Springer, Very exciting your site of examples.
>From my perspective I want to decompose the pieces even further to very basic tiddlywiki. Personally I already use a range of tiddlywiki native representation for connections. Lets say I have a tiddlywiki that documents fruit based recipes and all their relationships. Using tags, link references and back-links, I would favour a way we can use this to generate the data needed for a graphical tool to represent it using its features. Thus the graphing tool can represent native tiddlywiki relationships. Sure custom tags and special fields can help it along for advance features, but these should be avoided, if they can be implemented in a generic way. For example why could we not get the grapical tool to take the image from the icon field, the color from the color field and a tooltip from a tooltip field. The simplest approach is to make a graph generator. A Special tiddler that interrogates the raw tiddlywiki structure and translates it into the graphical language for one or more tools to then visualise. - Just as we have template tiddlers for static html, json, etc.. we should be able to build templates that generate the graphical tools "code language", the only challenging bit is identifying how to do this in a way we can pass titles and wiki text it the desired parts of the graphical representation like links the the relevant tiddlers. The idea solution would be if such a visualisation system (of raw tiddler relationships) came along side the graphical tool itself. Even better if each graphic tool could come with its own visualisation system so all tools can represent raw tiddlywiki data and relationships equally well. Such a solution will also be primed to permit further customisation and leveraging the graphical tools unique features, while keeping the data in raw tiddler forms. I hope this makes sense, I am speaking "conceptual design" here, the details still need work Regards Tony On Wednesday, June 10, 2020 at 1:51:43 AM UTC+10, springer wrote: > > Tony, > > Super! > > I would be very excited if it were possible to build a graphviz-style > result where the graph-building macro grabs a set of tiddler nodes based on > some filtered list, and pulls from some field in each tiddler for a list of > tiddlers (constrained to the filtered set) with "upstream-relation" to it, > and/or draws from some other field for a list of "downstream relations", > and perhaps a third field for non-directional relations. (There's some > redundancy if there's both an upstream and a downstream field, but that > would allow authors to add relations from wherever they notice them, and > bidirectional connections could simply result from both directions being > indicated, though some folks would want a "two-lane" style for > bi-directional connections.) > > The most difficult thing to replicate is the way in which diagramming > software actually treats the connections as entities with properties > (labels, line styles, colors, arrow styles). All the relevant info could in > theory be squeezed into the field of one of the tiddlers (for which the > connection is upstream, downstream, both, or neither). graphviz also allow > for labelled subgraph areas within a larger diagram, and for connections > that do all kind of complex things (whether or not incoming connections > should "merge" like tributaries before reaching their target), and some > tools track priority levels for handling connection lines that might > "cross", etc.). Still, if TW could harness something like the "smart" > layout engine of graphviz or mermaid while treating the nodes as linkable > tiddlers even in an otherwise "flat" diagram, that would be a fantastic > beginning. > > If the diagramming engine is truly integrated with TW, as you say, it > should be easy for the diagrammer to be flexible about what text to display > on the graph for each node (title, text, caption, and icon being the most > likely). Something like a node-shape field would help for flowcharts, > etc... presumably the linkstyle field (from plugin) would automatically > help with coloring label text, as well. > > In case anyone is curious, some examples of diagram styles from my > teaching site are here: http://ethicsatwes.tiddlyspot.com/#diagrams > > -Springer > > On Monday, June 8, 2020 at 11:47:27 PM UTC-4, TonyM wrote: >> >> Springer, >>> >> >> >>> I'm not sure I follow where your answer is going... Is your suggestion >>> that *graphviz* can be modified so as to show links, or are you simply >>> pointing out that there's no conceptual difficulty in the idea of getting >>> any kind of diagram to supervene on (map neatly onto) tiddler relations? >>> >> >> Yes, *any kind of diagram to supervene on (map neatly onto) tiddler >> relations?* but of course this requires a design component to accompany >> each diagramming tool >> >> >>> I do understand that "the data contains all the information necessary to >>> be represented in different graphical ways" ... So my question is whether >>> the graphviz plugin (or any other currently available tool) can be readily >>> modified to do that work... >>> >> >> Yes, I am asking that same question *graphviz plugin (or any other >> currently available tool) can be readily modified to do that work, * and >> suggesting we try and make the answer yes. >> >> >>> >>> (I also don't quite follow why you place special emphasis on "icons" in >>> your reply? >>> >> >> I was perhaps a bit loose with my words here, perhaps iconography, rather >> than icons is better, Whilst simple lines can be used for relationships, we >> can have different graphical representations of relationships and whilst >> these differ according to the graphical system you are using, a basic one >> would be helpful to represent the tiddlers (objects) and relationships in >> raw tiddlywiki. Eg line, directional (one to many many to one), broken or >> loose etc... >> >> >>> Am I missing something about the spirit of your reply? >>> >> >> No you are fairly close, to the spirit of my reply. >> >> I think my key take home is we need to provide a middle layer, between >> tiddlywiki and the graph or imaging solutions we have. If we have a robust >> set of standards any graphical representation can be given a set of macros >> to interact with the tiddlers to that standard. Most of the standards are >> already implied we just need to document them. >> >> Regards >> Tony >> > -- You received this message because you are subscribed to the Google Groups "TiddlyWiki" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/tiddlywiki/9b024f3e-164e-4d0f-be57-949828e01b99o%40googlegroups.com.
