We do not list the stylesheet verbatim here, however, the interested reader can find it in the TextGrid source code repository. Instead, we give a short summary of what the stylesheet does.
Insert the header and footer a correct GridWorkflowDL document needs
Format a place called “batchinput” with as many tokens as URIs were given in the tgwf document
For each entry in the <inputconstants> section, format a place holding a single token with its contents. The place will be a read-only place, i.e. its tokens will not be consumed
Format a number of default places every workflow has, e.g. for the CRUD's SessionID and LogParameter, input places for the StreamingEditor, and result places
Format the places between the services proper, e.g. for each <link>
Format a CRUDread transition: this will consume the URI and return data (to be handed over to the services proper) and metadata (to be given to the StreamingEditor).
Format a transition for every service given under <activities>
Format a transition for the StreamingEditor. The SE is used to create the metadata of the new TextGridObjects. It takes as input the old metadata and an XSLT stylesheet with rules.
Format a CRUDcreate transition. It will take as input the processed (text) data and the new metadata and will return the real metadata of the new TextGridObjects (i.e. with real URIs).
To get a better understanding of the resulting GridWorkflowDL document, we show the same graph from figure 2 in the Monitoring section , but now annotated with the meanings of the various places and control flow in figure 3 below.
Figure 3: visualization of data flow in the workflow graph