiler
... "Find" and "walk" functions for easily locating and navigating trees of data

Rather interesting LGPL lib from Mike Sweet, who is at the core of things like CUPS and FLTK, as well as htmldoc etc. It appears to actially be rather small indeed when compared with expat et al...

Neil Gaiman

21.07.2003, 09:45 · Martin · Permalink

Eisner Awards Keynote
Nobody wants a world of identikit comics. Do the comics only you can do. Tell the stories only you can tell. Do not lose sight of the fact that this is an industry that can create real art.

Andrea's Weblog: Two broken arms at the same time

19.07.2003, 22:35 · Martin · Permalink

Yesterday, André managed to trip over something and to fall in such a way that he broke both his arms.

best of wishes go youre way, André and Andrea!

more about feedster

19.07.2003, 00:00 · Martin · Permalink

... or rather the fulltext search-engine that lies at the heart of things. It's OpenSource after all, and in wide use as it seems.
[update 2003-07-24] They decided not to OpenSource it after all

Review of SCRIBUS 1.0

18.07.2003, 12:38 · Martin · Permalink

Scribus is a OpenSource Desktop Publishing tool. The review gives it a bright (potential) future and suggests it for 'lightweight DTP' now... Well it can do PDF nicely, and saves XML natively... I can already see some potential for that...

Feedster: Not Just A Big Pile of PHP Scripts

18.07.2003, 10:24 · Martin · Permalink

So if you ever thought that feedster was something you could pull off in your own 10$ PHP/MySQL enabled webaccount, go read this. Think again. Just because Scott admits his mistakes out in the open, 35 manyears of text-search engine experience between two people and a number of successfull startups sold, I guess you don't copy that in your lunch-breack folks...

BTW: Here's the blog of the second feedster guy, Francois. Link to him!

Linux for Tablet PC

17.07.2003, 17:15 · Martin · Permalink

Lycoris a 'Tablet PC Edition' of it's Linux based OS availeable...
Now the only reason that keapt me from droolling all about the new Tablet PC platform was - apart from the fact I coudn't afford one - the fact that it would force me to stick with Win XP for all time... no longer as it seems...

extending SharpReader, the other way

17.07.2003, 11:26 · Martin · Permalink

Now most people on Windows and into RSS agree that SharpReader is one of the best RSS Guis there is for Windows. But I was recently troubled with the home/office multiple computer problem and the hassle of keeping two Aggregators in synch... I had an idea of how to solve this, and one slightly hidden feature of SharpReader plays very nicely to this.
Ever noticed that SR saves ALL it's data in plain XML-files? All feeds, with all the posts you haven't deleted, with reade/unread status etc. Now this XML format is rather obious, very similar to RSS itself. I was thinking it should be a simple thing to build an alternative UI that uses these files as data. This could be done in PHP, server-side for example. Now if one devises a simple rsync upload of these files to a web-enabled dir on a webserver... Next step would be to write am alternative aggregator that can write the SR format... That would give us an webview (sans the advanced features of SR maybe), but with the possibility to mark feeds as read/unread etc... if we then sync those files back to the SharpReader installation... we could use the superior SR interface (threading etc.) when we have access to it, and use the simpler webinterface if we don't and still keep both synced...

just an idea...

16.07.2003, 15:58 · Martin · Permalink

connecting BlogChat and Jabber...
Class.Jabber.PHP : UsingTheClass

connecting dots via feedster

16.07.2003, 10:47 · Martin · Permalink

From my 'incubator' channels I found this intriguing idea:

Collaborative audio blogging | minimal verbosity

Instead of making the audio the content, why not setup some sort of system where the message is put on the website with a note saying "There is a new message to the blog, would you like to help transcribe it?" Put the audio next to a nice little form and let the visitor transcribe the text.

But, being here at work and not having access to my SharpReader at home I used feedster to re-find that post (this being a prime target for keyword search, as I remembered some keywords very well)

And now look what I find:

WombatNation: Transcription Portlet

is a voice portlet for transcribing telephony-based dictation. This technology makes large-vocabulary speech recognition technology available to telephony-based portal applications (portlets). It provides Java APIs through which developers can integrate transcription capabilities into a portlet application

Ok, that might not be in any way a solution for the original idea, but it's a very nice example of how a blog-.searchengine can be used to connect knowledge. esp. when someone (like me here) blogs that connection...

[Update:] Robert Stewart chimes in to point to some interesting discussion about this on his site. quote: And the connections continue...

TEK Homepage

16.07.2003, 10:32 · Martin · Permalink

TEK supports a different model of Web access. Think of inter-library loan. You need a book. Your library doesn't have it, and you can't afford to buy it. But you can get it on loan. You are willing to `pay' to access the infomation in that book with time: you will wait a week to see the book.

With the TEK search engine, the user submits a query which is emailed to Boston. The TEK Server, which is connected to the Internet backbone, searches the Web, locates some pages, selects which pages to send back, compresses them, and returns them back to the user. Because the search results are returned asynchronously, by email, the connectivity charges are lower. Post-processing the search results and selecting which pages to send back reduces the amount of information and addresses the bandwidth question.

By the way, TEK stands for "Time Equals Knowledge".

unsafe colomatch

15.07.2003, 00:06 · Martin · Permalink

Yeah! with template prview and all!

instant webfrontend to nntprss

14.07.2003, 17:21 · Martin · Permalink

Just a short notice: NewsPortal is a PHP based WebFrontend to read news via the nntp:// protocol (newsgroups)... combined with nntprss, one has an instant WebFrontend to that RSS-Aggregator ;)
I have a setup running localy with PHP reading from nntprss, and a setup where a PHP on my webserver talks to a nntp:// server... but for some reason the PHP on my webserver won't communicate with my nntprss instance on that same server...

holy shit!

14.07.2003, 15:01 · Martin · Permalink

Google must be kidding! I'm number 2 of 3 for "reverse keyword index" MuhwwAhAHhAhHAHAHAHA!

The ctypes module

14.07.2003, 14:43 · Martin · Permalink

ctypes is a Python package to create and manipulate C data types in Python, and to call functions in dynamic link libraries/shared dlls. It allows wrapping these libraries in pure Python.

It works on Windows, Linux and MacOS X (the latter two require that your machine is supported by libffi).

Sorting subscriptions

14.07.2003, 12:40 · Martin · Permalink

While I find the idea of an incubator category/folder nice, what I need even more dearly is a folder with 'feeds' I have already tried and decided NOT to follow, along with a short note of why I decided so. Some feeds just don't cut it for me in the long run and I drop them, no problem. But ever so often I hit a blog I would love to follow in my aggregator, but can't for some reason (truncated feeds is the biggest reason)... So I drop the feed from my shedule and forget why. And a few weeks later I got through the same process aginb, round and round the loop.
Now it wouldn't make sense to simply subscribe to such a feed and put it in some extra folder, from the bandwidth point of view. Better would be to put it in a folder and mark it 'no update for reason xy'... I guess I can do that already...

Tom Gilder's Blog: Blog Editing

14.07.2003, 10:34 · Martin · Permalink

Something I would really like to see in blog software is a "post history" though - similar to what you find in Wikis - that allows you to see any changes I've made.


alles Bild, Text und Tonmaterial ist © Martin Spernau, Verwendung und Reproduktion erfordert die Zustimmung des Authors