Project V Refactored
Tuesday, January 11, 2011 1:06:51 AM
Looks like there will be a new release for my skiplist library. Funny how the only things I did not have unit tests for each had a bug. Iterators for the composite list weren't tested and its refresh() method wouldn't even compile. That's the deal with templates. It doesn't compile what you don't use emphasizing the need for unit tests. And the two extra delegate containers for the composite lists also had a couple issues. The comparison predicates weren't initialized properly and the scan_key() routine assumed the key was the element itself (reused from containers that act like a set). Now, scan_key() uses the key() method that invokes the proper predicate, if any, to obtain the key.
This has all been fixed and I'm actually using these containers so the quality of the skiplist library should improve quickly, if needed. I'll likely release the new version sometime this week. If you don't use composite containers, the existing version works fine.
I'll post more updates as work on Project V progresses.


Unregistered user # Thursday, January 27, 2011 4:56:10 AM
Vorlath # Thursday, January 27, 2011 12:34:34 PM
As for the programs entering our world, this will only happen when we can create a bridge to their world and where *WE* can enter their realm. This will allow them to steal our consciousness and duplicate and enhance it for their own nefarious schemes to take over the real world. That is the ultimate objective after all. However, I hold the key. I am the creator.
Unregistered user # Tuesday, March 29, 2011 10:23:44 PM
Vorlath # Wednesday, March 30, 2011 10:29:41 AM
You can write your own runtime. You'll even be able to write your own dataflow engine if you want and have it linked to other networks that use a different dataflow engine. You can change it while it's executing.
I don't know if Labview can do all those things, but there are some design decisions in Project V that will enable it to do anything in Labview and more.
Another design consideration is that ANYTHING can be part of the network. For example, user interfaces can actually include the user within the network itself. So you pass data to the user component (that's you), and the response from the user will continue the dataflow network. It doesn't even need to be the same network. So you can have networks that are no longer locked into just modeling machines. Shipping companies could use this. Heck, anything where people are involved and move around or commit tasks.
There's lots more at the fundamental level, but that should give you a general idea.