MSK Software Development

I moved my development blog to schrievkrom.wordpress.com … my more private posting will stay at www.schrievkrom.de

I started work of an interface to the CouchDB (MSKCouchDB) - a document/json oriented database from the Apache project. You may get further information about CouchDB at CouchDB at Apache.

In addition to this early interface I updated the JSON wrapper to produce correct string representations of JSON objects.

The software is available at vastgoodies as usual.

And as usual - this is work in progress and it will change between the version, but the code gives you the ability to play with that database.

I’ve wrapped a C-based JSON parser for VASmalltalk. The source code is referenced from json.org and I created the dll library from that source available from there - therefore its a Windows only solution up to now.

The library has an interesting aspect - it was the first library where I had a useful example of calling an external library in a synchronous way - and expect callbacks into Smalltalk via EsEntry points.

The MSKJSONWrapper is available at vastgoodies as usual - and the needed DLL project (for Watcom C Compiler - but a dll is also included) is here.

Still amazing, how good recordings can be done using today camrecorders — I recordered all sessions at the VASTForum in Stuttgart this year. My recorder has a build-in memory module of 32 GByte and therefore I decided to switch the recording resolution to a low-level HDTV way: 1440×1080i. That would give me a recording time of around 7-8 hours. Using full 1920×1080i would just result into 4-5 hours.

The overall recording size of this day was about 23 GByte (H264 - MPEG4 AVC) for 9 sessions.

The surprising fact was, that my computer (mainly harddisc) was not able to give me the H264 recorded data in realtime while editing the video stream - therefore the video stream has to be converted to mpg2 (1:4 ratio) before being able to edit it. (As an example: Seaside-Talk: 3.6 GB against 13.3 GB).

The best conference videos I’ve seen so far were from the PDC2008 (Microsoft) - in a format of about 900×500 pixels they manage to show the presenter and the screen in two different windows within the video. They were recorded using two different camers: one for the presenter and one for the beamer-video (or they captured the video stream)..

The PDC2008 videos show a very good quality and therefore were also not small: between 180MB and 500 MB everything was possible. But the results were well done and are far superios than those videos viewable in general at youtube …

Looking at the PDC2008 videos one can see, that the presenters were very well lighted and the background was general dark. The arrangements gives a good contrast and a well looking picture. Camrecorders on the other hand may have problems with the light contrast of the talker and the presented video screen.

The large video screen (for the audience in the room) was far away from the presenter and they never use it directly in their talk (by e.g. going in front of the video screen and trying to show something on the large screen). The light of this screen also has no impact on the video of the presenter - due to the large distance.

The sound is pretty good - of course the audience is much larger - and each presenter has a microphone.

Just to give some experiences back to the community.

And this is the video of the last session: “Questions and Answers” - sorry for the bad sound, but actually I was sitting in the back of this room and I was also not able to hear or understand all words said or all questions asked.

VFM10-09-QA.mp4