- Hardware Accelerated indirect rendaering is unsuppoerted. - want to load "client-side" DRI drivers in the server - interface between libglx and libGLcore(in server) is very differnet from interface between libGL and DRI driver(in client)
- new generation XML configuration and code-generation.
Labor not cleanly divided between libglx and libGLcore - creation of GL visuals and drivers's internal representation mixed togher - reflects the pre-pbuffer nature of the code
- entire process is very convoluted.
Generate server side glx protocol code - solution: use same XML to generate the server side protocol code/ can generate up to 26000 of 30000 lines of server side protcol code. addtion: server side optimization.
- dispach table to call GL funtions. - it reduce size and load time. same as client-side.
impacted server side code. non-unix libGLcore implementations.
eache change to interface will break non-unix libGLcore implementation. darwin / cygwin ports must use dispatch table and new API.
IHV support. ML, keep aware of future API changes.- old interface is removed in X.org 7.0.
At the head of this week, OASIS, the international e-business standards consortium, annouced formally that member approved the Open Document Format for Office Applications (OpenDocument) v1.0 as an OASIS Standard.
This news had already Slashdoted in 4th May, but in Japan there are a few articles proviced by IT-oriented online news media such as IT media, Japan Internet.com and Nikkei IT PRO. Some blogs such as Placebo Effect also touch the story.
Most of Japanese market leaders seems not to understand the meaning of this news.
I'm now working for makeing guidelines of a government software/hardware procurement.
We had discussed and decided that government should make effort to procure a standard-based technology. It means if Linux have become an international standard, Gov. might make a RFP which refer that standard.
The situation may be same on the document format side. Now Open Document format becomes Defacto standard refferd with authority as same as SGML, XML and HTML.
So Gov. will refer it in RFPs.
This is exactly a BIG thing for not only software vendors and system integrator, but OSS community, I think.
FSIJ announced CodeFest Japan 2005 (only in Japanese).
I want to join it and hack kernel ACPI subsystem and IIIMF issue.
It is over-night hacker event that will be held first time in Japan. Recently in Asia-OSS and Beijing OSS Week, primary sponsor San-Wah and AIST held a Beijing CodeFest. I was also there at that time, but participated only 1-hour.
I had also been Korea at 3rd Dec. 2004 to attend to Northasia OSS promotion Forum. At same time and same place, CodeFest 2004 Korea had been held.
Every time, I had been recommended to join there by Mr. gNiibe, one of FSIJ leaders.
In this paper the author, Thaweesak Koanantakool, Ph.D, present an abstract of the input method and keyboard history in Thai. I think we should consider not only mother language but also the other languages.
A nessecity of Input method is common even for Eulopean. Because they have some accent or other special charactors to edit with ascii or localized keyboard.
Now X implementation uses XKB keyboard extension for such a situation, but in straight point of view it should be consolidated in the IM framework.