james g wrote:
The situation for computer programmes may well be even worse. If history is anything to go by, programmes may simply no longer run on newer operating systems after a while. The source code may have to be reverse engineered from the programme in order to retarget the programme for a modern platform, and this could be a real mission.
This is due to a fundamental design flaw in current operating systems, where the monolithic kernal or microkernal requires frequent updates rather than going with a modern exokernal core.
I have a cloud exokernal OS project called CHAOS, which is currently designed to get around the need for code retargeting just to update commands and functions, in fact it will only require a Java base, like Android, to host the core and then load from dynamically updated files. In essence, the platform does very few things very well, and leaves all the rest to the external system.
This is not new technology, unless you are a system engineer, in which case, 1990 doesn't have to be the source of the most advanced technology anymore.
When mainstream operating systems address this issue, and previous legacy code is no longer required, this may be fine, but in the meantime I think this is still a serious issue.
Well, Linux has a monolithic kernel, OS X has a micro kernel and Windows is completely fubar in it's kernel.
OS X is basically the best so far, with late 80's technology at the core. Linux is based on early 80's core design.
By comparision, exokernel designs have only been around for a decade or so, and allow for system function updates outside the core.
This allows for system updates to happen without restarts, without having to recode applications to take advantage of new functions.
In theory, Linux could fork over to an exokernel design easily (within a year), as there is enough academic examples to draw from.
OS X would take longer, possibly selling it as OS XI, as they would have to transfer the GUI links to the new exocore.
For this project and a few others, I would be happy to disclose the source. For some others, I am not going to disclose source code.
This is due a issue of why should I research and invest money in new programs, if I have to allow other people to piggyback on my work without contributing to my effort at all?
Ultimately, I am happy with a 5 year commercial copyright term and any academic credit arising from my research.
What I can't afford is to have to stay on a minimum wage job just because every talentless script kiddie gets my code automatically.
Lastly, I have a computational chemistry program which I am doing for my own research.
If source disclosure is the legal requirement for copyright grants, then I have to share the code to a purely private research project.
This would mean my personal research in the field would be up for grabs by any university or company which wants the information without my compensation.
I believe that I should be in the loop when my research is used, although it seems I may have to discuss this with victoria university soon.