GraphSpace - a graph programming IDE / News: Recent posts

The deep dark secret of the GraphSpace (GS) interpretor is that it is... well just that, an interpretor written in another language, and to get anything low-level done, GS has to use that host language. Yet there is an even deeper darker secret that we may have to employ an army of developers to hide, and that is the use of text to do the low level actions that execute when a node is visited and "Run." Well now it's out, and we should admit that graphs are good for some things and text is good for others. The real mission of GS is to use the graphs in the right places and use text for... writing the book. OK, so we are rooting for the graph end of things, but that is because we've been so ignored and we have so many text based languages. It has become a cause to work for. To try to balance the scales, and maybe the pendulum should swing across the mid line for a while. Miro Samek, wrote in his book "Practical Statecharts in C/C++", <em> 'The specific value of visual modeling lies in tapping the potential of high bandwidth spatial intelligence, as opposed to lexical intelligence used with textual information.' </em> Or why can't we use both, together, the fungus and the algae, work together to make lichen :(sorry that was nto the pritiest of analogies:).... read more

The GraphSpace interpretor is now able to dynamically add a new nodes during run time. It can also request information about adjacent nodes and their edges. More development is in the works to allow the interpreter to have full CRUD (creating, reading, updating and deleting) access to the interpreted graph. Not all of this is implemented at this time, but once it is, it will let data structures to be resized and restructured during run time. Until now the language has been limited (similar to scratch.mit.edu) in that there is no allocation of data nodes on the fly. It will probably be a good feature to let the developer control to what extent the graph can self restructure. For instance a graph may state read/copy/add/write/delete permissions for any set of nodes and edges, but for now we don't have this. more later.

A simple interpretor is being written in PHP, with other languages to follow. The interpretor looks for a start node and follows three steps before transitioning on to the next node, until an end node is reached. The three steps that are run on each visited node are: First, "Get" data from other nodes, via edges marked 'get'; Second, "Run" code or process the data; and Third, "Set" (or send out) data via edges marked 'set'. The interpretor then chooses the next node by testing edges marked 'transition' and evaluating a guard condition. The process of get, run, set and transition is repeated until a node marked 'end' is encountered.